Menu
▾
▴
tuningforkvp-users — Tuning Fork user mailing list
You can subscribe to this list here.
| 2008 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(10) |
Aug
(16) |
Sep
(3) |
Oct
(5) |
Nov
(4) |
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2009 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
|
From: Donnie J. <do...@da...> - 2008-08-01 23:36:49
|
Hello, I uploaded another screenshot for example. I would like to accomplish this same horizontal bar timeline graph but with each threads execution "states" on its own horizontal line. Screenshot: http://darthik.com/tuning_fork/ghc/images/screenshot-TfCapTimerEventsMultFeedlets.png Thank you. __ Donnie Jones On Fri, Aug 1, 2008 at 7:31 PM, Donnie Jones <do...@da...> wrote: > Hello, > > I have a follow-up question. If I have two threads executing, and I want > to create a timeline graph of the "state" changes of each thread. The > "states" are the same EventTypes for each thread, thus when I graph the > threads "state" changes the TimeIntervalStream that is created combines the > events into a single stream for each "state". I would like to graph an > individual horizontal bar with colored "states" for each thread. > > Any suggestions? > Thanks again. > __ > Donnie Jones > > > On Thu, Jul 31, 2008 at 9:24 PM, Donnie Jones <do...@da...> wrote: > >> Hello David, >> >> That worked. Thank you. >> __ >> Donnie >> >> >> On Thu, Jul 31, 2008 at 7:27 PM, David F. Bacon <df...@wa...>wrote: >> >>> select "cycle through display modes" from the dropdown in the figure (or >>> use the button with the same icon). i think this will give you what you >>> want. >>> david >>> >>> >>> On Jul 31, 2008, at 7:11 PM, Donnie Jones wrote: >>> >>> Hello, >>> >>> I am attempting to create a timeline graph of the changes in "states" of >>> execution for the Glasgow Haskell Compiler (GHC) Run-time System (RTS). I >>> would like to be able to create a timeline graph which has a single >>> horizontal bar (same location on the y-axis) in which the colors change for >>> the different "states". So far, I have only been able to accomplish >>> timeline graphs where the different states are on separate horizontal bars >>> on the y-axis in the timeline; I have linked a screenshot of such a graph >>> [1]. This graph was created as a TimeIntervalStream in which each >>> TimerEvent was a different "state" of execution. >>> Any suggestions would be great. >>> Thank you. >>> __ >>> Donnie Jones >>> >>> 1. >>> http://darthik.com/tuning_fork/ghc/images/screenshot-TfCapTimerEvents.png >>> >>> >>> >>> ------------------------------------------------------------------------- >>> This SF.Net email is sponsored by the Moblin Your Move Developer's >>> challenge >>> Build the coolest Linux based applications with Moblin SDK & win great >>> prizes >>> Grand prize is a trip for two to an Open Source event anywhere in the >>> world >>> >>> http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ >>> Tuningforkvp-users mailing list >>> Tun...@li... >>> https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users >>> >>> >>> >> > |
|
From: Donnie J. <do...@da...> - 2008-08-01 23:30:55
|
Hello, I have a follow-up question. If I have two threads executing, and I want to create a timeline graph of the "state" changes of each thread. The "states" are the same EventTypes for each thread, thus when I graph the threads "state" changes the TimeIntervalStream that is created combines the events into a single stream for each "state". I would like to graph an individual horizontal bar with colored "states" for each thread. Any suggestions? Thanks again. __ Donnie Jones On Thu, Jul 31, 2008 at 9:24 PM, Donnie Jones <do...@da...> wrote: > Hello David, > > That worked. Thank you. > __ > Donnie > > > On Thu, Jul 31, 2008 at 7:27 PM, David F. Bacon <df...@wa...>wrote: > >> select "cycle through display modes" from the dropdown in the figure (or >> use the button with the same icon). i think this will give you what you >> want. >> david >> >> >> On Jul 31, 2008, at 7:11 PM, Donnie Jones wrote: >> >> Hello, >> >> I am attempting to create a timeline graph of the changes in "states" of >> execution for the Glasgow Haskell Compiler (GHC) Run-time System (RTS). I >> would like to be able to create a timeline graph which has a single >> horizontal bar (same location on the y-axis) in which the colors change for >> the different "states". So far, I have only been able to accomplish >> timeline graphs where the different states are on separate horizontal bars >> on the y-axis in the timeline; I have linked a screenshot of such a graph >> [1]. This graph was created as a TimeIntervalStream in which each >> TimerEvent was a different "state" of execution. >> Any suggestions would be great. >> Thank you. >> __ >> Donnie Jones >> >> 1. >> http://darthik.com/tuning_fork/ghc/images/screenshot-TfCapTimerEvents.png >> >> >> >> ------------------------------------------------------------------------- >> This SF.Net email is sponsored by the Moblin Your Move Developer's >> challenge >> Build the coolest Linux based applications with Moblin SDK & win great >> prizes >> Grand prize is a trip for two to an Open Source event anywhere in the >> world >> >> http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ >> Tuningforkvp-users mailing list >> Tun...@li... >> https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users >> >> >> > |
|
From: Donnie J. <do...@da...> - 2008-08-01 01:24:05
|
Hello David, That worked. Thank you. __ Donnie On Thu, Jul 31, 2008 at 7:27 PM, David F. Bacon <df...@wa...> wrote: > select "cycle through display modes" from the dropdown in the figure (or > use the button with the same icon). i think this will give you what you > want. > david > > > On Jul 31, 2008, at 7:11 PM, Donnie Jones wrote: > > Hello, > > I am attempting to create a timeline graph of the changes in "states" of > execution for the Glasgow Haskell Compiler (GHC) Run-time System (RTS). I > would like to be able to create a timeline graph which has a single > horizontal bar (same location on the y-axis) in which the colors change for > the different "states". So far, I have only been able to accomplish > timeline graphs where the different states are on separate horizontal bars > on the y-axis in the timeline; I have linked a screenshot of such a graph > [1]. This graph was created as a TimeIntervalStream in which each > TimerEvent was a different "state" of execution. > Any suggestions would be great. > Thank you. > __ > Donnie Jones > > 1. > http://darthik.com/tuning_fork/ghc/images/screenshot-TfCapTimerEvents.png > > > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's > challenge > Build the coolest Linux based applications with Moblin SDK & win great > prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > > http://moblin-contest.org/redirect.php?banner_id=100&url=/_______________________________________________ > Tuningforkvp-users mailing list > Tun...@li... > https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users > > > |
|
From: David F. B. <df...@wa...> - 2008-07-31 23:28:36
|
select "cycle through display modes" from the dropdown in the figure (or use the button with the same icon). i think this will give you what you want. david On Jul 31, 2008, at 7:11 PM, Donnie Jones wrote: > Hello, > > I am attempting to create a timeline graph of the changes in > "states" of execution for the Glasgow Haskell Compiler (GHC) Run- > time System (RTS). I would like to be able to create a timeline > graph which has a single horizontal bar (same location on the y- > axis) in which the colors change for the different "states". So > far, I have only been able to accomplish timeline graphs where the > different states are on separate horizontal bars on the y-axis in > the timeline; I have linked a screenshot of such a graph [1]. This > graph was created as a TimeIntervalStream in which each TimerEvent > was a different "state" of execution. > > Any suggestions would be great. > Thank you. > __ > Donnie Jones > > 1. http://darthik.com/tuning_fork/ghc/images/screenshot- > TfCapTimerEvents.png > > > > ---------------------------------------------------------------------- > --- > This SF.Net email is sponsored by the Moblin Your Move Developer's > challenge > Build the coolest Linux based applications with Moblin SDK & win > great prizes > Grand prize is a trip for two to an Open Source event anywhere in > the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Tuningforkvp-users mailing list > Tun...@li... > https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users |
|
From: Donnie J. <do...@da...> - 2008-07-31 23:11:48
|
Hello, I am attempting to create a timeline graph of the changes in "states" of execution for the Glasgow Haskell Compiler (GHC) Run-time System (RTS). I would like to be able to create a timeline graph which has a single horizontal bar (same location on the y-axis) in which the colors change for the different "states". So far, I have only been able to accomplish timeline graphs where the different states are on separate horizontal bars on the y-axis in the timeline; I have linked a screenshot of such a graph [1]. This graph was created as a TimeIntervalStream in which each TimerEvent was a different "state" of execution. Any suggestions would be great. Thank you. __ Donnie Jones 1. http://darthik.com/tuning_fork/ghc/images/screenshot-TfCapTimerEvents.png |
|
From: Joshua A. <jo...@us...> - 2008-07-17 15:01:10
|
That sounds like the right explanation. I think what this illustrates is
that the CTraceGenerationLibrary is pretty low level and requires a certain
amount of side knowledge as you are discovering by trial and error. The
CXXTraceGenerationLibrary is much higher level and takes care of all of
those details for you, along with storage management, I/O, etc. But, for
historical reasons, the higher layer library is written in C++, which may
not be suitable for everyone's needs. It is actually only recently that
we even had a plain C library at all (the C++ library is older and used to
have its own low-level chunk management instead of being build on the C
library). When we developed the plain C library and rebased the C++
library on it, we decided to deliberately make the C layer very low level
to facilitate its being used in a very wide set of circumstances (including
device drivers, perhaps microcontrollers, etc). But, there is probably a
constituency for a somewhat higher level library in plain C which is still
very agnostic about I/O and storage management but takes care of a little
bit more for you. Two obvious candidates for added services are this
business of assigning feedlet and sequence information correctly to event
chunks and also the business of getting an appropriate timestamp to use.
If you have more thoughts about what would make C trace generation easier
without putting in lots of system-dependent stuff, please send them along.
Also, if you want to get more involved in the project from an open source
standpoint (contributing code as well as ideas) we should talk about that
as well.
Josh
"Donnie Jones"
<donnie@darthik.c
om> To
Sent by: tun...@li...
tuningforkvp-user e.net
s-bounces@lists.s cc
ourceforge.net Simon Marlow <mar...@gm...>
Subject
Re: [Tuningforkvp-users] Feedlet 1
07/17/2008 09:34 Missing -1 Chunks?
AM
Hello,
I believe I have figured out why I was receiving this error message. For
each time I flush the events to the log file, I needed to increment the
sequenceNumber for the eventChunk. Doing so also made the error go away.
Thank you.
__
Donnie
On Thu, Jul 17, 2008 at 8:07 AM, Donnie Jones <do...@da...> wrote:
Hello,
I am using the CTraceGenerationLibrary with Tuning Fork. I have attached
my test.trace file and test code TfTest.c.
I was wondering if you could possibly explain why I am getting the error
pasted below...
### begin error ###
Connected to
file: /home/donnie/UK/masters/thesis/tuning_fork_vp/ghc/test/test.trace
FiniteSeekableFeed.getCanonicalId: crc = 0
Sampled 108 of 323 KBytes. CRC took 1ms. CRC32 = 3725828339
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last
known sequence number is 0
Chunk indexing of 24 chunks took 0.0 seconds
Event indexing of 20680 events took 0.0 seconds
### end error ###
I have confirmed that I added 20680 events, so the "Event indexing of
20680 events" is correct.
Thank you.
__
Donnie
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's
challenge
Build the coolest Linux based applications with Moblin SDK & win great
prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Tuningforkvp-users mailing list
Tun...@li...
https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users
|
|
From: Donnie J. <do...@da...> - 2008-07-17 13:34:34
|
Hello, I believe I have figured out why I was receiving this error message. For each time I flush the events to the log file, I needed to increment the sequenceNumber for the eventChunk. Doing so also made the error go away. Thank you. __ Donnie On Thu, Jul 17, 2008 at 8:07 AM, Donnie Jones <do...@da...> wrote: > Hello, > > I am using the CTraceGenerationLibrary with Tuning Fork. I have attached > my test.trace file and test code TfTest.c. > I was wondering if you could possibly explain why I am getting the error > pasted below... > > ### begin error ### > Connected to file: > /home/donnie/UK/masters/thesis/tuning_fork_vp/ghc/test/test.trace > FiniteSeekableFeed.getCanonicalId: crc = 0 > Sampled 108 of 323 KBytes. CRC took 1ms. CRC32 = 3725828339 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last > known sequence number is 0 > Chunk indexing of 24 chunks took 0.0 seconds > Event indexing of 20680 events took 0.0 seconds > ### end error ### > > I have confirmed that I added 20680 events, so the "Event indexing of 20680 > events" is correct. > Thank you. > __ > Donnie > |
|
From: Donnie J. <do...@da...> - 2008-07-17 12:07:23
|
Hello, I am using the CTraceGenerationLibrary with Tuning Fork. I have attached my test.trace file and test code TfTest.c. I was wondering if you could possibly explain why I am getting the error pasted below... ### begin error ### Connected to file: /home/donnie/UK/masters/thesis/tuning_fork_vp/ghc/test/test.trace FiniteSeekableFeed.getCanonicalId: crc = 0 Sampled 108 of 323 KBytes. CRC took 1ms. CRC32 = 3725828339 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Feedlet 1 is missing -1 chunks. Current sequence number is 0 while last known sequence number is 0 Chunk indexing of 24 chunks took 0.0 seconds Event indexing of 20680 events took 0.0 seconds ### end error ### I have confirmed that I added 20680 events, so the "Event indexing of 20680 events" is correct. Thank you. __ Donnie |
|
From: Donnie J. <do...@da...> - 2008-07-16 10:04:06
|
Hello Joshua, On Tue, Jul 15, 2008 at 12:09 PM, Joshua Auerbach <jo...@us...> wrote: > Probably the best way to understand how to add storage management on top of > the C Trace Generation Library is to look at the C++ Trace Generation > library (project CXXTraceGenerationLibrary). This code is basically a > higher layer built on top of the C layer. In fact, if you are operating in > a code environment that tolerates the use of C++ and you have complex > storage management needs, you might just switch to using it. > > On the other hand, you can also just look at what the C++ code does and use > it as model for doing similar things in plain C. It uses tfStackChunk to > build up all the non-event chunks contiguously in one area (see occurrances > of tfStackChunk in Logger.cpp) and then has a separate pool of Event chunks > that go through a lifecycle of being either in use or free (see methods > like Logger::getEmptyEventChunk for instance). The reason that we > maintain this separation is to support logging to a socket as well as > file ... when something freshly connects to a socket, it will want to get > all the metadata chunks but only those event chunks that the trace library > can afford to keep around. If you are logging just to files you might not > need this finepoint but I think it's useful. > > Other strategies are possible, and tfInitChunk will certainly initialize a > chunk to the new, empty state. Use methods like tfGetStartAddress and > tfGetEndAddress to find the memory extent of either single chunks or > contiguous arrays of chunks. If this isn't enough information, let me know > more about what you're doing and I'll try to provide more specific advice. > > Josh > Thank you for the quick reply. I will take a closer look at the CXXTraceGenerationLibrary and let you know if I have any issues. __ Donnie Jones |
|
From: Joshua A. <jo...@us...> - 2008-07-15 17:16:30
|
Probably the best way to understand how to add storage management on top of the C Trace Generation Library is to look at the C++ Trace Generation library (project CXXTraceGenerationLibrary). This code is basically a higher layer built on top of the C layer. In fact, if you are operating in a code environment that tolerates the use of C++ and you have complex storage management needs, you might just switch to using it. On the other hand, you can also just look at what the C++ code does and use it as model for doing similar things in plain C. It uses tfStackChunk to build up all the non-event chunks contiguously in one area (see occurrances of tfStackChunk in Logger.cpp) and then has a separate pool of Event chunks that go through a lifecycle of being either in use or free (see methods like Logger::getEmptyEventChunk for instance). The reason that we maintain this separation is to support logging to a socket as well as file ... when something freshly connects to a socket, it will want to get all the metadata chunks but only those event chunks that the trace library can afford to keep around. If you are logging just to files you might not need this finepoint but I think it's useful. Other strategies are possible, and tfInitChunk will certainly initialize a chunk to the new, empty state. Use methods like tfGetStartAddress and tfGetEndAddress to find the memory extent of either single chunks or contiguous arrays of chunks. If this isn't enough information, let me know more about what you're doing and I'll try to provide more specific advice. Josh > > Hello, > > I am using the C Trace Generation library as a model for my C code > where I am generating traces for Tuning Fork. If the storage is > filled and tfHasRoom returns 0 for out of room, I would like to > write out the current chunks (headerChunk, feedletChunk, > eventTypeSpaceChunk, eventTypeChunk, and eventChunks) to a file with > fwrite() and then re-use the storage. Is it best to tfInitChunk() > the eventChunk so that the next call to fwrite() will only write out > the data in the eventChunks? > > Any help would be great! :) > Thank you. > __ > Donnie Jones > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Tuningforkvp-users mailing list > Tun...@li... > https://lists.sourceforge.net/lists/listinfo/tuningforkvp-users |
|
From: Donnie J. <do...@da...> - 2008-07-15 14:36:38
|
Hello, On Tue, Jul 15, 2008 at 9:59 AM, Donnie Jones <do...@da...> wrote: > Hello, > > I modified the example testCTracing.c to use tfAddIntEvent() instead of > tfAddDoubleEvent(), but when I load the test.trace file into Tuning Fork, I > get this error: > > ### begin error ### > Connected to file: (removed path)/CTraceGenerationLibrary/test.trace > FiniteSeekableFeed.getCanonicalId: crc = 0 > Sampled 0 of 0 KBytes. CRC took 0ms. CRC32 = 3688122119 > Dumping chunk 4. Limit = 88 position = 60 > 0: 00 00 00 01 00 00 00 00 10 e0 8e 2f 4d 87 1b 48 00 > 00 00 01 > 20: 00 00 00 00 10 e0 8e 2f 89 21 e5 48 00 00 00 01 00 > 00 00 01 > 40: 10 e0 8e 2f c4 bc af 48 00 00 00 01 00 00 00 02 10 > e0 8e 30 ../...H > 60: 00 57 79 48 00 00 00 01 00 00 00 03 10 e0 8e 30 > Unknown event type index 283151920 while parsing event. > Feedlet.indexChunkDescriptor died at chunk 4 of len 88 and position 60 with > an EventParserException: Unknown event type index 283151920 while parsing > event > Chunk indexing of 5 chunks took 0.0 seconds > Unknown event index 283151920 near position 60 > processEventChunk died at chunk 4 with an EventParserException: Unknown > event index 283151920 near position 60 > Event indexing of 1 events took 0.0 seconds > ### end error ### > > I am using Tuning Fork at revision 399 from the Subversion repository. > Thank you. > __ > Donnie Jones > I realized the error that I made... I had not changed the tfAddEventType() to reflect that an integer event type was being added, instead of a double, hence the parsing issue with the file. Thanks anyway. __ Donnie Jones |
|
From: Donnie J. <do...@da...> - 2008-07-15 13:59:40
|
Hello, I modified the example testCTracing.c to use tfAddIntEvent() instead of tfAddDoubleEvent(), but when I load the test.trace file into Tuning Fork, I get this error: ### begin error ### Connected to file: (removed path)/CTraceGenerationLibrary/test.trace FiniteSeekableFeed.getCanonicalId: crc = 0 Sampled 0 of 0 KBytes. CRC took 0ms. CRC32 = 3688122119 Dumping chunk 4. Limit = 88 position = 60 0: 00 00 00 01 00 00 00 00 10 e0 8e 2f 4d 87 1b 48 00 00 00 01 20: 00 00 00 00 10 e0 8e 2f 89 21 e5 48 00 00 00 01 00 00 00 01 40: 10 e0 8e 2f c4 bc af 48 00 00 00 01 00 00 00 02 10 e0 8e 30 ../...H 60: 00 57 79 48 00 00 00 01 00 00 00 03 10 e0 8e 30 Unknown event type index 283151920 while parsing event. Feedlet.indexChunkDescriptor died at chunk 4 of len 88 and position 60 with an EventParserException: Unknown event type index 283151920 while parsing event Chunk indexing of 5 chunks took 0.0 seconds Unknown event index 283151920 near position 60 processEventChunk died at chunk 4 with an EventParserException: Unknown event index 283151920 near position 60 Event indexing of 1 events took 0.0 seconds ### end error ### I am using Tuning Fork at revision 399 from the Subversion repository. Thank you. __ Donnie Jones |
|
From: Donnie J. <do...@da...> - 2008-07-15 11:05:24
|
Hello, I am using the C Trace Generation library as a model for my C code where I am generating traces for Tuning Fork. If the storage is filled and tfHasRoom returns 0 for out of room, I would like to write out the current chunks (headerChunk, feedletChunk, eventTypeSpaceChunk, eventTypeChunk, and eventChunks) to a file with fwrite() and then re-use the storage. Is it best to tfInitChunk() the eventChunk so that the next call to fwrite() will only write out the data in the eventChunks? Any help would be great! :) Thank you. __ Donnie Jones |
2 messages has been excluded from this view by a project administrator.
