You can subscribe to this list here.
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(1) |
Oct
(122) |
Nov
(152) |
Dec
(69) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
(6) |
Feb
(25) |
Mar
(73) |
Apr
(82) |
May
(24) |
Jun
(25) |
Jul
(10) |
Aug
(11) |
Sep
(10) |
Oct
(54) |
Nov
(203) |
Dec
(182) |
| 2004 |
Jan
(307) |
Feb
(305) |
Mar
(430) |
Apr
(312) |
May
(187) |
Jun
(342) |
Jul
(487) |
Aug
(637) |
Sep
(336) |
Oct
(373) |
Nov
(441) |
Dec
(210) |
| 2005 |
Jan
(385) |
Feb
(480) |
Mar
(636) |
Apr
(544) |
May
(679) |
Jun
(625) |
Jul
(810) |
Aug
(838) |
Sep
(634) |
Oct
(521) |
Nov
(965) |
Dec
(543) |
| 2006 |
Jan
(494) |
Feb
(431) |
Mar
(546) |
Apr
(411) |
May
(406) |
Jun
(322) |
Jul
(256) |
Aug
(401) |
Sep
(345) |
Oct
(542) |
Nov
(308) |
Dec
(481) |
| 2007 |
Jan
(427) |
Feb
(326) |
Mar
(367) |
Apr
(255) |
May
(244) |
Jun
(204) |
Jul
(223) |
Aug
(231) |
Sep
(354) |
Oct
(374) |
Nov
(497) |
Dec
(362) |
| 2008 |
Jan
(322) |
Feb
(482) |
Mar
(658) |
Apr
(422) |
May
(476) |
Jun
(396) |
Jul
(455) |
Aug
(267) |
Sep
(280) |
Oct
(253) |
Nov
(232) |
Dec
(304) |
| 2009 |
Jan
(486) |
Feb
(470) |
Mar
(458) |
Apr
(423) |
May
(696) |
Jun
(461) |
Jul
(551) |
Aug
(575) |
Sep
(134) |
Oct
(110) |
Nov
(157) |
Dec
(102) |
| 2010 |
Jan
(226) |
Feb
(86) |
Mar
(147) |
Apr
(117) |
May
(107) |
Jun
(203) |
Jul
(193) |
Aug
(238) |
Sep
(300) |
Oct
(246) |
Nov
(23) |
Dec
(75) |
| 2011 |
Jan
(133) |
Feb
(195) |
Mar
(315) |
Apr
(200) |
May
(267) |
Jun
(293) |
Jul
(353) |
Aug
(237) |
Sep
(278) |
Oct
(611) |
Nov
(274) |
Dec
(260) |
| 2012 |
Jan
(303) |
Feb
(391) |
Mar
(417) |
Apr
(441) |
May
(488) |
Jun
(655) |
Jul
(590) |
Aug
(610) |
Sep
(526) |
Oct
(478) |
Nov
(359) |
Dec
(372) |
| 2013 |
Jan
(467) |
Feb
(226) |
Mar
(391) |
Apr
(281) |
May
(299) |
Jun
(252) |
Jul
(311) |
Aug
(352) |
Sep
(481) |
Oct
(571) |
Nov
(222) |
Dec
(231) |
| 2014 |
Jan
(185) |
Feb
(329) |
Mar
(245) |
Apr
(238) |
May
(281) |
Jun
(399) |
Jul
(382) |
Aug
(500) |
Sep
(579) |
Oct
(435) |
Nov
(487) |
Dec
(256) |
| 2015 |
Jan
(338) |
Feb
(357) |
Mar
(330) |
Apr
(294) |
May
(191) |
Jun
(108) |
Jul
(142) |
Aug
(261) |
Sep
(190) |
Oct
(54) |
Nov
(83) |
Dec
(22) |
| 2016 |
Jan
(49) |
Feb
(89) |
Mar
(33) |
Apr
(50) |
May
(27) |
Jun
(34) |
Jul
(53) |
Aug
(53) |
Sep
(98) |
Oct
(206) |
Nov
(93) |
Dec
(53) |
| 2017 |
Jan
(65) |
Feb
(82) |
Mar
(102) |
Apr
(86) |
May
(187) |
Jun
(67) |
Jul
(23) |
Aug
(93) |
Sep
(65) |
Oct
(45) |
Nov
(35) |
Dec
(17) |
| 2018 |
Jan
(26) |
Feb
(35) |
Mar
(38) |
Apr
(32) |
May
(8) |
Jun
(43) |
Jul
(27) |
Aug
(30) |
Sep
(43) |
Oct
(42) |
Nov
(38) |
Dec
(67) |
| 2019 |
Jan
(32) |
Feb
(37) |
Mar
(53) |
Apr
(64) |
May
(49) |
Jun
(18) |
Jul
(14) |
Aug
(53) |
Sep
(25) |
Oct
(30) |
Nov
(49) |
Dec
(31) |
| 2020 |
Jan
(87) |
Feb
(45) |
Mar
(37) |
Apr
(51) |
May
(99) |
Jun
(36) |
Jul
(11) |
Aug
(14) |
Sep
(20) |
Oct
(24) |
Nov
(40) |
Dec
(23) |
| 2021 |
Jan
(14) |
Feb
(53) |
Mar
(85) |
Apr
(15) |
May
(19) |
Jun
(3) |
Jul
(14) |
Aug
(1) |
Sep
(57) |
Oct
(73) |
Nov
(56) |
Dec
(22) |
| 2022 |
Jan
(3) |
Feb
(22) |
Mar
(6) |
Apr
(55) |
May
(46) |
Jun
(39) |
Jul
(15) |
Aug
(9) |
Sep
(11) |
Oct
(34) |
Nov
(20) |
Dec
(36) |
| 2023 |
Jan
(79) |
Feb
(41) |
Mar
(99) |
Apr
(169) |
May
(48) |
Jun
(16) |
Jul
(16) |
Aug
(57) |
Sep
(32) |
Oct
|
Nov
|
Dec
|
|
From: Nicholas N. <nj...@so...> - 2023-03-23 08:15:25
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=3f8494c8a4aaeb0176c38d22bb52ba221930454d commit 3f8494c8a4aaeb0176c38d22bb52ba221930454d Author: Nicholas Nethercote <n.n...@gm...> Date: Tue Mar 21 11:04:25 2023 +1100 Just use one decimal place for all percentages. Simpler, shorter, and good enough in practice. Diff: --- cachegrind/cg_annotate.in | 7 +-- cachegrind/tests/ann2.post.exp | 36 +++++------ cachegrind/tests/ann3.post.exp | 136 ++++++++++++++++++++--------------------- 3 files changed, 87 insertions(+), 92 deletions(-) diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index e7d094b50d..ac2f2b792a 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -528,12 +528,7 @@ class CcPrinter: perc = " (n/a)" else: p = cc.counts[i] * 100 / summary_count - # Use just one decimal place for large percentages. - if abs(p) < 100: - w = 2 - else: - w = 1 - perc = f" ({p:.{w}f}%)" + perc = f" ({p:.1f}%)" else: perc = "" diff --git a/cachegrind/tests/ann2.post.exp b/cachegrind/tests/ann2.post.exp index 8d33c9e79a..4428346813 100644 --- a/cachegrind/tests/ann2.post.exp +++ b/cachegrind/tests/ann2.post.exp @@ -18,30 +18,30 @@ Dw Dr Ir 18,005 (100.0%) 4,057,955 (100.0%) 5,229,753 (100.0%) PROGRAM TOTALS -------------------------------------------------------------------------------- -Dw Dr Ir file:function --------------------------------------------------------------------------------- - 3 (0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) a.c:main -4,543 (25.23%) 17,566 (0.43%) 47,993 (0.92%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:do_lookup_x -3,083 (17.12%) 5,750 (0.14%) 28,534 (0.55%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:_dl_lookup_symbol_x - 8 (0.04%) 5,521 (0.14%) 28,136 (0.54%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-tunables.c:__GI___tunables_init -2,490 (13.83%) 5,219 (0.13%) 21,821 (0.42%) /build/glibc-OTsEL5/glibc-2.27/elf/../sysdeps/x86_64/dl-machine.h:_dl_relocate_object - 0 5,158 (0.13%) 25,408 (0.49%) /build/glibc-OTsEL5/glibc-2.27/string/../sysdeps/x86_64/strcmp.S:strcmp +Dw Dr Ir file:function +-------------------------------------------------------------------------------- + 3 (0.0%) 4,000,004 (98.6%) 5,000,015 (95.6%) a.c:main +4,543 (25.2%) 17,566 (0.4%) 47,993 (0.9%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:do_lookup_x +3,083 (17.1%) 5,750 (0.1%) 28,534 (0.5%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:_dl_lookup_symbol_x + 8 (0.0%) 5,521 (0.1%) 28,136 (0.5%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-tunables.c:__GI___tunables_init +2,490 (13.8%) 5,219 (0.1%) 21,821 (0.4%) /build/glibc-OTsEL5/glibc-2.27/elf/../sysdeps/x86_64/dl-machine.h:_dl_relocate_object + 0 5,158 (0.1%) 25,408 (0.5%) /build/glibc-OTsEL5/glibc-2.27/string/../sysdeps/x86_64/strcmp.S:strcmp -------------------------------------------------------------------------------- -- User-annotated source: a.c -------------------------------------------------------------------------------- -Dw Dr Ir +Dw Dr Ir - 1 (0.01%) 0 2 (0.00%) int main(void) { - 1 (0.01%) 0 1 (0.00%) int z = 0; - 1 (0.01%) 2,000,001 (49.29%) 3,000,004 (57.36%) for (int i = 0; i < 1000000; i++) { - 0 2,000,000 (49.29%) 2,000,000 (38.24%) z += i; - . . . } - 0 1 (0.00%) 6 (0.00%) return z % 256; - 0 2 (0.00%) 2 (0.00%) } + 1 (0.0%) 0 2 (0.0%) int main(void) { + 1 (0.0%) 0 1 (0.0%) int z = 0; + 1 (0.0%) 2,000,001 (49.3%) 3,000,004 (57.4%) for (int i = 0; i < 1000000; i++) { + 0 2,000,000 (49.3%) 2,000,000 (38.2%) z += i; + . . . } + 0 1 (0.0%) 6 (0.0%) return z % 256; + 0 2 (0.0%) 2 (0.0%) } -------------------------------------------------------------------------------- -Dw Dr Ir +Dw Dr Ir -------------------------------------------------------------------------------- - 3 (0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) events annotated + 3 (0.0%) 4,000,004 (98.6%) 5,000,015 (95.6%) events annotated diff --git a/cachegrind/tests/ann3.post.exp b/cachegrind/tests/ann3.post.exp index 52c5b87503..063715aa6f 100644 --- a/cachegrind/tests/ann3.post.exp +++ b/cachegrind/tests/ann3.post.exp @@ -18,48 +18,48 @@ A SomeCount VeryLongEventName 100,000 (100.0%) 100,000 (100.0%) 0 PROGRAM TOTALS -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName file:function --------------------------------------------------------------------------------- -70,491 (70.49%) 90,491 (90.49%) 0 ann3-basic.rs:f0 -15,000 (15.00%) 600 (0.60%) 0 ann3-basic.rs:f1 - 9,000 (9.00%) 6,000 (6.00%) 0 ann3-could-not-be-found.rs:f1 - 2,000 (2.00%) 100 (0.10%) 0 ann3-basic.rs:f2 - 1,000 (1.00%) 500 (0.50%) 0 ann3-via-I.rs:f1 - 1,000 (1.00%) 300 (0.30%) -1,000 (n/a) ann3-past-the-end.rs:f1 --1,000 (-1.00%) 0 0 ann3-negatives.rs:neg3 --1,000 (-1.00%) 0 0 ann3-negatives.rs:neg2 - 1,000 (1.00%) 0 0 ann3-more-recent-than-cgout.rs:new - 1,000 (1.00%) 0 0 ???:unknown - 500 (0.50%) 0 0 ann3-basic.rs:f6 - 500 (0.50%) 0 0 ann3-basic.rs:f4 +A SomeCount VeryLongEventName file:function +-------------------------------------------------------------------------------- +70,491 (70.5%) 90,491 (90.5%) 0 ann3-basic.rs:f0 +15,000 (15.0%) 600 (0.6%) 0 ann3-basic.rs:f1 + 9,000 (9.0%) 6,000 (6.0%) 0 ann3-could-not-be-found.rs:f1 + 2,000 (2.0%) 100 (0.1%) 0 ann3-basic.rs:f2 + 1,000 (1.0%) 500 (0.5%) 0 ann3-via-I.rs:f1 + 1,000 (1.0%) 300 (0.3%) -1,000 (n/a) ann3-past-the-end.rs:f1 +-1,000 (-1.0%) 0 0 ann3-negatives.rs:neg3 +-1,000 (-1.0%) 0 0 ann3-negatives.rs:neg2 + 1,000 (1.0%) 0 0 ann3-more-recent-than-cgout.rs:new + 1,000 (1.0%) 0 0 ???:unknown + 500 (0.5%) 0 0 ann3-basic.rs:f6 + 500 (0.5%) 0 0 ann3-basic.rs:f4 -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-basic.rs -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName +A SomeCount VeryLongEventName -- line 2 ---------------------------------------- - . . . two - . . . three - 5,000 (5.00%) 500 (0.50%) 0 four - 5,000 (5.00%) 100 (0.10%) 0 five - . . . six -70,491 (70.49%) 90,491 (90.49%) 0 seven - . . . eight - 110 (0.11%) 9 (0.01%) 0 nine - . . . ten - . . . eleven - 200 (0.20%) 0 0 twelve - 200 (0.20%) 0 0 thirteen - 100 (0.10%) 0 0 fourteen - 0 0 0 fifteen - 0 0 0 sixteen - 0 0 0 seventeen - 0 0 0 eighteen - 499 (0.50%) 2,000 (2.00%) 0 nineteen - 300 (0.30%) 0 0 twenty - - 7,100 (7.10%) 100 (0.10%) 0 <counts for unidentified lines in ann3-basic.rs> + . . . two + . . . three + 5,000 (5.0%) 500 (0.5%) 0 four + 5,000 (5.0%) 100 (0.1%) 0 five + . . . six +70,491 (70.5%) 90,491 (90.5%) 0 seven + . . . eight + 110 (0.1%) 9 (0.0%) 0 nine + . . . ten + . . . eleven + 200 (0.2%) 0 0 twelve + 200 (0.2%) 0 0 thirteen + 100 (0.1%) 0 0 fourteen + 0 0 0 fifteen + 0 0 0 sixteen + 0 0 0 seventeen + 0 0 0 eighteen + 499 (0.5%) 2,000 (2.0%) 0 nineteen + 300 (0.3%) 0 0 twenty + + 7,100 (7.1%) 100 (0.1%) 0 <counts for unidentified lines in ann3-basic.rs> -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-more-recent-than-cgout.rs @@ -71,49 +71,49 @@ A SomeCount VeryLongEventName @ Annotations may not be correct. @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -A SomeCount VeryLongEventName +A SomeCount VeryLongEventName - . . . one -1,000 (1.00%) 0 0 two - . . . three - . . . four + . . . one +1,000 (1.0%) 0 0 two + . . . three + . . . four -- line 4 ---------------------------------------- -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-negatives.rs -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName - - 2,000 (2.00%) 2,000 (2.00%) 2,000 (n/a) one - -1,000 (-1.00%) -1,000 (-1.00%) 0 two - . . . three - . . . four - 999,000 (999.0%) 0 -150,000 (n/a) five --1,000,000 (-1000.0%) 0 150,000 (n/a) six - . . . seven - . . . eight - . . . nine - -10,000 (-10.00%) 0 10 (n/a) ten - 10,000 (10.00%) 0 -20 (n/a) eleven - . . . twelve - . . . thirteen +A SomeCount VeryLongEventName + + 2,000 (2.0%) 2,000 (2.0%) 2,000 (n/a) one + -1,000 (-1.0%) -1,000 (-1.0%) 0 two + . . . three + . . . four + 999,000 (999.0%) 0 -150,000 (n/a) five +-1,000,000 (-1000.0%) 0 150,000 (n/a) six + . . . seven + . . . eight + . . . nine + -10,000 (-10.0%) 0 10 (n/a) ten + 10,000 (10.0%) 0 -20 (n/a) eleven + . . . twelve + . . . thirteen -- line 13 ---------------------------------------- - -2,000 (-2.00%) -1,000 (-1.00%) -990 (n/a) <counts for unidentified lines in ann3-negatives.rs> + -2,000 (-2.0%) -1,000 (-1.0%) -990 (n/a) <counts for unidentified lines in ann3-negatives.rs> -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-past-the-end.rs -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName +A SomeCount VeryLongEventName -200 (0.20%) 100 (0.10%) 0 one - . . . two - . . . three +200 (0.2%) 100 (0.1%) 0 one + . . . two + . . . three -- line 3 ---------------------------------------- -- line 18 ---------------------------------------- -300 (0.30%) 100 (0.10%) 0 <bogus line 20> -300 (0.30%) 100 (0.10%) 0 <bogus line 21> -200 (0.20%) 0 -1,000 (n/a) <bogus line 22> +300 (0.3%) 100 (0.1%) 0 <bogus line 20> +300 (0.3%) 100 (0.1%) 0 <bogus line 21> +200 (0.2%) 0 -1,000 (n/a) <bogus line 22> @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ @@ -129,9 +129,9 @@ A SomeCount VeryLongEventName -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-aux/ann3-via-I.rs -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName +A SomeCount VeryLongEventName -1,000 (1.00%) 500 (0.50%) 0 one +1,000 (1.0%) 500 (0.5%) 0 one -------------------------------------------------------------------------------- The following files chosen for auto-annotation could not be found: @@ -140,7 +140,7 @@ The following files chosen for auto-annotation could not be found: ann3-no-such-file.rs -------------------------------------------------------------------------------- -A SomeCount VeryLongEventName +A SomeCount VeryLongEventName -------------------------------------------------------------------------------- -84,100 (84.10%) 94,700 (94.70%) 1,990 (n/a) events annotated +84,100 (84.1%) 94,700 (94.7%) 1,990 (n/a) events annotated |
|
From: Nicholas N. <nj...@so...> - 2023-03-23 08:15:24
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=05d01cd6817613331b35955c84b23757e27db392 commit 05d01cd6817613331b35955c84b23757e27db392 Author: Nicholas Nethercote <n.n...@gm...> Date: Tue Mar 21 11:58:36 2023 +1100 Change `Threshold:` to `Thresholds:`. Because all the thresholds other than the first one were always 100, due to historical reasons. Diff: --- cachegrind/cg_annotate.in | 17 ++--------------- cachegrind/tests/ann1.post.exp | 2 +- cachegrind/tests/ann2.post.exp | 2 +- cachegrind/tests/ann3.post.exp | 2 +- cachegrind/tests/diff.post.exp | 2 +- 5 files changed, 6 insertions(+), 19 deletions(-) diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index ac2f2b792a..48bf4f1aab 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -233,12 +233,6 @@ class Events: # Like `sort_events`, but indices into `events`, rather than names. sort_indices: list[int] - # Threshold percentages, one per sort event. Dictates when we stop printing - # functions. Positions correspond to positions in `sort_events`. Only - # `thresholds[0]` is actually used for thresholding, for historical - # reasons. - threshold_percs: list[float] - def __init__(self, text: str) -> None: self.events = text.split() self.num_events = len(self.events) @@ -269,11 +263,6 @@ class Events: self.sort_indices = [event_indices[event] for event in self.sort_events] - # The primary sort event gets the --threshold value, and all other sort - # events get 100% (i.e. ignored). - self.threshold_percs = [100] * len(self.sort_events) - self.threshold_percs[0] = args.threshold - def mk_cc(self, text: str) -> Cc: # This is slightly faster than a list comprehension. counts = list(map(int, text.split())) @@ -561,7 +550,7 @@ def print_header(desc: str, cmd: str, events: Events) -> None: print("Events recorded: ", *events.events) print("Events shown: ", *events.show_events) print("Event sort order:", *events.sort_events) - print("Thresholds: ", *events.threshold_percs) + print("Threshold: ", args.threshold) if len(args.include) == 0: print("Include dirs: ") @@ -598,9 +587,7 @@ def print_flfn_ccs( threshold_index = events.sort_indices[0] # Convert the threshold from a percentage to an event count. - threshold = ( - events.threshold_percs[0] * abs(summary_cc.counts[threshold_index]) / 100 - ) + threshold = args.threshold * abs(summary_cc.counts[threshold_index]) / 100 def meets_threshold(flfn_and_cc: tuple[Flfn, Cc]) -> bool: cc = flfn_and_cc[1] diff --git a/cachegrind/tests/ann1.post.exp b/cachegrind/tests/ann1.post.exp index 10196377a7..e946b0a743 100644 --- a/cachegrind/tests/ann1.post.exp +++ b/cachegrind/tests/ann1.post.exp @@ -7,7 +7,7 @@ Data file: cgout-test Events recorded: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw Events shown: Ir I1mr ILmr Event sort order: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw -Thresholds: 0.1 100 100 100 100 100 100 100 100 +Threshold: 0.1 Include dirs: User annotated: Auto-annotation: on diff --git a/cachegrind/tests/ann2.post.exp b/cachegrind/tests/ann2.post.exp index 4428346813..a4884e6152 100644 --- a/cachegrind/tests/ann2.post.exp +++ b/cachegrind/tests/ann2.post.exp @@ -7,7 +7,7 @@ Data file: cgout-test Events recorded: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw Events shown: Dw Dr Ir Event sort order: Dr -Thresholds: 0.1 +Threshold: 0.1 Include dirs: User annotated: a.c Auto-annotation: off diff --git a/cachegrind/tests/ann3.post.exp b/cachegrind/tests/ann3.post.exp index 063715aa6f..a5520b5af3 100644 --- a/cachegrind/tests/ann3.post.exp +++ b/cachegrind/tests/ann3.post.exp @@ -4,7 +4,7 @@ Data file: cgout-test3 Events recorded: A SomeCount VeryLongEventName Events shown: A SomeCount VeryLongEventName Event sort order: A SomeCount VeryLongEventName -Thresholds: 0.5 100 100 +Threshold: 0.5 Include dirs: ann3-no-such-dir ann3-no-such-dir-2 ann3-aux diff --git a/cachegrind/tests/diff.post.exp b/cachegrind/tests/diff.post.exp index 81fbe3e920..7d3b8aeeee 100644 --- a/cachegrind/tests/diff.post.exp +++ b/cachegrind/tests/diff.post.exp @@ -5,7 +5,7 @@ Data file: cgout-diff Events recorded: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw Events shown: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw Event sort order: Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw -Thresholds: 0.1 100 100 100 100 100 100 100 100 +Threshold: 0.1 Include dirs: User annotated: Auto-annotation: on |
|
From: Nicholas N. <nj...@so...> - 2023-03-23 08:15:21
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=b8a90ee9682b5dbed81d4dd8ee2d7bc3b77eb906 commit b8a90ee9682b5dbed81d4dd8ee2d7bc3b77eb906 Author: Nicholas Nethercote <n.n...@gm...> Date: Mon Mar 20 13:58:04 2023 +1100 Improve formatting of percentage columns. Currently their width is mostly hard-wired in a quick and dirty fashion. This commit does them properly, so: - all columns are always the right width, even ones with really large percentages - things like `( 1.00%)` are now `(1.00%)` - any percentages that would involve a division by zero now show as `(n/a)` rather than `( 0.00%)` Diff: --- cachegrind/cg_annotate.in | 99 +++++++++++++++------------- cachegrind/tests/ann2.post.exp | 28 ++++---- cachegrind/tests/ann3.post.exp | 144 ++++++++++++++++++++--------------------- cachegrind/tests/cgout-test3 | 2 +- cachegrind/tests/diff.post.exp | 16 ++--- 5 files changed, 148 insertions(+), 141 deletions(-) diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index 20969f0f92..e7d094b50d 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -456,10 +456,6 @@ def read_cgout_file() -> tuple[str, str, Events, DictFlfnCc, DictFlDictLineCc, C return (desc, cmd, events, dict_flfn_cc, dict_fl_dict_line_cc, summary_cc) -def safe_perc(m: int, n: int) -> float: - return 0 if n == 0 else m * 100 / n - - class CcPrinter: # Note: every `CcPrinter` gets the same `Events` object. events: Events @@ -467,9 +463,14 @@ class CcPrinter: # Note: every `CcPrinter` gets the same summary CC. summary_cc: Cc - # The width of each event column. For simplicity, its length matches - # `events.events`, even though not all events are necessarily shown. - widths: list[int] + # The width of each event count column. (This column is also used for event + # names.) For simplicity, its length matches `events.events`, even though + # not all events are necessarily shown. + count_widths: list[int] + + # The width of each percentage column. Zero if --show-percs is disabled. + # Its length matches `count_widths`. + perc_widths: list[int] def __init__(self, events: Events, ccs: list[Cc], summary_cc: Cc) -> None: self.events = events @@ -488,61 +489,67 @@ class CcPrinter: min_cc.counts[i] = count # Find maximum width for each column. - self.widths = [0] * len(events.events) + self.count_widths = [0] * len(events.events) + self.perc_widths = [0] * len(events.events) for i, event in enumerate(events.events): - # Get widest of the min and max, accounting for commas that will be - # added, and a possible percentage. - width = max(len(str(min_cc.counts[i])), len(str(max_cc.counts[i]))) - width += (width - 1) // 3 - if args.show_percs: - width += 9 # e.g. " (12.34%)" is 9 chars. + # Get count and perc widths of the min and max CCs. + (min_count, min_perc) = self.count_and_perc(min_cc, i) + (max_count, max_perc) = self.count_and_perc(max_cc, i) - # Account for the event name, too. - self.widths[i] = max(width, len(event)) + # The event name goes in the count column. + self.count_widths[i] = max(len(min_count), len(max_count), len(event)) + self.perc_widths[i] = max(len(min_perc), len(max_perc)) def print_events(self, suffix: str) -> None: for i in self.events.show_indices: - # +1 is for the single space between columns. - print(f"{self.events.events[i]:{self.widths[i] + 1}}", end="") + # The event name goes in the count column. + event = self.events.events[i] + nwidth = self.count_widths[i] + pwidth = self.perc_widths[i] + empty_perc = "" + print(f"{event:<{nwidth}}{empty_perc:>{pwidth}} ", end="") print(suffix) - def print_count(self, i: int, text: str) -> None: - print(f"{text:>{self.widths[i]}}", end=" ") + def print_count_and_perc(self, i: int, count: str, perc: str) -> None: + nwidth = self.count_widths[i] + pwidth = self.perc_widths[i] + print(f"{count:>{nwidth}}{perc:>{pwidth}} ", end="") - def print_cc(self, cc: Cc, suffix: str) -> None: - for i in self.events.show_indices: - nstr = f"{cc.counts[i]:,d}" # commify - if args.show_percs: - if cc.counts[i] != 0: - # Try our best to keep the number fitting into 5 chars. This - # requires dropping a digit after the decimal place if it's - # sufficiently negative (e.g. "-10.0") or positive (e.g. - # "100.0"). Thanks to diffs it's possible to have even more - # extreme values, like "-100.0" or "1000.0"; those rare case - # will end up with slightly wrong indenting, oh well. - p = safe_perc(cc.counts[i], self.summary_cc.counts[i]) - normal = -9.995 < p < 99.995 - perc = f" ({p:5.{2 if normal else 1}f}%)" - else: - # Don't show percentages for "0" entries, it's just clutter. - perc = " " - else: + def count_and_perc(self, cc: Cc, i: int) -> tuple[str, str]: + count = f"{cc.counts[i]:,d}" # commify + if args.show_percs: + if cc.counts[i] == 0: + # Don't show percentages for "0" entries, it's just clutter. perc = "" + else: + summary_count = self.summary_cc.counts[i] + if summary_count == 0: + perc = " (n/a)" + else: + p = cc.counts[i] * 100 / summary_count + # Use just one decimal place for large percentages. + if abs(p) < 100: + w = 2 + else: + w = 1 + perc = f" ({p:.{w}f}%)" + else: + perc = "" - self.print_count(i, nstr + perc) + return (count, perc) + + def print_cc(self, cc: Cc, suffix: str) -> None: + for i in self.events.show_indices: + (count, perc) = self.count_and_perc(cc, i) + self.print_count_and_perc(i, count, perc) print("", suffix) def print_missing_cc(self, suffix: str) -> None: - if args.show_percs: - # Don't show percentages for "." entries, it's just clutter. - text = ". " - else: - text = "." - + # Don't show percentages for "." entries, it's just clutter. for i in self.events.show_indices: - self.print_count(i, text) + self.print_count_and_perc(i, ".", "") print("", suffix) diff --git a/cachegrind/tests/ann2.post.exp b/cachegrind/tests/ann2.post.exp index ac12bf87a4..8d33c9e79a 100644 --- a/cachegrind/tests/ann2.post.exp +++ b/cachegrind/tests/ann2.post.exp @@ -20,28 +20,28 @@ Dw Dr Ir -------------------------------------------------------------------------------- Dw Dr Ir file:function -------------------------------------------------------------------------------- - 3 ( 0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) a.c:main -4,543 (25.23%) 17,566 ( 0.43%) 47,993 ( 0.92%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:do_lookup_x -3,083 (17.12%) 5,750 ( 0.14%) 28,534 ( 0.55%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:_dl_lookup_symbol_x - 8 ( 0.04%) 5,521 ( 0.14%) 28,136 ( 0.54%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-tunables.c:__GI___tunables_init -2,490 (13.83%) 5,219 ( 0.13%) 21,821 ( 0.42%) /build/glibc-OTsEL5/glibc-2.27/elf/../sysdeps/x86_64/dl-machine.h:_dl_relocate_object - 0 5,158 ( 0.13%) 25,408 ( 0.49%) /build/glibc-OTsEL5/glibc-2.27/string/../sysdeps/x86_64/strcmp.S:strcmp + 3 (0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) a.c:main +4,543 (25.23%) 17,566 (0.43%) 47,993 (0.92%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:do_lookup_x +3,083 (17.12%) 5,750 (0.14%) 28,534 (0.55%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-lookup.c:_dl_lookup_symbol_x + 8 (0.04%) 5,521 (0.14%) 28,136 (0.54%) /build/glibc-OTsEL5/glibc-2.27/elf/dl-tunables.c:__GI___tunables_init +2,490 (13.83%) 5,219 (0.13%) 21,821 (0.42%) /build/glibc-OTsEL5/glibc-2.27/elf/../sysdeps/x86_64/dl-machine.h:_dl_relocate_object + 0 5,158 (0.13%) 25,408 (0.49%) /build/glibc-OTsEL5/glibc-2.27/string/../sysdeps/x86_64/strcmp.S:strcmp -------------------------------------------------------------------------------- -- User-annotated source: a.c -------------------------------------------------------------------------------- Dw Dr Ir -1 ( 0.01%) 0 2 ( 0.00%) int main(void) { -1 ( 0.01%) 0 1 ( 0.00%) int z = 0; -1 ( 0.01%) 2,000,001 (49.29%) 3,000,004 (57.36%) for (int i = 0; i < 1000000; i++) { -0 2,000,000 (49.29%) 2,000,000 (38.24%) z += i; -. . . } -0 1 ( 0.00%) 6 ( 0.00%) return z % 256; -0 2 ( 0.00%) 2 ( 0.00%) } + 1 (0.01%) 0 2 (0.00%) int main(void) { + 1 (0.01%) 0 1 (0.00%) int z = 0; + 1 (0.01%) 2,000,001 (49.29%) 3,000,004 (57.36%) for (int i = 0; i < 1000000; i++) { + 0 2,000,000 (49.29%) 2,000,000 (38.24%) z += i; + . . . } + 0 1 (0.00%) 6 (0.00%) return z % 256; + 0 2 (0.00%) 2 (0.00%) } -------------------------------------------------------------------------------- Dw Dr Ir -------------------------------------------------------------------------------- -3 ( 0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) events annotated + 3 (0.02%) 4,000,004 (98.57%) 5,000,015 (95.61%) events annotated diff --git a/cachegrind/tests/ann3.post.exp b/cachegrind/tests/ann3.post.exp index a648bb2e78..52c5b87503 100644 --- a/cachegrind/tests/ann3.post.exp +++ b/cachegrind/tests/ann3.post.exp @@ -1,9 +1,9 @@ -------------------------------------------------------------------------------- Command: ann3 Data file: cgout-test3 -Events recorded: A SomeCount ThisIsAVeryLongEventName -Events shown: A SomeCount ThisIsAVeryLongEventName -Event sort order: A SomeCount ThisIsAVeryLongEventName +Events recorded: A SomeCount VeryLongEventName +Events shown: A SomeCount VeryLongEventName +Event sort order: A SomeCount VeryLongEventName Thresholds: 0.5 100 100 Include dirs: ann3-no-such-dir ann3-no-such-dir-2 @@ -13,53 +13,53 @@ User annotated: ann3-unmentioned.rs Auto-annotation: on -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName -------------------------------------------------------------------------------- -100,000 (100.0%) 100,000 (100.0%) 0 PROGRAM TOTALS +100,000 (100.0%) 100,000 (100.0%) 0 PROGRAM TOTALS -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName file:function +A SomeCount VeryLongEventName file:function -------------------------------------------------------------------------------- -70,491 (70.49%) 90,491 (90.49%) 0 ann3-basic.rs:f0 -15,000 (15.00%) 600 ( 0.60%) 0 ann3-basic.rs:f1 - 9,000 ( 9.00%) 6,000 ( 6.00%) 0 ann3-could-not-be-found.rs:f1 - 2,000 ( 2.00%) 100 ( 0.10%) 0 ann3-basic.rs:f2 - 1,000 ( 1.00%) 500 ( 0.50%) 0 ann3-via-I.rs:f1 - 1,000 ( 1.00%) 300 ( 0.30%) -1,000 ( 0.00%) ann3-past-the-end.rs:f1 --1,000 (-1.00%) 0 0 ann3-negatives.rs:neg3 --1,000 (-1.00%) 0 0 ann3-negatives.rs:neg2 - 1,000 ( 1.00%) 0 0 ann3-more-recent-than-cgout.rs:new - 1,000 ( 1.00%) 0 0 ???:unknown - 500 ( 0.50%) 0 0 ann3-basic.rs:f6 - 500 ( 0.50%) 0 0 ann3-basic.rs:f4 +70,491 (70.49%) 90,491 (90.49%) 0 ann3-basic.rs:f0 +15,000 (15.00%) 600 (0.60%) 0 ann3-basic.rs:f1 + 9,000 (9.00%) 6,000 (6.00%) 0 ann3-could-not-be-found.rs:f1 + 2,000 (2.00%) 100 (0.10%) 0 ann3-basic.rs:f2 + 1,000 (1.00%) 500 (0.50%) 0 ann3-via-I.rs:f1 + 1,000 (1.00%) 300 (0.30%) -1,000 (n/a) ann3-past-the-end.rs:f1 +-1,000 (-1.00%) 0 0 ann3-negatives.rs:neg3 +-1,000 (-1.00%) 0 0 ann3-negatives.rs:neg2 + 1,000 (1.00%) 0 0 ann3-more-recent-than-cgout.rs:new + 1,000 (1.00%) 0 0 ???:unknown + 500 (0.50%) 0 0 ann3-basic.rs:f6 + 500 (0.50%) 0 0 ann3-basic.rs:f4 -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-basic.rs -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName -- line 2 ---------------------------------------- - . . . two - . . . three - 5,000 ( 5.00%) 500 ( 0.50%) 0 four - 5,000 ( 5.00%) 100 ( 0.10%) 0 five - . . . six -70,491 (70.49%) 90,491 (90.49%) 0 seven - . . . eight - 110 ( 0.11%) 9 ( 0.01%) 0 nine - . . . ten - . . . eleven - 200 ( 0.20%) 0 0 twelve - 200 ( 0.20%) 0 0 thirteen - 100 ( 0.10%) 0 0 fourteen - 0 0 0 fifteen - 0 0 0 sixteen - 0 0 0 seventeen - 0 0 0 eighteen - 499 ( 0.50%) 2,000 ( 2.00%) 0 nineteen - 300 ( 0.30%) 0 0 twenty - - 7,100 ( 7.10%) 100 ( 0.10%) 0 <counts for unidentified lines in ann3-basic.rs> + . . . two + . . . three + 5,000 (5.00%) 500 (0.50%) 0 four + 5,000 (5.00%) 100 (0.10%) 0 five + . . . six +70,491 (70.49%) 90,491 (90.49%) 0 seven + . . . eight + 110 (0.11%) 9 (0.01%) 0 nine + . . . ten + . . . eleven + 200 (0.20%) 0 0 twelve + 200 (0.20%) 0 0 thirteen + 100 (0.10%) 0 0 fourteen + 0 0 0 fifteen + 0 0 0 sixteen + 0 0 0 seventeen + 0 0 0 eighteen + 499 (0.50%) 2,000 (2.00%) 0 nineteen + 300 (0.30%) 0 0 twenty + + 7,100 (7.10%) 100 (0.10%) 0 <counts for unidentified lines in ann3-basic.rs> -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-more-recent-than-cgout.rs @@ -71,49 +71,49 @@ A SomeCount ThisIsAVeryLongEventName @ Annotations may not be correct. @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName - . . . one -1,000 ( 1.00%) 0 0 two - . . . three - . . . four + . . . one +1,000 (1.00%) 0 0 two + . . . three + . . . four -- line 4 ---------------------------------------- -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-negatives.rs -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName - - 2,000 ( 2.00%) 2,000 ( 2.00%) 2,000 ( 0.00%) one - -1,000 (-1.00%) -1,000 (-1.00%) 0 two - . . . three - . . . four - 999,000 (999.0%) 0 -150,000 ( 0.00%) five --1,000,000 (-1000.0%) 0 150,000 ( 0.00%) six - . . . seven - . . . eight - . . . nine - -10,000 (-10.0%) 0 10 ( 0.00%) ten - 10,000 (10.00%) 0 -20 ( 0.00%) eleven - . . . twelve - . . . thirteen +A SomeCount VeryLongEventName + + 2,000 (2.00%) 2,000 (2.00%) 2,000 (n/a) one + -1,000 (-1.00%) -1,000 (-1.00%) 0 two + . . . three + . . . four + 999,000 (999.0%) 0 -150,000 (n/a) five +-1,000,000 (-1000.0%) 0 150,000 (n/a) six + . . . seven + . . . eight + . . . nine + -10,000 (-10.00%) 0 10 (n/a) ten + 10,000 (10.00%) 0 -20 (n/a) eleven + . . . twelve + . . . thirteen -- line 13 ---------------------------------------- - -2,000 (-2.00%) -1,000 (-1.00%) -990 ( 0.00%) <counts for unidentified lines in ann3-negatives.rs> + -2,000 (-2.00%) -1,000 (-1.00%) -990 (n/a) <counts for unidentified lines in ann3-negatives.rs> -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-past-the-end.rs -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName -200 ( 0.20%) 100 ( 0.10%) 0 one - . . . two - . . . three +200 (0.20%) 100 (0.10%) 0 one + . . . two + . . . three -- line 3 ---------------------------------------- -- line 18 ---------------------------------------- -300 ( 0.30%) 100 ( 0.10%) 0 <bogus line 20> -300 ( 0.30%) 100 ( 0.10%) 0 <bogus line 21> -200 ( 0.20%) 0 -1,000 ( 0.00%) <bogus line 22> +300 (0.30%) 100 (0.10%) 0 <bogus line 20> +300 (0.30%) 100 (0.10%) 0 <bogus line 21> +200 (0.20%) 0 -1,000 (n/a) <bogus line 22> @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ @@ -129,9 +129,9 @@ A SomeCount ThisIsAVeryLongEventName -------------------------------------------------------------------------------- -- Auto-annotated source: ann3-aux/ann3-via-I.rs -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName -1,000 ( 1.00%) 500 ( 0.50%) 0 one +1,000 (1.00%) 500 (0.50%) 0 one -------------------------------------------------------------------------------- The following files chosen for auto-annotation could not be found: @@ -140,7 +140,7 @@ The following files chosen for auto-annotation could not be found: ann3-no-such-file.rs -------------------------------------------------------------------------------- -A SomeCount ThisIsAVeryLongEventName +A SomeCount VeryLongEventName -------------------------------------------------------------------------------- -84,100 (84.10%) 94,700 (94.70%) 1,990 ( 0.00%) events annotated +84,100 (84.10%) 94,700 (94.70%) 1,990 (n/a) events annotated diff --git a/cachegrind/tests/cgout-test3 b/cachegrind/tests/cgout-test3 index 7a3e188a8f..d8023eef1d 100644 --- a/cachegrind/tests/cgout-test3 +++ b/cachegrind/tests/cgout-test3 @@ -1,5 +1,5 @@ cmd: ann3 -events: A SomeCount ThisIsAVeryLongEventName +events: A SomeCount VeryLongEventName # A file testing various things. fl=ann3-basic.rs diff --git a/cachegrind/tests/diff.post.exp b/cachegrind/tests/diff.post.exp index aa88967e26..81fbe3e920 100644 --- a/cachegrind/tests/diff.post.exp +++ b/cachegrind/tests/diff.post.exp @@ -11,25 +11,25 @@ User annotated: Auto-annotation: on -------------------------------------------------------------------------------- -Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw +Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw -------------------------------------------------------------------------------- -5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 PROGRAM TOTALS +5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 PROGRAM TOTALS -------------------------------------------------------------------------------- -Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw file:function +Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw file:function -------------------------------------------------------------------------------- -5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 a.c:main +5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 a.c:main -------------------------------------------------------------------------------- -- Auto-annotated source: a.c -------------------------------------------------------------------------------- -Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw +Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw -5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 <counts for unidentified lines in a.c> +5,000,000 (100.0%) 0 0 -2,000,000 (100.0%) 0 0 0 0 0 <counts for unidentified lines in a.c> -------------------------------------------------------------------------------- -Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw +Ir I1mr ILmr Dr D1mr DLmr Dw D1mw DLmw -------------------------------------------------------------------------------- -0 0 0 0 0 0 0 0 0 events annotated + 0 0 0 0 0 0 0 0 0 events annotated |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-22 23:09:11
|
All the Rust projects I work on use GitHub, and in terms of usability and productivity it's miles ahead of how Valgrind development works. If I were king of the world here's how I would drag Valgrind's development practices forward by 10-20 years. - Move the repository to GitHub. Require all changes to be done via pull requests, with no direct pushing. - Set up some CI testing via GitHub Actions. Require that all pull requests pass these tests before merging. - Lots of projects require a review approval before a pull request can be merged. But that might be too hard for Valgrind to start with, given the small number of active contributors. - Switch from KDE bugzilla to GitHub issues for bug reporting. Not sure what I'd do with existing open bug reports, whether it would be worth importing them to GitHub issues somehow or not. - Use auto-formatting tools, such as clang-format. (Possibly even moving from 3 space indents in C code to 2 or 4!) - Change the docs from that XML-based thing we use (groan) to something nicer, probably involving Markdown. - Website: not sure... a lot of it could be naturally hosted on the main GitHub page. It might be nice to still have valgrind.org, though, but perhaps greatly stripped back. I understand the concerns about GitHub and commercialization, but I also worry about Valgrind's future viability if it doesn't attract some level of new contributors. The Linux kernel will never have that problem, but I suspect lots of GNU projects also face that risk. Nick On Thu, 23 Mar 2023 at 08:19, Paul Floyd <pj...@wa...> wrote: > > > On 22-03-23 21:31, Nicholas Nethercote wrote: > > Thanks to Paul and Mark for a couple of small fixes to my commit. > > > > Yesterday I was idly dreaming about the quality-of-life improvements > > that would be available if Valgrind was hosted on GitHub: > > - ability to upload commits ahead of time, in a fashion nicer than > > "attach patch to bugzilla" > > - ability to do reviews > > - CI support for pre-merge testing runs > > - easier entry for newcomers > > > > Does sourceware.org <http://sourceware.org> have support for any of > > these things? > > Hi > > GH does have a lot going for it, and when they add code browsing it will > be even better. One concern though is their owner and creeping > commercialization. > > As far as I know, a lot of GNU projects and Linux still work mostly by > by patches sent to mailing lists. > > One other fairly common system is phabricator (used by LLVM and FreeBSD > amonst others). However, the company that developed phabricator closed > down so I don't know where that is going (there is a fork). FWIW FreeBSD > is increasingly using GH. > > A+ > Paul > > > > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers > |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-22 22:52:55
|
Thanks for the suggestion. I have removed the use of `TypeAlias` like you suggested, and documented version expectations. Nick On Thu, 23 Mar 2023 at 08:03, Paul Floyd <pj...@wa...> wrote: > > > On 22-03-23 00:22, Nicholas Nethercote wrote: > > I have merged the new version of `cg_annotate`: > > > https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444 > < > https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444 > > > > Hi Nick > > I've seen two problems. The first I've already fixed with "env". > > The other is with typing / TypeAlias. This was added in Python 3.10. On > my FreeBSD system "python3" defaults to Python 3.9, so I get > > ImportError: cannot import name 'TypeAlias' from 'typing' > (/usr/local/lib/python3.9/typing.py) > > That will probably also cause problems on old Linux systems as well. > > I don't know much Python, but would it be OK to use non-explicit > (implicit?) type aliases like in the diff below: > > The alternative would be to add python3.10 detection, something like > > AC_CHECK_PROGS([PYTHON3],[python3.10 python3] > > and then to use @PYTHON3@ in cg_annotate.in and the 3 ann[123]/vgtest > files. > > A+ > Paul > > > > diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in > index 91d75aecd..c3d5f71d4 100755 > --- a/cachegrind/cg_annotate.in > +++ b/cachegrind/cg_annotate.in > @@ -73,7 +73,7 @@ import re > import sys > from argparse import ArgumentParser, BooleanOptionalAction, Namespace > from collections import defaultdict > -from typing import Callable, DefaultDict, NewType, NoReturn, TextIO, > TypeAlias > +from typing import Callable, DefaultDict, NewType, NoReturn, TextIO > > > class Args(Namespace): > @@ -323,11 +323,11 @@ class Cc: > Flfn = NewType("Flfn", tuple[str, str]) > > # Per-function CCs. > -DictFlfnCc: TypeAlias = DefaultDict[Flfn, Cc] > +DictFlfnCc = DefaultDict[Flfn, Cc] > > # Per-line CCs, organised by filename and line number. > -DictLineCc: TypeAlias = DefaultDict[int, Cc] > -DictFlDictLineCc: TypeAlias = DefaultDict[str, DictLineCc] > +DictLineCc = DefaultDict[int, Cc] > +DictFlDictLineCc = DefaultDict[str, DictLineCc] > > > def die(msg: str) -> NoReturn: > > > > A+ > Paul > > > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers > |
|
From: Nicholas N. <nj...@so...> - 2023-03-22 22:51:47
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=e9e7b663fb69075a110963d85726389f469fcf5b commit e9e7b663fb69075a110963d85726389f469fcf5b Author: Nicholas Nethercote <n.n...@gm...> Date: Thu Mar 23 09:50:21 2023 +1100 Make `cg_annotate` work with Python 3.9, by avoiding `TypeAlias`. Diff: --- cachegrind/cg_annotate.in | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index 91d75aecdf..20969f0f92 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -34,6 +34,10 @@ This script reads Cachegrind output files and produces human-readable reports. # formatters, type-checkers, and linters on `cg_annotate.in` and then generates # `cg_annotate`. # +# Python versions: Currently this script targets Python 3.9 and later versions. +# Consequences of this: +# - No use of `TypeAlias` for explicit type aliases, which requires 3.10. +# # The following Python tools are used. All can be installed with `pip3 install # $NAME`, except `cProfile` which is built into Python. # @@ -73,7 +77,7 @@ import re import sys from argparse import ArgumentParser, BooleanOptionalAction, Namespace from collections import defaultdict -from typing import Callable, DefaultDict, NewType, NoReturn, TextIO, TypeAlias +from typing import Callable, DefaultDict, NewType, NoReturn, TextIO class Args(Namespace): @@ -323,11 +327,13 @@ class Cc: Flfn = NewType("Flfn", tuple[str, str]) # Per-function CCs. -DictFlfnCc: TypeAlias = DefaultDict[Flfn, Cc] +# Note: not using `TypeAlias`. See "Python versions" comment above. +DictFlfnCc = DefaultDict[Flfn, Cc] # Per-line CCs, organised by filename and line number. -DictLineCc: TypeAlias = DefaultDict[int, Cc] -DictFlDictLineCc: TypeAlias = DefaultDict[str, DictLineCc] +# Note: not using `TypeAlias`. See "Python versions" comment above. +DictLineCc = DefaultDict[int, Cc] +DictFlDictLineCc = DefaultDict[str, DictLineCc] def die(msg: str) -> NoReturn: |
|
From: Paul F. <pj...@wa...> - 2023-03-22 21:18:25
|
On 22-03-23 21:31, Nicholas Nethercote wrote: > Thanks to Paul and Mark for a couple of small fixes to my commit. > > Yesterday I was idly dreaming about the quality-of-life improvements > that would be available if Valgrind was hosted on GitHub: > - ability to upload commits ahead of time, in a fashion nicer than > "attach patch to bugzilla" > - ability to do reviews > - CI support for pre-merge testing runs > - easier entry for newcomers > > Does sourceware.org <http://sourceware.org> have support for any of > these things? Hi GH does have a lot going for it, and when they add code browsing it will be even better. One concern though is their owner and creeping commercialization. As far as I know, a lot of GNU projects and Linux still work mostly by by patches sent to mailing lists. One other fairly common system is phabricator (used by LLVM and FreeBSD amonst others). However, the company that developed phabricator closed down so I don't know where that is going (there is a fork). FWIW FreeBSD is increasingly using GH. A+ Paul |
|
From: Paul F. <pj...@wa...> - 2023-03-22 21:02:53
|
On 22-03-23 00:22, Nicholas Nethercote wrote: > I have merged the new version of `cg_annotate`: > https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444 <https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444> Hi Nick I've seen two problems. The first I've already fixed with "env". The other is with typing / TypeAlias. This was added in Python 3.10. On my FreeBSD system "python3" defaults to Python 3.9, so I get ImportError: cannot import name 'TypeAlias' from 'typing' (/usr/local/lib/python3.9/typing.py) That will probably also cause problems on old Linux systems as well. I don't know much Python, but would it be OK to use non-explicit (implicit?) type aliases like in the diff below: The alternative would be to add python3.10 detection, something like AC_CHECK_PROGS([PYTHON3],[python3.10 python3] and then to use @PYTHON3@ in cg_annotate.in and the 3 ann[123]/vgtest files. A+ Paul diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index 91d75aecd..c3d5f71d4 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -73,7 +73,7 @@ import re import sys from argparse import ArgumentParser, BooleanOptionalAction, Namespace from collections import defaultdict -from typing import Callable, DefaultDict, NewType, NoReturn, TextIO, TypeAlias +from typing import Callable, DefaultDict, NewType, NoReturn, TextIO class Args(Namespace): @@ -323,11 +323,11 @@ class Cc: Flfn = NewType("Flfn", tuple[str, str]) # Per-function CCs. -DictFlfnCc: TypeAlias = DefaultDict[Flfn, Cc] +DictFlfnCc = DefaultDict[Flfn, Cc] # Per-line CCs, organised by filename and line number. -DictLineCc: TypeAlias = DefaultDict[int, Cc] -DictFlDictLineCc: TypeAlias = DefaultDict[str, DictLineCc] +DictLineCc = DefaultDict[int, Cc] +DictFlDictLineCc = DefaultDict[str, DictLineCc] def die(msg: str) -> NoReturn: A+ Paul |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-22 20:32:11
|
Thanks to Paul and Mark for a couple of small fixes to my commit. Yesterday I was idly dreaming about the quality-of-life improvements that would be available if Valgrind was hosted on GitHub: - ability to upload commits ahead of time, in a fashion nicer than "attach patch to bugzilla" - ability to do reviews - CI support for pre-merge testing runs - easier entry for newcomers Does sourceware.org have support for any of these things? Nick On Wed, 22 Mar 2023 at 10:22, Nicholas Nethercote <n.n...@gm...> wrote: > I have merged the new version of `cg_annotate`: > > https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444 > > Nick > > On Fri, 17 Mar 2023 at 16:24, Nicholas Nethercote <n.n...@gm...> > wrote: > >> I have finished the rewrite. I am happy with the new code, it is much >> better than the old code. You can see it at >> https://bugs.kde.org/show_bug.cgi?id=467472. I plan to merge it by the >> end of next week, and I am happy to hear any suggestions. >> >> I also have some good news about the `cg_annotate.in`/`cg_annotate` >> split. I learned that you can generate the latter from the former very >> quickly with `config.status cachegrind/cg_annotate.in`. Also, this can >> be done automatically from some make targets. So I ended up creating a new >> make target `make ann` that can be run within the `cachegrind` directory. >> It runs the various Python formatters, type-checkers, and linters I am >> using on `cg_annotate.in` and then generates `cg_annotate`. It's a >> one-step "build" command that runs quickly, which is great. >> >> Nick >> >> On Wed, 15 Mar 2023 at 06:15, Nicholas Nethercote <n.n...@gm...> >> wrote: >> >>> On Sun, 12 Mar 2023 at 23:01, Paul Floyd <pj...@wa...> wrote: >>> >>>> >>>> The only think I can think of to get the version is to use something >>>> like >>>> >>>> pkg-config --modversion valgrind >>>> >>> >>> Thanks for the suggestion. Unfortunately this could cause misleading >>> results. E.g. if I have Valgrind installed on my system but I also have a >>> development version, when I run the development version of `cg_annotate >>> --version` it will claim to be the installed version. I think the `@VERSION@` >>> junk is unavoidable. >>> >>> Nick >>> >> |
|
From: Paul F. <pa...@so...> - 2023-03-22 19:44:49
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=bb35cd572f7f805eb55160ad13e5bf8f1c174499 commit bb35cd572f7f805eb55160ad13e5bf8f1c174499 Author: Paul Floyd <pj...@wa...> Date: Wed Mar 22 20:43:30 2023 +0100 Make cg_annotate independent of python3 install location Diff: --- cachegrind/cg_annotate.in | 2 +- cachegrind/tests/ann1.vgtest | 2 +- cachegrind/tests/ann2.vgtest | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in index 247026f1fa..91d75aecdf 100755 --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -1,4 +1,4 @@ -#! /usr/bin/python3 +#! /usr/bin/env python3 # pyright: strict # -------------------------------------------------------------------- diff --git a/cachegrind/tests/ann1.vgtest b/cachegrind/tests/ann1.vgtest index e3e574276a..660f524f09 100644 --- a/cachegrind/tests/ann1.vgtest +++ b/cachegrind/tests/ann1.vgtest @@ -2,5 +2,5 @@ # the post-processing of the cgout-test file. prog: ../../tests/true vgopts: --cachegrind-out-file=cachegrind.out -post: touch cgout-test && perl ../../cachegrind/cg_annotate --show=Ir,I1mr,ILmr --show-percs=no cgout-test +post: touch cgout-test && python3 ../../cachegrind/cg_annotate --show=Ir,I1mr,ILmr --show-percs=no cgout-test cleanup: rm cachegrind.out diff --git a/cachegrind/tests/ann2.vgtest b/cachegrind/tests/ann2.vgtest index 7cf1b7fcd3..9ef76e38e9 100644 --- a/cachegrind/tests/ann2.vgtest +++ b/cachegrind/tests/ann2.vgtest @@ -2,5 +2,5 @@ # the post-processing of the cgout-test file. prog: ../../tests/true vgopts: --cachegrind-out-file=cachegrind.out -post: touch cgout-test && perl ../../cachegrind/cg_annotate --sort=Dr --show=Dw,Dr,Ir --auto=no cgout-test a.c +post: touch cgout-test && python3 ../../cachegrind/cg_annotate --sort=Dr --show=Dw,Dr,Ir --auto=no cgout-test a.c cleanup: rm cachegrind.out |
|
From: Mark W. <ma...@so...> - 2023-03-22 10:42:36
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=3bb907290aa97ba6b5d776844444c04fcb3a71aa commit 3bb907290aa97ba6b5d776844444c04fcb3a71aa Author: Mark Wielaard <ma...@kl...> Date: Wed Mar 22 11:41:53 2023 +0100 cachegrind/tests/Makefile.am EXTRA_DIST add ann3.{{post,stderr}.exp,vgtest} Diff: --- cachegrind/tests/Makefile.am | 1 + 1 file changed, 1 insertion(+) diff --git a/cachegrind/tests/Makefile.am b/cachegrind/tests/Makefile.am index 2ce7426677..ded05c8fdb 100644 --- a/cachegrind/tests/Makefile.am +++ b/cachegrind/tests/Makefile.am @@ -16,6 +16,7 @@ EXTRA_DIST = \ cgout-test \ ann1.post.exp ann1.stderr.exp ann1.vgtest \ ann2.post.exp ann2.stderr.exp ann2.vgtest \ + ann3.post.exp ann3.stderr.exp ann3.vgtest \ chdir.vgtest chdir.stderr.exp \ clreq.vgtest clreq.stderr.exp \ diff.post.exp diff.stderr.exp diff.vgtest \ |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-21 23:23:08
|
I have merged the new version of `cg_annotate`: https://sourceware.org/git/?p=valgrind.git;a=commit;h=4650b7949ae3a41326e52ae454a9202493c41444 Nick On Fri, 17 Mar 2023 at 16:24, Nicholas Nethercote <n.n...@gm...> wrote: > I have finished the rewrite. I am happy with the new code, it is much > better than the old code. You can see it at > https://bugs.kde.org/show_bug.cgi?id=467472. I plan to merge it by the > end of next week, and I am happy to hear any suggestions. > > I also have some good news about the `cg_annotate.in`/`cg_annotate` > split. I learned that you can generate the latter from the former very > quickly with `config.status cachegrind/cg_annotate.in`. Also, this can be > done automatically from some make targets. So I ended up creating a new > make target `make ann` that can be run within the `cachegrind` directory. > It runs the various Python formatters, type-checkers, and linters I am > using on `cg_annotate.in` and then generates `cg_annotate`. It's a > one-step "build" command that runs quickly, which is great. > > Nick > > On Wed, 15 Mar 2023 at 06:15, Nicholas Nethercote <n.n...@gm...> > wrote: > >> On Sun, 12 Mar 2023 at 23:01, Paul Floyd <pj...@wa...> wrote: >> >>> >>> The only think I can think of to get the version is to use something like >>> >>> pkg-config --modversion valgrind >>> >> >> Thanks for the suggestion. Unfortunately this could cause misleading >> results. E.g. if I have Valgrind installed on my system but I also have a >> development version, when I run the development version of `cg_annotate >> --version` it will claim to be the installed version. I think the `@VERSION@` >> junk is unavoidable. >> >> Nick >> > |
|
From: Nicholas N. <nj...@so...> - 2023-03-21 23:15:05
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=4650b7949ae3a41326e52ae454a9202493c41444 commit 4650b7949ae3a41326e52ae454a9202493c41444 Author: Nicholas Nethercote <n.n...@gm...> Date: Thu Mar 9 15:41:31 2023 +1100 Rewrite `cg_annotate` in Python. Perl was a reasonable choice for `cg_annotate` in 2002, but not in 2023. Also, the existing structure of the code is not good. These two things make it hard to modify `cg_annotate` in any significant way. Benefits of the change: - Now written in a language that is (a) nice, and (b) not moribund. - Easier to maintain, due to (a) abovementioned better language, (b) better code structure, and (c) better language tooling, such as formatters, type checkers, and linters. - The new version is a little shorter. - It runs about 2x faster. - Argument handling is more standard. E.g. things like `--context 2`, `--auto`, `--no-auto` are supported. (The old forms that require `=` are still supported, though the `=yes`/`=no` forms are deprecated.) The behaviour and output of the new version is identical for typical uses, but there are some very minor changes for edge cases, which nobody is likely to notice. For example: - The file format is slightly changed: I removed support for '.' counts, which had the same meaning as '0'. This was a feature that Cachegrind never used, and the old script handled it inconsistently. - The new version will abort on a malformed data line. The old version would just print a warning and continue. The commit also adds a new test `ann3` that tests many parts of `cg_annotate` that weren't tested previously, and tweaks the existing `ann2` test. Diff: --- NEWS | 5 + cachegrind/Makefile.am | 30 + cachegrind/cg_annotate.in | 1759 +++++++++++------------ cachegrind/docs/cg-manual.xml | 33 +- cachegrind/pylintrc | 53 + cachegrind/tests/ann2.post.exp | 14 +- cachegrind/tests/ann2.vgtest | 2 +- cachegrind/tests/ann3-aux/ann3-via-I.rs | 1 + cachegrind/tests/ann3-basic.rs | 20 + cachegrind/tests/ann3-more-recent-than-cgout.rs | 5 + cachegrind/tests/ann3-negatives.rs | 15 + cachegrind/tests/ann3-past-the-end.rs | 3 + cachegrind/tests/ann3-unmentioned.rs | 1 + cachegrind/tests/ann3.post.exp | 146 ++ cachegrind/tests/ann3.stderr.exp | 17 + cachegrind/tests/ann3.vgtest | 13 + cachegrind/tests/cgout-test3 | 93 ++ cachegrind/tests/diff.post.exp | 4 +- 18 files changed, 1258 insertions(+), 956 deletions(-) diff --git a/NEWS b/NEWS index 8ede199d67..990043fe63 100644 --- a/NEWS +++ b/NEWS @@ -86,6 +86,11 @@ AMD64/macOS 10.13 and nanoMIPS/Linux. - Valgrind now contains python code that defines GDB helgrind front end monitor commands. See CORE CHANGES. +* Cachegrind: + - `cg_annotate` has been rewritten from Perl into Python. The new + version is twice as fast, has more flexible argument parsing, and + will make future improvements easier. + * Callgrind: - Valgrind now contains python code that defines GDB callgrind front end monitor commands. See CORE CHANGES. diff --git a/cachegrind/Makefile.am b/cachegrind/Makefile.am index f8447a17ce..8ea99ca529 100644 --- a/cachegrind/Makefile.am +++ b/cachegrind/Makefile.am @@ -88,3 +88,33 @@ cachegrind_@VGCONF_ARCH_SEC@_@VGCONF_OS@_LINK = \ $(cachegrind_@VGCONF_ARCH_SEC@_@VGCONF_OS@_CFLAGS) \ $(cachegrind_@VGCONF_ARCH_SEC@_@VGCONF_OS@_LDFLAGS) endif + +#---------------------------------------------------------------------------- +# Miscellaneous +#---------------------------------------------------------------------------- + +# Run the formatters, type checkers, and linters on `cg_annotate.in`, then +# generate `cg_annotate`. +# +# Note: `pyright` refuses to check any file without a `.py` extension, hence +# the copying to `/tmp/tmp.py`. +ann: + @echo "== black ==" + @black cg_annotate.in + @echo + @echo "== isort ==" + @isort cg_annotate.in + @echo + @echo "== mypy ==" + @mypy --strict cg_annotate.in + @echo + @echo "== pyright ==" + @cp cg_annotate.in /tmp/tmp.py && pyright /tmp/tmp.py && rm /tmp/tmp.py + @echo + @echo "== ruff ==" + @ruff cg_annotate.in + @echo + @echo "== pylint ==" + @pylint cg_annotate.in + @echo "== config.status ==" + $(MAKE) cg_annotate diff --git a/cachegrind/cg_annotate.in b/cachegrind/cg_annotate.in old mode 100644 new mode 100755 index 9111fbe7ef..247026f1fa --- a/cachegrind/cg_annotate.in +++ b/cachegrind/cg_annotate.in @@ -1,942 +1,855 @@ -#! @PERL@ +#! /usr/bin/python3 +# pyright: strict -##--------------------------------------------------------------------## -##--- Cachegrind's annotator. cg_annotate.in ---## -##--------------------------------------------------------------------## +# -------------------------------------------------------------------- +# --- Cachegrind's annotator. cg_annotate.in --- +# -------------------------------------------------------------------- -# This file is part of Cachegrind, a Valgrind tool for cache -# profiling programs. +# This file is part of Cachegrind, a Valgrind tool for cache +# profiling programs. # -# Copyright (C) 2002-2017 Nicholas Nethercote -# nj...@va... +# Copyright (C) 2002-2023 Nicholas Nethercote +# nj...@va... # -# This program is free software; you can redistribute it and/or -# modify it under the terms of the GNU General Public License as -# published by the Free Software Foundation; either version 2 of the -# License, or (at your option) any later version. +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License as +# published by the Free Software Foundation; either version 2 of the +# License, or (at your option) any later version. # -# This program is distributed in the hope that it will be useful, but -# WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# General Public License for more details. +# This program is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. # -# You should have received a copy of the GNU General Public License -# along with this program; if not, see <http://www.gnu.org/licenses/>. +# You should have received a copy of the GNU General Public License +# along with this program; if not, see <http://www.gnu.org/licenses/>. # -# The GNU General Public License is contained in the file COPYING. - -#---------------------------------------------------------------------------- -# The file format is simple, basically printing the cost centre for every -# source line, grouped by files and functions. The details are in -# Cachegrind's manual. - -#---------------------------------------------------------------------------- -# Performance improvements record, using cachegrind.out for cacheprof, doing no -# source annotation (irrelevant ones removed): -# user time -# 1. turned off warnings in add_hash_a_to_b() 3.81 --> 3.48s -# [now add_array_a_to_b()] -# 6. make line_to_CC() return a ref instead of a hash 3.01 --> 2.77s +# The GNU General Public License is contained in the file COPYING. + +""" +This script reads Cachegrind output files and produces human-readable reports. +""" + +# Use `make ann` to "build" this script every time it is changed. This runs the +# formatters, type-checkers, and linters on `cg_annotate.in` and then generates +# `cg_annotate`. # -#10. changed file format to avoid file/fn name repetition 2.40s -# (not sure why higher; maybe due to new '.' entries?) -#11. changed file format to drop unnecessary end-line "."s 2.36s -# (shrunk file by about 37%) -#12. switched from hash CCs to array CCs 1.61s -#13. only adding b[i] to a[i] if b[i] defined (was doing it if -# either a[i] or b[i] was defined, but if b[i] was undefined -# it just added 0) 1.48s -#14. Stopped converting "." entries to undef and then back 1.16s -#15. Using foreach $i (x..y) instead of for ($i = 0...) in -# add_array_a_to_b() 1.11s +# The following Python tools are used. All can be installed with `pip3 install +# $NAME`, except `cProfile` which is built into Python. # -# Auto-annotating primes: -#16. Finding count lengths by int((length-1)/3), not by -# commifying (halves the number of commify calls) 1.68s --> 1.47s - -use warnings; -use strict; - -#---------------------------------------------------------------------------- -# Overview: the running example in the comments is for: -# - events = A,B,C,D -# - --show=C,A,D -# - --sort=D,C -#---------------------------------------------------------------------------- - -#---------------------------------------------------------------------------- -# Global variables, main data structures -#---------------------------------------------------------------------------- -# CCs are arrays, the counts corresponding to @events, with 'undef' -# representing '.'. This makes things fast (faster than using hashes for CCs) -# but we have to use @sort_order and @show_order below to handle the --sort and -# --show options, which is a bit tricky. -#---------------------------------------------------------------------------- - -# Total counts for summary (an array reference). -my $summary_CC; - -# Totals for each function, for overall summary. -# hash(filename:fn_name => CC array) -my %fn_totals; - -# Individual CCs, organised by filename and line_num for easy annotation. -# hash(filename => hash(line_num => CC array)) -my %allCCs; - -# Files chosen for annotation on the command line. -# key = basename (trimmed of any directory), value = full filename -my %user_ann_files; - -# Generic description string. -my $desc = ""; - -# Command line of profiled program. -my $cmd; - -# Events in input file, eg. (A,B,C,D) -my @events; - -# Events to show, from command line, eg. (C,A,D) -my @show_events; - -# Map from @show_events indices to @events indices, eg. (2,0,3). Gives the -# order in which we must traverse @events in order to show the @show_events, -# eg. (@events[$show_order[1]], @events[$show_order[2]]...) = @show_events. -# (Might help to think of it like a hash (0 => 2, 1 => 0, 2 => 3).) -my @show_order; - -# Print out the function totals sorted by these events, eg. (D,C). -my @sort_events; - -# Map from @sort_events indices to @events indices, eg. (3,2). Same idea as -# for @show_order. -my @sort_order; - -# Thresholds, one for each sort event (or default to 1 if no sort events -# specified). We print out functions and do auto-annotations until we've -# handled this proportion of all the events thresholded. -my @thresholds; - -my $default_threshold = 0.1; - -my $single_threshold = $default_threshold; - -# If on, show a percentage for each non-zero count. -my $show_percs = 1; - -# If on, automatically annotates all files that are involved in getting over -# all the threshold counts. -my $auto_annotate = 1; - -# Number of lines to show around each annotated line. -my $context = 8; - -# Directories in which to look for annotation files. -my @include_dirs = (""); - -# Input file name -my $input_file = undef; - -# Version number -my $version = "@VERSION@"; - -# Usage message. -my $usage = <<END -usage: cg_annotate [options] cachegrind-out-file [source-files...] - - options for the user, with defaults in [ ], are: - -h --help show this message - --version show version - --show=A,B,C only show figures for events A,B,C [all] - --sort=A,B,C sort columns by events A,B,C [event column order] - --threshold=<0--20> a function is shown if it accounts for more than x% of - the counts of the primary sort event [$default_threshold] - --show-percs=yes|no show a percentage for each non-zero count [yes] - --auto=yes|no annotate all source files containing functions - that helped reach the event count threshold [yes] - --context=N print N lines of context before and after - annotated lines [8] - -I<d> --include=<d> add <d> to list of directories to search for - source files - - cg_annotate is Copyright (C) 2002-2017 Nicholas Nethercote. - and licensed under the GNU General Public License, version 2. - Bug reports, feedback, admiration, abuse, etc, to: njn\@valgrind.org. - -END -; - -# Used in various places of output. -my $fancy = '-' x 80 . "\n"; - -sub safe_div($$) -{ - my ($x, $y) = @_; - return ($y == 0 ? 0 : $x / $y); -} - -#----------------------------------------------------------------------------- -# Argument and option handling -#----------------------------------------------------------------------------- -sub process_cmd_line() -{ - for my $arg (@ARGV) { - - # Option handling - if ($arg =~ /^-/) { - - # --version - if ($arg =~ /^--version$/) { - die("cg_annotate-$version\n"); - - # --show=A,B,C - } elsif ($arg =~ /^--show=(.*)$/) { - @show_events = split(/,/, $1); - - # --sort=A,B,C - # Nb: You can specify thresholds individually, eg. - # --sort=A:99,B:95,C:90. These will override any --threshold - # argument. - } elsif ($arg =~ /^--sort=(.*)$/) { - @sort_events = split(/,/, $1); - my $th_specified = 0; - foreach my $i (0 .. scalar @sort_events - 1) { - if ($sort_events[$i] =~ /.*:([\d\.]+)%?$/) { - my $th = $1; - ($th >= 0 && $th <= 100) or die($usage); - $sort_events[$i] =~ s/:.*//; - $thresholds[$i] = $th; - $th_specified = 1; - } else { - $thresholds[$i] = 0; - } - } - if (not $th_specified) { - @thresholds = (); - } - - # --threshold=X (tolerates a trailing '%') - } elsif ($arg =~ /^--threshold=([\d\.]+)%?$/) { - $single_threshold = $1; - ($1 >= 0 && $1 <= 20) or die($usage); - - # --show-percs=yes|no - } elsif ($arg =~ /^--show-percs=yes$/) { - $show_percs = 1; - } elsif ($arg =~ /^--show-percs=no$/) { - $show_percs = 0; - - # --auto=yes|no - } elsif ($arg =~ /^--auto=yes$/) { - $auto_annotate = 1; - } elsif ($arg =~ /^--auto=no$/) { - $auto_annotate = 0; - - # --context=N - } elsif ($arg =~ /^--context=([\d\.]+)$/) { - $context = $1; - if ($context < 0) { - die($usage); - } - - # We don't handle "-I name" -- there can be no space. - } elsif ($arg =~ /^-I$/) { - die("Sorry, no space is allowed after a -I flag\n"); - - # --include=A,B,C. Allow -I=name for backwards compatibility. - } elsif ($arg =~ /^(-I=|-I|--include=)(.*)$/) { - my $inc = $2; - $inc =~ s|/$||; # trim trailing '/' - push(@include_dirs, "$inc/"); - - } else { # -h and --help fall under this case - die($usage); - } - - # Argument handling -- annotation file checking and selection. - # Stick filenames into a hash for quick 'n easy lookup throughout. - } else { - if (not defined $input_file) { - # First non-option argument is the output file. - $input_file = $arg; - } else { - # Subsequent non-option arguments are source files. - my $readable = 0; - foreach my $include_dir (@include_dirs) { - if (-r $include_dir . $arg) { - $readable = 1; - } - } - $readable or die("File $arg not found in any of: @include_dirs\n"); - $user_ann_files{$arg} = 1; - } - } - } - - # Must have chosen an input file - if (not defined $input_file) { - die($usage); - } -} - -#----------------------------------------------------------------------------- -# Reading of input file -#----------------------------------------------------------------------------- -sub max ($$) -{ - my ($x, $y) = @_; - return ($x > $y ? $x : $y); -} - -# Add the two arrays; any '.' entries are ignored. Two tricky things: -# 1. If $a2->[$i] is undefined, it defaults to 0 which is what we want; we turn -# off warnings to allow this. This makes things about 10% faster than -# checking for definedness ourselves. -# 2. We don't add an undefined count or a ".", even though it's value is 0, -# because we don't want to make an $a2->[$i] that is undef become 0 -# unnecessarily. -sub add_array_a_to_b ($$) -{ - my ($a1, $a2) = @_; - - my $n = max(scalar @$a1, scalar @$a2); - $^W = 0; - foreach my $i (0 .. $n-1) { - $a2->[$i] += $a1->[$i] if (defined $a1->[$i] && "." ne $a1->[$i]); - } - $^W = 1; -} - -# Add each event count to the CC array. '.' counts become undef, as do -# missing entries (implicitly). -sub line_to_CC ($) -{ - my @CC = (split /\s+/, $_[0]); - (@CC <= @events) or die("Line $.: too many event counts\n"); - return \@CC; -} - -sub read_input_file() -{ - open(INPUTFILE, "< $input_file") - || die "Cannot open $input_file for reading\n"; - - # Read "desc:" lines. - my $line; - while ($line = <INPUTFILE>) { - if ($line =~ s/desc:\s+//) { - $desc .= $line; - } else { - last; - } - } - - # Read "cmd:" line (Nb: will already be in $line from "desc:" loop above). - ($line =~ s/^cmd:\s+//) or die("Line $.: missing command line\n"); - $cmd = $line; - chomp($cmd); # Remove newline - - # Read "events:" line. We make a temporary hash in which the Nth event's - # value is N, which is useful for handling --show/--sort options below. - $line = <INPUTFILE>; - (defined $line && $line =~ s/^events:\s+//) - or die("Line $.: missing events line\n"); - @events = split(/\s+/, $line); - my %events; - my $n = 0; - foreach my $event (@events) { - $events{$event} = $n; - $n++ - } - - # If no --show arg give, default to showing all events in the file. - # If --show option is used, check all specified events appeared in the - # "events:" line. Then initialise @show_order. - if (@show_events) { - foreach my $show_event (@show_events) { - (defined $events{$show_event}) or - die("--show event `$show_event' did not appear in input\n"); - } - } else { - @show_events = @events; - } - foreach my $show_event (@show_events) { - push(@show_order, $events{$show_event}); - } - - # Do as for --show, but if no --sort arg given, default to sorting by - # column order (ie. first column event is primary sort key, 2nd column is - # 2ndary key, etc). - if (@sort_events) { - foreach my $sort_event (@sort_events) { - (defined $events{$sort_event}) or - die("--sort event `$sort_event' did not appear in input\n"); - } - } else { - @sort_events = @events; - } - foreach my $sort_event (@sort_events) { - push(@sort_order, $events{$sort_event}); - } - - # If multiple threshold args weren't given via --sort, stick in the single - # threshold (either from --threshold if used, or the default otherwise) for - # the primary sort event, and 0% for the rest. - if (not @thresholds) { - foreach my $e (@sort_order) { - push(@thresholds, 100); - } - $thresholds[0] = $single_threshold; - } - - my $currFileName; - my $currFileFuncName; - - my $currFuncCC; - my $currFileCCs = {}; # hash(line_num => CC) - - # Read body of input file. - while (<INPUTFILE>) { - # Skip comments and empty lines. - next if /^\s*$/ || /^\#/; - - if (s/^(-?\d+)\s+//) { - my $lineNum = $1; - my $CC = line_to_CC($_); - defined($currFuncCC) || die; - add_array_a_to_b($CC, $currFuncCC); - - # If currFileName is selected, add CC to currFileName list. We look for - # full filename matches; or, if auto-annotating, we have to - # remember everything -- we won't know until the end what's needed. - defined($currFileCCs) || die; - if ($auto_annotate || defined $user_ann_files{$currFileName}) { - my $currLineCC = $currFileCCs->{$lineNum}; - if (not defined $currLineCC) { - $currLineCC = []; - $currFileCCs->{$lineNum} = $currLineCC; - } - add_array_a_to_b($CC, $currLineCC); - } - - } elsif (s/^fn=(.*)$//) { - $currFileFuncName = "$currFileName:$1"; - $currFuncCC = $fn_totals{$currFileFuncName}; - if (not defined $currFuncCC) { - $currFuncCC = []; - $fn_totals{$currFileFuncName} = $currFuncCC; - } - - } elsif (s/^fl=(.*)$//) { - $currFileName = $1; - $currFileCCs = $allCCs{$currFileName}; - if (not defined $currFileCCs) { - $currFileCCs = {}; - $allCCs{$currFileName} = $currFileCCs; - } - # Assume that a "fn=" line is followed by a "fl=" line. - $currFileFuncName = undef; - - } elsif (s/^summary:\s+//) { - $summary_CC = line_to_CC($_); - (scalar(@$summary_CC) == @events) - or die("Line $.: summary event and total event mismatch\n"); - - } else { - warn("WARNING: line $. malformed, ignoring\n"); - } - } - - # Check if summary line was present - if (not defined $summary_CC) { - die("missing final summary line, aborting\n"); - } - - close(INPUTFILE); -} - -#----------------------------------------------------------------------------- -# Print options used -#----------------------------------------------------------------------------- -sub print_options () -{ - print($fancy); - print($desc); - print("Command: $cmd\n"); - print("Data file: $input_file\n"); - print("Events recorded: @events\n"); - print("Events shown: @show_events\n"); - print("Event sort order: @sort_events\n"); - print("Thresholds: @thresholds\n"); - - my @include_dirs2 = @include_dirs; # copy @include_dirs - shift(@include_dirs2); # remove "" entry, which is always the first - unshift(@include_dirs2, "") if (0 == @include_dirs2); - my $include_dir = shift(@include_dirs2); - print("Include dirs: $include_dir\n"); - foreach my $include_dir (@include_dirs2) { - print(" $include_dir\n"); - } - - my @user_ann_files = keys %user_ann_files; - unshift(@user_ann_files, "") if (0 == @user_ann_files); - my $user_ann_file = shift(@user_ann_files); - print("User annotated: $user_ann_file\n"); - foreach $user_ann_file (@user_ann_files) { - print(" $user_ann_file\n"); - } - - my $is_on = ($auto_annotate ? "on" : "off"); - print("Auto-annotation: $is_on\n"); - print("\n"); -} - -#----------------------------------------------------------------------------- -# Print summary and sorted function totals -#----------------------------------------------------------------------------- -sub mycmp ($$) -{ - my ($c, $d) = @_; - - # Iterate through sort events (eg. 3,2); return result if two are different - foreach my $i (@sort_order) { - my ($x, $y); - $x = $c->[$i]; - $y = $d->[$i]; - $x = -1 unless defined $x; - $y = -1 unless defined $y; - - my $cmp = abs($y) <=> abs($x); # reverse sort of absolute size - if (0 != $cmp) { - return $cmp; - } - } - # Exhausted events, equal - return 0; -} - -sub commify ($) { - my ($val) = @_; - 1 while ($val =~ s/^(-?\d+)(\d{3})/$1,$2/); - return $val; -} - -# Because the counts can get very big, and we don't want to waste screen space -# and make lines too long, we compute exactly how wide each column needs to be -# by finding the widest entry for each one. -sub compute_CC_col_widths (@) -{ - my @CCs = @_; - my $CC_col_widths = []; - - # Initialise with minimum widths (from event names) - foreach my $event (@events) { - push(@$CC_col_widths, length($event)); - } - - # Find maximum width count for each column. @CC_col_width positions - # correspond to @CC positions. - foreach my $CC (@CCs) { - foreach my $i (0 .. scalar(@$CC)-1) { - if (defined $CC->[$i]) { - # Find length, accounting for commas that will be added, and - # possibly a percentage. - my $length = length $CC->[$i]; - my $width = $length + int(($length - 1) / 3); - if ($show_percs) { - $width += 9; # e.g. " (12.34%)" is 9 chars - } - $CC_col_widths->[$i] = max($CC_col_widths->[$i], $width); - } - } - } - return $CC_col_widths; -} - -# Print the CC with each column's size dictated by $CC_col_widths. -sub print_CC ($$) -{ - my ($CC, $CC_col_widths) = @_; - - foreach my $i (@show_order) { - my $count = (defined $CC->[$i] ? commify($CC->[$i]) : "."); - - my $perc = ""; - if ($show_percs) { - if (defined $CC->[$i] && $CC->[$i] != 0) { - # Try our best to keep the number fitting into 5 chars. This - # requires dropping a digit after the decimal place if it's - # sufficiently negative (e.g. "-10.0") or positive (e.g. - # "100.0"). Thanks to diffs it's possible to have even more - # extreme values, like "-100.0" or "1000.0"; those rare case - # will end up with slightly wrong indenting, oh well. - $perc = safe_div($CC->[$i] * 100, $summary_CC->[$i]); - $perc = (-9.995 < $perc && $perc < 99.995) - ? sprintf(" (%5.2f%%)", $perc) - : sprintf(" (%5.1f%%)", $perc); - } else { - # Don't show percentages for "." and "0" entries. - $perc = " "; - } - } - - # $reps will be negative for the extreme values mentioned above. The - # use of max() avoids a possible warning about a negative repeat count. - my $text = $count . $perc; - my $len = length($text); - my $reps = $CC_col_widths->[$i] - length($text); - my $space = ' ' x max($reps, 0); - print("$space$text "); - } -} - -sub print_events ($) -{ - my ($CC_col_widths) = @_; - - foreach my $i (@show_order) { - my $event = $events[$i]; - my $event_width = length($event); - my $col_width = $CC_col_widths->[$i]; - my $space = ' ' x ($col_width - $event_width); - print("$event$space "); - } -} - -# Prints summary and function totals (with separate column widths, so that -# function names aren't pushed over unnecessarily by huge summary figures). -# Also returns a hash containing all the files that are involved in getting the -# events count above the thresholds (ie. all the interesting ones). -sub print_summary_and_fn_totals () -{ - my @fn_fullnames = keys %fn_totals; - - # Work out the size of each column for printing (summary and functions - # separately). - my $summary_CC_col_widths = compute_CC_col_widths($summary_CC); - my $fn_CC_col_widths = compute_CC_col_widths(values %fn_totals); - - # Header and counts for summary - print($fancy); - print_events($summary_CC_col_widths); - print("\n"); - print($fancy); - print_CC($summary_CC, $summary_CC_col_widths); - print(" PROGRAM TOTALS\n"); - print("\n"); - - # Header for functions - print($fancy); - print_events($fn_CC_col_widths); - print(" file:function\n"); - print($fancy); - - # Sort function names into order dictated by --sort option. - @fn_fullnames = sort { - mycmp($fn_totals{$a}, $fn_totals{$b}) - } @fn_fullnames; - - - # Assertion - (scalar @sort_order == scalar @thresholds) or - die("sort_order length != thresholds length:\n", - " @sort_order\n @thresholds\n"); - - my $threshold_files = {}; - # @curr_totals has the same shape as @sort_order and @thresholds - my @curr_totals = (); - foreach my $e (@thresholds) { - push(@curr_totals, 0); - } - - # Print functions, stopping when the threshold has been reached. - foreach my $fn_name (@fn_fullnames) { - - my $fn_CC = $fn_totals{$fn_name}; - - # Stop when we've reached all the thresholds - my $any_thresholds_exceeded = 0; - foreach my $i (0 .. scalar @thresholds - 1) { - my $prop = safe_div(abs($fn_CC->[$sort_order[$i]] * 100), - abs($summary_CC->[$sort_order[$i]])); - $any_thresholds_exceeded ||= ($prop >= $thresholds[$i]); - } - last if not $any_thresholds_exceeded; - - # Print function results - print_CC($fn_CC, $fn_CC_col_widths); - print(" $fn_name\n"); - - # Update the threshold counts - my $filename = $fn_name; - $filename =~ s/:.+$//; # remove function name - $threshold_files->{$filename} = 1; - foreach my $i (0 .. scalar @sort_order - 1) { - $curr_totals[$i] += $fn_CC->[$sort_order[$i]] - if (defined $fn_CC->[$sort_order[$i]]); - } - } - print("\n"); - - return $threshold_files; -} - -#----------------------------------------------------------------------------- -# Annotate selected files -#----------------------------------------------------------------------------- - -# Issue a warning that the source file is more recent than the input file. -sub warning_on_src_more_recent_than_inputfile ($) -{ - my $src_file = $_[0]; - - my $warning = <<END -@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -@@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ -@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -@ Source file '$src_file' is more recent than input file '$input_file'. -@ Annotations may not be correct. -@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ +# - Formatters: +# - `black`, for general formatting. This avoids the need for style checkers +# like `flake8`. Note that `black` allows a max line length of 88, which is +# a mild but common PEP-8 violation. +# - `isort`, for import sorting. +# +# - Type-checkers: +# - `mypy --strict`. This is the most commonly used Python type checker. +# - `pyright`. This is another good type checker. The `pyright: strict` +# comment above forces strict checking. +# - Sometimes one type-checker will complain about something the other does +# not. The goal is to keep both type checkers happy. +# +# - Linters: +# - `ruff`. Sometimes useful, and very fast to run. +# - `pylint`. Sometimes annoying, sometimes useful. The `pylintrc` +# modifies/disables the more annoying lints. +# +# - Profilers: +# - `cProfile` + `snakeviz`: Typically run with +# `python3 -m cProfile -o cg.prof cg_annotate $INPUT && snakeviz cg.prof`. +# - `scalene`. Typically run with `scalene ./cg_annotate $INPUT`. +# +# - Packager: +# - `cp` is used for distribution. This is possible because this program is a +# single file and only uses the Python Standard Library. This avoids the +# needs for any of the million different Python package management tools. + + +from __future__ import annotations + +import os +import re +import sys +from argparse import ArgumentParser, BooleanOptionalAction, Namespace +from collections import defaultdict +from typing import Callable, DefaultDict, NewType, NoReturn, TextIO, TypeAlias + + +class Args(Namespace): + """ + A typed wrapper for parsed args. + + None of these fields are modified after arg parsing finishes. + """ + + show: list[str] + sort: list[str] + threshold: float # a percentage + show_percs: bool + auto: bool + context: int + include: list[str] + cgout_filename: list[str] + src_filenames: list[str] + + @staticmethod + def parse() -> Args: + def comma_separated_list(values: str) -> list[str]: + return values.split(",") + + def threshold(n: str) -> float: + f = float(n) + if 0 <= f <= 20: + return f + raise ValueError + + def add_bool_argument(p: ArgumentParser, name: str, help: str) -> None: + """ + Add a bool argument that defaults to true. + + Supports these forms: `--foo`, `--no-foo`, `--foo=yes`, `--foo=no`. + The latter two were the forms supported by the old Perl version of + `cg_annotate`, and are now deprecated. + """ + flag = "--" + name + dest = name.replace("-", "_") + + # Note: the default value is always printed with `BooleanOptionalAction`, + # due to an argparse bug: https://github.com/python/cpython/issues/83137. + p.add_argument( + flag, + default=True, + action=BooleanOptionalAction, + help=help, + ) + p.add_argument( + f"{flag}=yes", + dest=dest, + action="store_true", + help=f"(deprecated) same as --{name}", + ) + p.add_argument( + f"{flag}=no", + dest=dest, + action="store_false", + help=f"(deprecated) same as --no-{name}", + ) + + p = ArgumentParser(description="Process Cachegrind output files.") + + p.add_argument("--version", action="version", version="%(prog)s-@VERSION@") + + p.add_argument( + "--show", + type=comma_separated_list, + metavar="A,B,C", + help="only show figures for events A,B,C (default: all events)", + ) + + p.add_argument( + "--sort", + type=comma_separated_list, + metavar="A,B,C", + help="sort functions by events A,B,C (default: event column order)", + ) + + p.add_argument( + "--threshold", + type=threshold, + default=0.1, + metavar="N:[0,20]", + help="only show functions with more than N%% of primary sort event " + "counts (default: %(default)s)", + ) + add_bool_argument( + p, + "show-percs", + "show a percentage for each non-zero count", + ) + add_bool_argument( + p, + "auto", + "annotate all source files containing functions that reached the " + "event count threshold", + ) + p.add_argument( + "--context", + type=int, + default=8, + metavar="N", + help="print N lines of context before and after annotated lines " + "(default: %(default)s)", + ) + p.add_argument( + "-I", + "--include", + action="append", + default=[], + metavar="D", + help="add D to the list of searched source file directories", + ) + p.add_argument( + "cgout_filename", + nargs=1, + metavar="cachegrind-out-file", + help="file produced by Cachegrind", + ) + p.add_argument( + "src_filenames", + nargs="*", + metavar="source-files", + help="source files to annotate (usually not needed due to --auto)", + ) + + return p.parse_args(namespace=Args()) + + +# Args are stored in a global for easy access. +args = Args.parse() + + +# A single instance of this class is constructed, from `args` and the `events:` +# line in the cgout file. +class Events: + # The event names. + events: list[str] + + # The order in which we must traverse events for --show. Can be shorter + # than `events`. + show_events: list[str] + + # Like `show_events`, but indices into `events`, rather than names. + show_indices: list[int] + + # The order in which we must traverse events for --sort. Can be shorter + # than `events`. + sort_events: list[str] + + # Like `sort_events`, but indices into `events`, rather than names. + sort_indices: list[int] + + # Threshold percentages, one per sort event. Dictates when we stop printing + # functions. Positions correspond to positions in `sort_events`. Only + # `thresholds[0]` is actually used for thresholding, for historical + # reasons. + threshold_percs: list[float] + + def __init__(self, text: str) -> None: + self.events = text.split() + self.num_events = len(self.events) + + # A temporary dict mapping events to indices, [0, n-1]. + event_indices = {event: n for n, event in enumerate(self.events)} + + # If --show is given, check it is valid. If --show is not given, + # default to all events in the standard order. + if args.show: + for event in args.show: + if event not in event_indices: + die(f"--show event `{event}` did not appear in `events:` line") + self.show_events = args.show + else: + self.show_events = self.events + + self.show_indices = [event_indices[event] for event in self.show_events] + + # Likewise for --sort. + if args.sort: + for event in args.sort: + if event not in event_indices: + die(f"--sort event `{event}` did not appear in `events:` line") + self.sort_events = args.sort + else: + self.sort_events = self.events + + self.sort_indices = [event_indices[event] for event in self.sort_events] + + # The primary sort event gets the --threshold value, and all other sort + # events get 100% (i.e. ignored). + self.threshold_percs = [100] * len(self.sort_events) + self.threshold_percs[0] = args.threshold + + def mk_cc(self, text: str) -> Cc: + # This is slightly faster than a list comprehension. + counts = list(map(int, text.split())) + + if len(counts) == self.num_events: + pass + elif len(counts) < self.num_events: + # Add zeroes at the end for any missing numbers. + counts.extend([0] * (self.num_events - len(counts))) + else: + raise ValueError + + return Cc(counts) + + def mk_empty_cc(self) -> Cc: + # This is much faster than a list comprehension. + return Cc([0] * self.num_events) + + +class Cc: + """ + This is a dumb container for counts. + + It doesn't know anything about events, i.e. what each count means. It can + do basic operations like `__iadd__` and `__eq__`, and anything more must be + done elsewhere. `Events.mk_cc` and `Events.mk_empty_cc` are used for + construction. + """ + + # Always the same length as `Events.events`. + counts: list[int] + + def __init__(self, counts: list[int]) -> None: + self.counts = counts + + def __repr__(self) -> str: + return str(self.counts) + + def __eq__(self, other: object) -> bool: + if not isinstance(other, Cc): + return NotImplemented + return self.counts == other.counts + + def __iadd__(self, other: Cc) -> Cc: + for i, other_count in enumerate(other.counts): + self.counts[i] += other_count + return self + + +# A paired filename and function name. +Flfn = NewType("Flfn", tuple[str, str]) + +# Per-function CCs. +DictFlfnCc: TypeAlias = DefaultDict[Flfn, Cc] + +# Per-line CCs, organised by filename and line number. +DictLineCc: TypeAlias = DefaultDict[int, Cc] +DictFlDictLineCc: TypeAlias = DefaultDict[str, DictLineCc] + + +def die(msg: str) -> NoReturn: + print("cg_annotate: error:", msg, file=sys.stderr) + sys.exit(1) + + +def read_cgout_file() -> tuple[str, str, Events, DictFlfnCc, DictFlDictLineCc, Cc]: + # The file format is described in Cachegrind's manual. + try: + cgout_file = open(args.cgout_filename[0], "r", encoding="utf-8") + except OSError as err: + die(f"{err}") + + with cgout_file: + cgout_line_num = 0 + + def parse_die(msg: str) -> NoReturn: + die(f"{cgout_file.name}:{cgout_line_num}: {msg}") + + def readline() -> str: + nonlocal cgout_line_num + cgout_line_num += 1 + return cgout_file.readline() + + # Read "desc:" lines. + desc = "" + while line := readline(): + if m := re.match(r"desc:\s+(.*)", line): + desc += m.group(1) + "\n" + else: + break + + # Read "cmd:" line. (`line` is already set from the "desc:" loop.) + if m := re.match(r"cmd:\s+(.*)", line): + cmd = m.group(1) + else: + parse_die("missing a `command:` line") + + # Read "events:" line. + line = readline() + if m := re.match(r"events:\s+(.*)", line): + events = Events(m.group(1)) + else: + parse_die("missing an `events:` line") + + def mk_empty_dict_line_cc() -> DictLineCc: + return defaultdict(events.mk_empty_cc) + + curr_fl = "" + curr_flfn = Flfn(("", "")) + + # Three different places where we accumulate CC data. + dict_flfn_cc: DictFlfnCc = defaultdict(events.mk_empty_cc) + dict_fl_dict_line_cc: DictFlDictLineCc = defaultdict(mk_empty_dict_line_cc) + summary_cc = None + + # Compile the one hot regex. + count_pat = re.compile(r"(\d+)\s+(.*)") + + # Line matching is done in order of pattern frequency, for speed. + while True: + line = readline() + + if m := count_pat.match(line): + line_num = int(m.group(1)) + try: + cc = events.mk_cc(m.group(2)) + except ValueError: + parse_die("malformed or too many event counts") + + # Record this CC at the function level. + flfn_cc = dict_flfn_cc[curr_flfn] + flfn_cc += cc + + # Record this CC at the file/line level. + line_cc = dict_fl_dict_line_cc[curr_fl][line_num] + line_cc += cc + + elif line.startswith("fn="): + curr_flfn = Flfn((curr_fl, line[3:-1])) + + elif line.startswith("fl="): + curr_fl = line[3:-1] + # A `fn=` line should follow, overwriting the "???". + curr_flfn = Flfn((curr_fl, "???")) + + elif m := re.match(r"summary:\s+(.*)", line): + try: + summary_cc = events.mk_cc(m.group(1)) + except ValueError: + parse_die("too many event counts") + + elif line == "": + break # EOF + + elif line == "\n" or line.startswith("#"): + # Skip empty lines and comment lines. + pass + + else: + parse_die(f"malformed line: {line[:-1]}") + + # Check if summary line was present. + if not summary_cc: + parse_die("missing `summary:` line, aborting") + + # Check summary is correct. + total_cc = events.mk_empty_cc() + for flfn_cc in dict_flfn_cc.values(): + total_cc += flfn_cc + if summary_cc != total_cc: + msg = ( + "`summary:` line doesn't match compute total\n" + f"- summary: {summary_cc}\n" + f"- total: {total_cc}" + ) + parse_die(msg) + + return (desc, cmd, events, dict_flfn_cc, dict_fl_dict_line_cc, summary_cc) + + +def safe_perc(m: int, n: int) -> float: + return 0 if n == 0 else m * 100 / n + + +class CcPrinter: + # Note: every `CcPrinter` gets the same `Events` object. + events: Events + + # Note: every `CcPrinter` gets the same summary CC. + summary_cc: Cc + + # The width of each event column. For simplicity, its length matches + # `events.events`, even though not all events are necessarily shown. + widths: list[int] + + def __init__(self, events: Events, ccs: list[Cc], summary_cc: Cc) -> None: + self.events = events + self.summary_cc = summary_cc + + # Find min and max value for each event. One of them will be the + # widest value. + min_cc = events.mk_empty_cc() + max_cc = events.mk_empty_cc() + for cc in ccs: + for i, _ in enumerate(events.events): + count = cc.counts[i] + if count > max_cc.counts[i]: + max_cc.counts[i] = count + elif count < min_cc.counts[i]: + min_cc.counts[i] = count + + # Find maximum width for each column. + self.widths = [0] * len(events.events) + for i, event in enumerate(events.events): + # Get widest of the min and max, accounting for commas that will be + # added, and a possible percentage. + width = max(len(str(min_cc.counts[i])), len(str(max_cc.counts[i]))) + width += (width - 1) // 3 + if args.show_percs: + width += 9 # e.g. " (12.34%)" is 9 chars. + + # Account for the event name, too. + self.widths[i] = max(width, len(event)) + + def print_events(self, suffix: str) -> None: + for i in self.events.show_indices: + # +1 is for the single space between columns. + print(f"{self.events.events[i]:{self.widths[i] + 1}}", end="") + + print(suffix) + + def print_count(self, i: int, text: str) -> None: + print(f"{text:>{self.widths[i]}}", end=" ") + + def print_cc(self, cc: Cc, suffix: str) -> None: + for i in self.events.show_indices: + nstr = f"{cc.counts[i]:,d}" # commify + if args.show_percs: + if cc.counts[i] != 0: + # Try our best to keep the number fitting into 5 chars. This + # requires dropping a digit after the decimal place if it's + # sufficiently negative (e.g. "-10.0") or positive (e.g. + # "100.0"). Thanks to diffs it's possible to have even more + # extreme values, like "-100.0" or "1000.0"; those rare case + # will end up with slightly wrong indenting, oh well. + p = safe_perc(cc.counts[i], self.summary_cc.counts[i]) + normal = -9.995 < p < 99.995 + perc = f" ({p:5.{2 if normal else 1}f}%)" + else: + # Don't show percentages for "0" entries, it's just clutter. + perc = " " + else: + perc = "" + + self.print_count(i, nstr + perc) + + print("", suffix) + + def print_missing_cc(self, suffix: str) -> None: + if args.show_percs: + # Don't show percentages for "." entries, it's just clutter. + text = ". " + else: + text = "." + + for i in self.events.show_indices: + self.print_count(i, text) + + print("", suffix) + + +# Used in various places in the output. +FANCY: str = "-" * 80 + + +def print_header(desc: str, cmd: str, events: Events) -> None: + print(FANCY) + print(desc, end="") + print("Command: ", cmd) + print("Data file: ", args.cgout_filename[0]) + print("Events recorded: ", *events.events) + print("Events shown: ", *events.show_events) + print("Event sort order:", *events.sort_events) + print("Thresholds: ", *events.threshold_percs) + + if len(args.include) == 0: + print("Include dirs: ") + else: + print(f"Include dirs: {args.include[0]}") + for include_dirname in args.include[1:]: + print(f" {include_dirname}") + + if len(args.src_filenames) == 0: + print("User annotated: ") + else: + print(f"User annotated: {args.src_filenames[0]}") + for src_filename in args.src_filenames[1:]: + print(f" {src_filename}") + + print("Auto-annotation: ", "on" if args.auto else "off") + print() + + +def print_summary_cc(events: Events, summary_cc: Cc) -> None: + printer = CcPrinter(events, [summary_cc], summary_cc) + + print(FANCY) + printer.print_events("") + print(FANCY) + printer.print_cc(summary_cc, "PROGRAM TOTALS") + print() + + +def print_flfn_ccs( + events: Events, dict_flfn_cc: DictFlfnCc, summary_cc: Cc +) -> set[str]: + # Only the first threshold percentage is actually used. + threshold_index = events.sort_indices[0] + + # Convert the threshold from a percentage to an event count. + threshold = ( + events.threshold_percs[0] * abs(summary_cc.counts[threshold_index]) / 100 + ) + + def meets_threshold(flfn_and_cc: tuple[Flfn, Cc]) -> bool: + cc = flfn_and_cc[1] + return abs(cc.counts[threshold_index]) >= threshold + + # Create a list with the counts in sort order, so that left-to-right list + # comparison does the right thing. Plus the `Flfn` at the end for + # deterministic output when all the event counts are identical in two CCs. + def key(flfn_and_cc: tuple[Flfn, Cc]) -> tuple[list[int], Flfn]: + cc = flfn_and_cc[1] + return ([abs(cc.counts[i]) for i in events.sort_indices], flfn_and_cc[0]) + + # Filter out functions for which the primary sort event count is below the + # threshold, and sort the remainder. + filtered_flfns_and_ccs = filter(meets_threshold, dict_flfn_cc.items()) + sorted_flfns_and_ccs = sorted(filtered_flfns_and_ccs, key=key, reverse=True) + sorted_ccs = list(map(lambda flfn_and_cc: flfn_and_cc[1], sorted_flfns_and_ccs)) + + printer = CcPrinter(events, sorted_ccs, summary_cc) + + print(FANCY) + printer.print_events(" file:function") + print(FANCY) -END -; - print($warning); -} - -# If there is information about lines not in the file, issue a warning -# explaining possible causes. -sub warning_on_nonexistent_lines ($$$) -{ - my ($src_more_recent_than_inputfile, $src_file, $excess_line_nums) = @_; - my $cause_and_solution; - - if ($src_more_recent_than_inputfile) { - $cause_and_solution = <<END -@@ cause: '$src_file' has changed since information was gathered. -@@ If so, a warning will have already been issued about this. -@@ solution: Recompile program and rerun under "valgrind --cachesim=yes" to -@@ gather new information. -END - # We suppress warnings about .h files - } elsif ($src_file =~ /\.h$/) { - $cause_and_solution = <<END -@@ cause: bug in the Valgrind's debug info reader that screws up with .h -@@ files sometimes -@@ solution: none, sorry -END - } else { - $cause_and_solution = <<END -@@ cause: not sure, sorry -END - } - - my $warning = <<END + # Print per-function counts. + for flfn, flfn_cc in sorted_flfns_and_ccs: + printer.print_cc(flfn_cc, f"{flfn[0]}:{flfn[1]}") + + print() + + # Files containing a function that met the threshold. + return set(flfn_and_cc[0][0] for flfn_and_cc in sorted_flfns_and_ccs) + + +def mk_warning(msg: str) -> str: + return f"""\ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ WARNING @@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -@@ -@@ Information recorded about lines past the end of '$src_file'. -@@ -@@ Probable cause and solution: -$cause_and_solution@@ +{msg}\ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ -END -; - print($warning); -} - -sub annotate_ann_files($) -{ - my ($threshold_files) = @_; - - my %all_ann_files; - my @unfound_auto_annotate_files; - my $printed_totals_CC = []; - - # If auto-annotating, add interesting files (but not "???") - if ($auto_annotate) { - delete $threshold_files->{"???"}; - %all_ann_files = (%user_ann_files, %$threshold_files) - } else { - %all_ann_files = %user_ann_files; - } - - # Track if we did any annotations. - my $did_annotations = 0; - - LOOP: - foreach my $src_file (keys %all_ann_files) { - - my $opened_file = ""; - my $full_file_name = ""; - # Nb: include_dirs already includes "", so it works in the case - # where the filename has the full path. - foreach my $include_dir (@include_dirs) { - my $try_name = $include_dir . $src_file; - if (open(INPUTFILE, "< $try_name")) { - $opened_file = $try_name; - $full_file_name = ($include_dir eq "" - ? $src_file - : "$include_dir + $src_file"); - last; - } - } - - if (not $opened_file) { - # Failed to open the file. If chosen on the command line, die. - # If arose from auto-annotation, print a little message. - if (defined $user_ann_files{$src_file}) { - die("File $src_file not opened in any of: @include_dirs\n"); - - } else { - push(@unfound_auto_annotate_files, $src_file); - } - - } else { - # File header (distinguish between user- and auto-selected files). - print("$fancy"); - my $ann_type = - (defined $user_ann_files{$src_file} ? "User" : "Auto"); - print("-- $ann_type-annotated source: $full_file_name\n"); - print("$fancy"); - - # Get file's CCs - my $src_file_CCs = $allCCs{$src_file}; - if (!defined $src_file_CCs) { - print(" No information has been collected for $src_file\n\n"); - next LOOP; - } - - $did_annotations = 1; - - # Numeric, not lexicographic sort! - my @line_nums = sort {$a <=> $b} keys %$src_file_CCs; - - # If $src_file more recent than cachegrind.out, issue warning - my $src_more_recent_than_inputfile = 0; - if ((stat $opened_file)[9] > (stat $input_file)[9]) { - $src_more_recent_than_inputfile = 1; - warning_on_src_more_recent_than_inputfile($src_file); - } - - # Work out the size of each column for printing - my $CC_col_widths = compute_CC_col_widths(values %$src_file_CCs); - - # Events header - print_events($CC_col_widths); - print("\n\n"); - - # Shift out 0 if it's in the line numbers (from unknown entries, - # likely due to bugs in Valgrind's stabs debug info reader) - shift(@line_nums) if (0 == $line_nums[0]); - - # Finds interesting line ranges -- all lines with a CC, and all - # lines within $context lines of a line with a CC. - my $n = @line_nums; - my @pairs; - for (my $i = 0; $i < $n; $i++) { - push(@pairs, $line_nums[$i] - $context); # lower marker - while ($i < $n-1 && - $line_nums[$i] + 2*$context >= $line_nums[$i+1]) { - $i++; - } - push(@pairs, $line_nums[$i] + $context); # upper marker - } - - # Annotate chosen lines, tracking total counts of lines printed - if (@pairs) { - $pairs[0] = 1 if ($pairs[0] < 1); - while (@pairs) { - my $low = shift @pairs; - my $high = shift @pairs; - while ($. < $low-1) { - my $tmp = <INPUTFILE>; - last unless (defined $tmp); # hack to detect EOF - } - my $src_line; - # Print line number, unless start of file - print("-- line $low " . '-' x 40 . "\n") if ($low != 1); - while (($. < $high) && ($src_line = <INPUTFILE>)) { - if (defined $line_nums[0] && $. == $line_nums[0]) { - print_CC($src_file_CCs->{$.}, $CC_col_widths); - add_array_a_to_b($src_file_CCs->{$.}, - $printed_totals_CC); - shift(@line_nums); - - } else { - print_CC([], $CC_col_widths); - } - - print(" $src_line"); - } - # Print line number, unless EOF - if ($src_line) { - print("-- line $high " . '-' x 40 . "\n"); - } else { - last; - } - } - } - - # If there was info on lines past the end of the file... - if (@line_nums) { - foreach my $line_num (@line_nums) { - print_CC($src_file_CCs->{$line_num}, $CC_col_widths); - print(" <bogus line $line_num>\n"); - } - print("\n"); - warning_on_nonexistent_lines($src_more_recent_than_inputfile, - $src_file, \@line_nums); - } - print("\n"); - - # Print summary of counts attributed to file but not to any - # particular line (due to incomplete debug info). - if ($src_file_CCs->{0}) { - print_CC($src_file_CCs->{0}, $CC_col_widths); - print(" <counts for unidentified lines in $src_file>\n\n"); - } - - close(INPUTFILE); - } - } - - # Print list of unfound auto-annotate selected files. - if (@unfound_auto_annotate_files) { - print("$fancy"); - print("The following files chosen for auto-annotation could not be found:\n"); - print($fancy); - foreach my $f (sort @unfound_auto_annotate_files) { - print(" $f\n"); - } - print("\n"); - } +""" + +def warn_src_file_is_newer(src_filename: str, cgout_filename: str) -> None: + msg = f"""\ +@ Source file '{src_filename}' is more recent than input file '{cgout_filename}'. +@ Annotations may not be correct. +""" + print(mk_warning(msg)) + + +def warn_bogus_lines(src_filename: str) -> None: +... [truncated message content] |
|
From: Paul F. <pa...@so...> - 2023-03-21 21:11:59
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=394fa9223aff48b6b344f0d55f8bc409a5425486 commit 394fa9223aff48b6b344f0d55f8bc409a5425486 Author: Paul Floyd <pj...@wa...> Date: Tue Mar 21 22:11:11 2023 +0100 Solaris: fix build, another aligned alloc typo Diff: --- coregrind/m_replacemalloc/vg_replace_malloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/coregrind/m_replacemalloc/vg_replace_malloc.c b/coregrind/m_replacemalloc/vg_replace_malloc.c index d3f33e66cf..db2fc5f309 100644 --- a/coregrind/m_replacemalloc/vg_replace_malloc.c +++ b/coregrind/m_replacemalloc/vg_replace_malloc.c @@ -1617,7 +1617,7 @@ extern int *___errno (void) __attribute__((weak)); #endif #if defined(VGO_solaris) -#define VG_MEMALIGD_ALIGN_FACTOR_FOUR 1 +#define VG_MEMALIGN_ALIGN_FACTOR_FOUR 1 #else #define VG_MEMALIGN_ALIGN_FACTOR_FOUR 0 #endif |
|
From: Paul F. <pa...@so...> - 2023-03-17 18:49:15
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=022a9caf40ceb32b8dfe258a8f2777f6299457d3 commit 022a9caf40ceb32b8dfe258a8f2777f6299457d3 Author: Paul Floyd <pj...@wa...> Date: Fri Mar 17 19:48:08 2023 +0100 Bug 467482 - Build failure on aarch64 Alpine Patch submitted by fan...@gm... Diff: --- NEWS | 9 ++++++--- coregrind/m_replacemalloc/vg_replace_malloc.c | 2 +- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/NEWS b/NEWS index cd20f7773a..8ede199d67 100644 --- a/NEWS +++ b/NEWS @@ -10,9 +10,6 @@ AMD64/macOS 10.13 and nanoMIPS/Linux. * ==================== CORE CHANGES =================== -* Make the address space limit on FreeBSD amd64 128Gbytes - (the same as Linux and Solaris, it was 32Gbytes) - * When GDB is used to debug a program running under valgrind using the valgrind gdbserver, GDB will automatically load some python code provided in valgrind defining GDB front end commands @@ -56,6 +53,11 @@ AMD64/macOS 10.13 and nanoMIPS/Linux. option to change the behaviour of Valgrind: --realloc-zero-bytes-frees=yes|no [yes on Linux glibc, no otherwise] +* ================== PLATFORM CHANGES ================= + +* Make the address space limit on FreeBSD amd64 128Gbytes + (the same as Linux and Solaris, it was 32Gbytes) + * ==================== TOOL CHANGES =================== * Memcheck: @@ -125,6 +127,7 @@ are not entered into bugzilla tend to get forgotten about or ignored. 464969 D language demangling 465435 m_libcfile.c:66 (vgPlain_safe_fd): Assertion 'newfd >= VG_(fd_hard_limit)' failed. 466104 aligned_alloc problems, part 1 +467482 Build failure on aarch64 Alpine n-i-bz FreeBSD rfork syscall fail with EINVAL or ENOSYS rather than VG_(unimplemented) To see details of a given bug, visit diff --git a/coregrind/m_replacemalloc/vg_replace_malloc.c b/coregrind/m_replacemalloc/vg_replace_malloc.c index c46e719c94..d3f33e66cf 100644 --- a/coregrind/m_replacemalloc/vg_replace_malloc.c +++ b/coregrind/m_replacemalloc/vg_replace_malloc.c @@ -1607,7 +1607,7 @@ extern int *___errno (void) __attribute__((weak)); #if defined(VGO_freebsd) #define VG_MEMALIGN_MAKE_SIZE_MULTIPLE_ALIGN 1 #else -#define VG_MEMALIG_MAKE_SIZE_MULTIPLE_ALIGN 0 +#define VG_MEMALIGN_MAKE_SIZE_MULTIPLE_ALIGN 0 #endif #if defined(VGO_solaris) |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-17 05:24:31
|
I have finished the rewrite. I am happy with the new code, it is much better than the old code. You can see it at https://bugs.kde.org/show_bug.cgi?id=467472. I plan to merge it by the end of next week, and I am happy to hear any suggestions. I also have some good news about the `cg_annotate.in`/`cg_annotate` split. I learned that you can generate the latter from the former very quickly with `config.status cachegrind/cg_annotate.in`. Also, this can be done automatically from some make targets. So I ended up creating a new make target `make ann` that can be run within the `cachegrind` directory. It runs the various Python formatters, type-checkers, and linters I am using on `cg_annotate.in` and then generates `cg_annotate`. It's a one-step "build" command that runs quickly, which is great. Nick On Wed, 15 Mar 2023 at 06:15, Nicholas Nethercote <n.n...@gm...> wrote: > On Sun, 12 Mar 2023 at 23:01, Paul Floyd <pj...@wa...> wrote: > >> >> The only think I can think of to get the version is to use something like >> >> pkg-config --modversion valgrind >> > > Thanks for the suggestion. Unfortunately this could cause misleading > results. E.g. if I have Valgrind installed on my system but I also have a > development version, when I run the development version of `cg_annotate > --version` it will claim to be the installed version. I think the `@VERSION@` > junk is unavoidable. > > Nick > |
|
From: Nicholas N. <n.n...@gm...> - 2023-03-14 19:16:17
|
On Sun, 12 Mar 2023 at 23:01, Paul Floyd <pj...@wa...> wrote: > > The only think I can think of to get the version is to use something like > > pkg-config --modversion valgrind > Thanks for the suggestion. Unfortunately this could cause misleading results. E.g. if I have Valgrind installed on my system but I also have a development version, when I run the development version of `cg_annotate --version` it will claim to be the installed version. I think the `@VERSION@` junk is unavoidable. Nick |
|
From: Paul F. <pj...@wa...> - 2023-03-12 15:56:06
|
Hi
I'm starting to look at adding alignment checks for the various aligned
allocators in memcheck.
That will give us errors like ASAN:
==89579==ERROR: AddressSanitizer: invalid alignment requested in
posix_memalign: -1, alignment must be a power of two and a multiple of
sizeof(void*) == 8 (thread T0)
#0 0x28e487 in posix_memalign
/usr/src/contrib/llvm-project/compiler-rt/lib/asan/asan_malloc_linux.cpp:210:3
#1 0x2b7c00 in main
/usr/home/paulf/scratch/valgrind/memcheck/tests/posix_memalign.c:39:10
#2 0x23542f in _start /usr/src/lib/csu/amd64/crt1_c.c:75:7
#3 0x8002dd007 (<unknown module>)
There are two problems. Firstly we map 4 C aligned allocation functions
(memalign, posix_memalign, aligned_alloc and valloc) to our one wrapper
interface, tl_memalign. So we don't know which function to check for.
Secondly we do argument checking / fixup in vg_replace_malloc. That
means that at the moment invalid args often cause a fail before getting
to memcheck.
To discriminate between the functions, the easiest thing will probably
be to add an enum tag (splitting tl_memalign into 3 will be a lot of
effort).
I see two possible solutions for the checks.
1/ move all the checking later so that memcheck can see the arguments.
2/ use a client req before all of the argument twiddling
I'm leaning more to using a client req as it has less impact on the
existing flow.
Any thoughts?
A+
Paul
|
|
From: Paul F. <pj...@wa...> - 2023-03-12 12:01:02
|
On 06-03-23 12:03, Nicholas Nethercote wrote: > Hi, > > Perl was a reasonable choice for `cg_annotate` when I first wrote it 20+ > years ago. But it's unfortunate now, with Perl being (a) a pretty weird > and horrible language, and (b) moribund. > > I'd like to rewrite it (and `cg_diff`) in Python, which will make > maintenance easier. I see that we already have some Python in Valgrind: > `coregrind/m_gdbserver/valgrind-monitor.py` and > `coregrind/m_gdbserver/valgrind-monitor-def.py`. Therefore I don't think > this should be controversial. > > But I might as well ask, just in case: any objections or advice? Because > these are single file scripts, we avoid all the usual problems of Python > packaging, and just use `cp` as the package manager :) > > On a related note, the `cg_annotate.in > <http://cg_annotate.in>`/`cg_annotate` split is annoying. The only > reason for it now is to auto-embed the version number into the script, > via the configure `@VERSION@` variable, for `cg_annotate --version` > output. Does anyone know of a way to achieve that without requiring > configure? Hi Nick Scripting languages aren't my strong point, but I'd say for me Python is the least worst choice. The only think I can think of to get the version is to use something like pkg-config --modversion valgrind Cheers Paul |
|
From: Paul F. <pa...@so...> - 2023-03-12 10:13:08
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=50f65336127b5f63c8af55544ce3e6a9bc834355 commit 50f65336127b5f63c8af55544ce3e6a9bc834355 Author: Paul Floyd <pj...@wa...> Date: Sun Mar 12 11:12:29 2023 +0100 coverity: a couple of potential buffer overruns Diff: --- coregrind/launcher-freebsd.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/coregrind/launcher-freebsd.c b/coregrind/launcher-freebsd.c index 792a4bd239..030b183b57 100644 --- a/coregrind/launcher-freebsd.c +++ b/coregrind/launcher-freebsd.c @@ -83,7 +83,7 @@ static const char *find_client(const char *clientname) while (path) { if ((colon = strchr(path, ':')) == NULL) { - strcpy(fullname, path); + strlcpy(fullname, path, PATH_MAX); path = NULL; } else { memcpy(fullname, path, colon - path); @@ -91,8 +91,8 @@ static const char *find_client(const char *clientname) path = colon + 1; } - strcat(fullname, "/"); - strcat(fullname, clientname); + strlcat(fullname, "/", PATH_MAX); + strlcat(fullname, clientname, PATH_MAX); if (access(fullname, R_OK|X_OK) == 0) { return fullname; |
|
From: Paul F. <pa...@so...> - 2023-03-12 08:40:29
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=d3a49926ad4c35e7d72bd435cf111d4cf507be54 commit d3a49926ad4c35e7d72bd435cf111d4cf507be54 Author: Paul Floyd <pj...@wa...> Date: Sun Mar 12 09:39:35 2023 +0100 coverity: printf args format mismatch Diff: --- coregrind/vgdb-invoker-freebsd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/coregrind/vgdb-invoker-freebsd.c b/coregrind/vgdb-invoker-freebsd.c index 4a8a97cb7c..27e1943685 100644 --- a/coregrind/vgdb-invoker-freebsd.c +++ b/coregrind/vgdb-invoker-freebsd.c @@ -315,7 +315,7 @@ Bool attach (pid_t pid, const char *msg) res = ptrace (PT_ATTACH, pid, 0, 0); if (res != 0) { if (output_error || debuglevel > 0) { - ERROR(errno, "%s PT_ATTACH pid %d %ld\n", msg, pid, res); + ERROR(errno, "%s PT_ATTACH pid %d %d\n", msg, pid, res); if (initial_attach) { output_error = False; } |
|
From: Paul F. <pa...@so...> - 2023-03-12 08:22:41
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=18c278f386ba0e028800faa9eac658dfe6e727b1 commit 18c278f386ba0e028800faa9eac658dfe6e727b1 Author: Paul Floyd <pj...@wa...> Date: Sun Mar 12 09:22:10 2023 +0100 coverity: copy paste error Diff: --- coregrind/m_syswrap/syswrap-amd64-freebsd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/coregrind/m_syswrap/syswrap-amd64-freebsd.c b/coregrind/m_syswrap/syswrap-amd64-freebsd.c index e54b157dd7..00c353ae41 100644 --- a/coregrind/m_syswrap/syswrap-amd64-freebsd.c +++ b/coregrind/m_syswrap/syswrap-amd64-freebsd.c @@ -962,7 +962,7 @@ POST(sys_wait6) } if (ARG6) { - POST_MEM_WRITE(ARG5, sizeof(vki_siginfo_t)); + POST_MEM_WRITE(ARG6, sizeof(vki_siginfo_t)); } } |
|
From: Paul F. <pa...@so...> - 2023-03-12 07:22:33
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=a1bb70bbf1f627461c30c626cbb853fce9411926 commit a1bb70bbf1f627461c30c626cbb853fce9411926 Author: Paul Floyd <pj...@wa...> Date: Sun Mar 12 08:21:55 2023 +0100 FreeBSD: fix a syswrap compiler warning Diff: --- coregrind/m_syswrap/syswrap-freebsd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/coregrind/m_syswrap/syswrap-freebsd.c b/coregrind/m_syswrap/syswrap-freebsd.c index 0d64d57605..fd4dff4da4 100644 --- a/coregrind/m_syswrap/syswrap-freebsd.c +++ b/coregrind/m_syswrap/syswrap-freebsd.c @@ -4898,7 +4898,7 @@ POST(sys_rtprio_thread) PRE(sys_sctp_generic_sendmsg) { *flags |= SfMayBlock; - PRINT("sys_sctp_generic_sendmsg ( %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d )",SARG1,ARG2,ARG3,ARG4,SARG5,ARG6,SARG7); + PRINT("sys_sctp_generic_sendmsg ( %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d, %#" FMT_REGWORD "x, %" FMT_REGWORD "d )",SARG1,ARG2,SARG3,ARG4,SARG5,ARG6,SARG7); PRE_REG_READ7(ssize_t, "sctp_generic_sendmsg", int, s, void *, msg, int, msglen, struct sockaddr *, to, socklen_t, len, |
|
From: Paul F. <pa...@so...> - 2023-03-12 07:21:47
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=8532c8b5e43ba4453ebfbe3169e4e9fc973fef3b commit 8532c8b5e43ba4453ebfbe3169e4e9fc973fef3b Author: Paul Floyd <pj...@wa...> Date: Sun Mar 12 08:20:57 2023 +0100 regtest: add an s390 expected for sized_aligned_new_delete_args Diff: --- memcheck/tests/Makefile.am | 1 + .../sized_aligned_new_delete_args.stderr.exp-s390 | 96 ++++++++++++++++++++++ 2 files changed, 97 insertions(+) diff --git a/memcheck/tests/Makefile.am b/memcheck/tests/Makefile.am index 7600981687..0509d45869 100644 --- a/memcheck/tests/Makefile.am +++ b/memcheck/tests/Makefile.am @@ -143,6 +143,7 @@ EXTRA_DIST = \ sized_aligned_new_delete_args.vgtest \ sized_aligned_new_delete_args.stderr.exp_32 \ sized_aligned_new_delete_args.stderr.exp-gcc \ + sized_aligned_new_delete_args.stderr.exp-s390 \ sized_aligned_new_delete_misaligned.stderr.exp \ sized_aligned_new_delete_misaligned.vgtest \ deep-backtrace.vgtest deep-backtrace.stderr.exp \ diff --git a/memcheck/tests/sized_aligned_new_delete_args.stderr.exp-s390 b/memcheck/tests/sized_aligned_new_delete_args.stderr.exp-s390 new file mode 100644 index 0000000000..ea74f0f3d7 --- /dev/null +++ b/memcheck/tests/sized_aligned_new_delete_args.stderr.exp-s390 @@ -0,0 +1,96 @@ +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:16) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:16) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:16) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete(void*, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:17) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:19) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:19) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:19) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete[](void*, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:20) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:24) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:24) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:24) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new(unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:24) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete(void*, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:25) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:27) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:27) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:27) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator new[](unsigned long, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:27) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete[](void*, std::align_val_t, std::nothrow_t const&) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:28) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: ...operator delete... (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:31) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete[](void*, unsigned long) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:34) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete(void*, unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:37) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete(void*, unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:37) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete[](void*, unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:40) + +Conditional jump or move depends on uninitialised value(s) + at 0x........: operator delete[](void*, unsigned long, std::align_val_t) (vg_replace_malloc.c:...) + by 0x........: main (sized_aligned_new_delete_args.cpp:40) + |
|
From: Paul F. <pa...@so...> - 2023-03-11 15:49:13
|
https://sourceware.org/git/gitweb.cgi?p=valgrind.git;h=96a66dfda6e0f4c43e633aa4bd2a60389073a41c commit 96a66dfda6e0f4c43e633aa4bd2a60389073a41c Author: Paul Floyd <pj...@wa...> Date: Sat Mar 11 16:47:56 2023 +0100 doc: Add asan equivalent to --realloc-zero-bytes-frees to manual-core Diff: --- docs/xml/manual-core.xml | 3 +++ 1 file changed, 3 insertions(+) diff --git a/docs/xml/manual-core.xml b/docs/xml/manual-core.xml index 20886fe49e..296fc1c2a9 100644 --- a/docs/xml/manual-core.xml +++ b/docs/xml/manual-core.xml @@ -1802,6 +1802,9 @@ that can report errors, e.g. Memcheck, but not Cachegrind.</para> consider using <computeroutput>--realloc-zero-bytes-frees=no</computeroutput>. </para> + <para>Address Sanitizer has a similar and even wordier option + <computeroutput>allocator_frees_and_returns_null_on_realloc_zero</computeroutput>. + </para> </listitem> </varlistentry> |