|
From: Daniel V. <vei...@re...> - 2003-10-13 23:30:29
|
To me Valgrind allows me to automate very thorough regression tests and get a really high level of confidence in the code that passed though it. But I have still a concern about the code that did not get exercized though my valgrind regression tests. So I am wondering if someone made some code coverage analysis tool based on valgrind. It would require building, saving and loading annotated maps of the text segment and a report generator. Is there anything similar ? I think it would make a very nice addition to kcachegrind which does already most of the UI work needed to provide such information, maybe working with a (potentially large) set of cachegrind files this could be possible. Does this exists ? Daniel -- Daniel Veillard | Red Hat Network https://rhn.redhat.com/ vei...@re... | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ |
|
From: Jeremy F. <je...@go...> - 2003-10-13 23:46:38
|
On Mon, 2003-10-13 at 16:30, Daniel Veillard wrote: > To me Valgrind allows me to automate very thorough regression tests > and get a really high level of confidence in the code that passed though > it. But I have still a concern about the code that did not get exercized > though my valgrind regression tests. So I am wondering if someone made > some code coverage analysis tool based on valgrind. It would require > building, saving and loading annotated maps of the text segment and > a report generator. Is there anything similar ? I think it would make > a very nice addition to kcachegrind which does already most of the > UI work needed to provide such information, maybe working with a > (potentially large) set of cachegrind files this could be possible. > Does this exists ? vgprof (see http://www.goop.org/~jeremy/valgrind) does this. It generates gprof-style output, which seemed like a good idea at the time, but there's probably some more useful format it could be generating (gcov?). J |
|
From: Daniel V. <vei...@re...> - 2003-10-13 23:54:14
|
On Mon, Oct 13, 2003 at 04:44:28PM -0700, Jeremy Fitzhardinge wrote: > On Mon, 2003-10-13 at 16:30, Daniel Veillard wrote: > > To me Valgrind allows me to automate very thorough regression tests > > and get a really high level of confidence in the code that passed though > > it. But I have still a concern about the code that did not get exercized > > though my valgrind regression tests. So I am wondering if someone made > > some code coverage analysis tool based on valgrind. It would require > > building, saving and loading annotated maps of the text segment and > > a report generator. Is there anything similar ? I think it would make > > a very nice addition to kcachegrind which does already most of the > > UI work needed to provide such information, maybe working with a > > (potentially large) set of cachegrind files this could be possible. > > Does this exists ? > > vgprof (see http://www.goop.org/~jeremy/valgrind) does this. It > generates gprof-style output, which seemed like a good idea at the time, > but there's probably some more useful format it could be generating > (gcov?). Cool :-), I will check this out (TODO++), the multiple vgmon.out input should allow to do what I want. Thanks ! Daniel -- Daniel Veillard | Red Hat Network https://rhn.redhat.com/ vei...@re... | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ |
|
From: Nicholas N. <nj...@ca...> - 2003-10-14 07:52:39
|
On Tue, 13 Oct 2003, Jeremy Fitzhardinge wrote: > > So I am wondering if someone made some code coverage analysis tool > > based on valgrind. > > vgprof (see http://www.goop.org/~jeremy/valgrind) does this. Cachegrind (normal version, not KCachegrind) also gives you this information as a side-effect of its normal workings, since the "Ir" column gives the number of instructions executed per line. Use the "--show=Ir" option to cg_annotate to only show this column. N |
|
From: Daniel V. <vei...@re...> - 2003-10-14 08:03:56
|
On Tue, Oct 14, 2003 at 08:52:32AM +0100, Nicholas Nethercote wrote: > On Tue, 13 Oct 2003, Jeremy Fitzhardinge wrote: > > > > So I am wondering if someone made some code coverage analysis tool > > > based on valgrind. > > > > vgprof (see http://www.goop.org/~jeremy/valgrind) does this. > > Cachegrind (normal version, not KCachegrind) also gives you this > information as a side-effect of its normal workings, since the "Ir" column > gives the number of instructions executed per line. Use the "--show=Ir" > option to cg_annotate to only show this column. Okay, and I assume kcachegrind uses it but only for a single cachegrind information file. I wonder how hard it would be to generate a coverage mode for kcachegrind based on a set of files, that would be a nice addition from my perspective :-) thanks for the informations ! Daniel -- Daniel Veillard | Red Hat Network https://rhn.redhat.com/ vei...@re... | libxml GNOME XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/ |