|
From: Stefan K. <en...@ho...> - 2009-12-26 16:53:08
|
Hi, I am quite interested in a coverage tool that does not need recompilation. There seems to be a patched qemu somewhere, but then I found valgrind-vcov. I had some issue building the branch. mv -f .deps/memcheck_x86_linux-mc_main.Tpo .deps/memcheck_x86_linux-mc_main.Po gcc -DHAVE_CONFIG_H -I. -I.. -I.. -I../include -I../VEX/pub -DVGA_x86=1 -DVGO_linux=1 -DVGP_x86_linux=1 -m32 -mpreferred-stack-boundary=2 -O2 -g -Wall -Wmissing-prototypes -Wshadow -Wpointer-arith -Wstrict-prototypes -Wmissing-declarations -Wno-format-zero-length -fno-strict-aliasing -O2 -Wno-long-long -Wno-pointer-sign -Wdeclaration-after-statement -fno-stack-protector -MT memcheck_x86_linux-mc_translate.o -MD -MP -MF .deps/memcheck_x86_linux-mc_translate.Tpo -c -o memcheck_x86_linux-mc_translate.o `test -f 'mc_translate.c' || echo './'`mc_translate.c mc_translate.c: In function ‘vgMemCheck_instrument’: mc_translate.c:4202: error: ‘struct <anonymous>’ has no member named ‘resSC’ mc_translate.c:4204: error: ‘struct <anonymous>’ has no member named ‘resSC’ mc_translate.c:4208: error: ‘struct <anonymous>’ has no member named ‘resSC’ mc_translate.c: In function ‘schemeS’: mc_translate.c:4949: error: ‘struct <anonymous>’ has no member named ‘resSC’ mc_translate.c:4950: error: ‘struct <anonymous>’ has no member named ‘resSC’ make[4]: *** [memcheck_x86_linux-mc_translate.o] Fehler 1 make[4]: Leaving directory `/home/ensonic/projects/valgrind-vcov/memcheck' make[3]: *** [all-recursive] Fehler 1 make[3]: Leaving directory `/home/ensonic/projects/valgrind-vcov/memcheck' make[2]: *** [all] Fehler 2 make[2]: Leaving directory `/home/ensonic/projects/valgrind-vcov/memcheck' make[1]: *** [all-recursive] Fehler 1 make[1]: Leaving directory `/home/ensonic/projects/valgrind-vcov' make: *** [all] Fehler 2 After skipping memcheck, I could get the rest to build, but get issues running it. Any ideas? I would be interested to figure if it can be made useable together with lcov or is vcov abandoned? Stefan /vg-in-place --tool=exp-vcov /path/to/testapp ==26250== exp-VCov, a test coverage tool. ==26250== NOTE: This is an Experimental-Class Valgrind Tool. ==26250== Copyright (C) 2002-2008, and GNU GPL'd, by Nicholas Nethercote. ==26250== Using Valgrind-3.5.0.SVN and LibVEX; rerun with -h for copyright info ==26250== exp-VCov: vc_main.c:385 (doOneInstr): the 'impossible' happened. exp-VCov: didn't find 43 in lineCCs (/usr/src/debug//////////glibc-2.10.1/setjmp/../sysdeps/i386/setjmp.S) ==26250== at 0x38002AB0: report_and_quit (m_libcassert.c:145) ==26250== by 0x38002CEC: vgPlain_assert_fail (m_libcassert.c:217) ==26250== by 0x38000B83: doOneInstr (vc_main.c:384) ==26250== by 0x38000C8F: vc_instrument (vc_main.c:435) ==26250== by 0x38095912: LibVEX_Translate (main_main.c:495) ==26250== by 0x38018CE1: vgPlain_translate (m_translate.c:1517) ==26250== by 0x3803D991: vgPlain_scheduler (scheduler.c:844) ==26250== by 0x38069414: run_a_thread_NORETURN (syswrap-linux.c:91) sched status: running_tid=1 Thread 1: status = VgTs_Runnable ==26250== at 0x466BD80: __sigsetjmp (setjmp.S:34) ==26250== by 0x46622B8: setlocale (setlocale.c:333) ==26250== by 0x8049E3A: main (bt-cmd.c:65) |
|
From: Bart V. A. <bar...@gm...> - 2009-12-27 14:29:56
|
On Sat, Dec 26, 2009 at 5:52 PM, Stefan Kost <en...@ho...>wrote: > I am quite interested in a coverage tool that does not need recompilation. > There > seems to be a patched qemu somewhere, but then I found valgrind-vcov. I had > some > issue building the branch. > That is because at the time a Valgrind branch is created, only Valgrind is branched but not VEX. If you look up the last modification date of the Valgrind code on the Valgrind branch and update the VEX source code to that date, then the build should succeed. Bart. |
|
From: Stefan K. <en...@ho...> - 2009-12-27 23:16:41
|
Am 27.12.2009 16:29, schrieb Bart Van Assche:
> On Sat, Dec 26, 2009 at 5:52 PM, Stefan Kost <en...@ho...
> <mailto:en...@ho...>> wrote:
>
> I am quite interested in a coverage tool that does not need
> recompilation. There
> seems to be a patched qemu somewhere, but then I found
> valgrind-vcov. I had some
> issue building the branch.
>
>
> That is because at the time a Valgrind branch is created, only Valgrind
> is branched but not VEX. If you look up the last modification date of
> the Valgrind code on the Valgrind branch and update the VEX source code
> to that date, then the build should succeed.
>
> Bart.
Thanks that helps, would be nice to put this to README_DEVELOPERS. This is what
I did:
> svn log -r PREV
------------------------------------------------------------------------
r10643 | njn | 2009-07-28 03:39:43 +0300 (Di, 28. Jul 2009) | 3 Zeilen
Merged all the changes from the trunk between r7367:10642, updated VCov for
various changes, and fixed a few other minor things.
------------------------------------------------------------------------
> cd VEX
> svn up -r{2009-07-28}
...
> cd ..
Unfortunately this does not magically fix the VCOV issues.
Stefan
|
|
From: Nicholas N. <n.n...@gm...> - 2010-01-04 23:21:59
|
On Mon, Dec 28, 2009 at 10:16 AM, Stefan Kost <en...@ho...> wrote: > Am 27.12.2009 16:29, schrieb Bart Van Assche: >> On Sat, Dec 26, 2009 at 5:52 PM, Stefan Kost <en...@ho... >> <mailto:en...@ho...>> wrote: >> >> I am quite interested in a coverage tool that does not need >> recompilation. There >> seems to be a patched qemu somewhere, but then I found >> valgrind-vcov. I had some >> issue building the branch. >> >> That is because at the time a Valgrind branch is created, only Valgrind >> is branched but not VEX. If you look up the last modification date of >> the Valgrind code on the Valgrind branch and update the VEX source code >> to that date, then the build should succeed. > > Thanks that helps, would be nice to put this to README_DEVELOPERS. VCov is on a branch, I haven't done anything with it for a while and probably won't any time soon. Nick |
|
From: Stefan K. <en...@ho...> - 2010-01-10 13:12:17
|
Am 05.01.2010 01:21, schrieb Nicholas Nethercote: > On Mon, Dec 28, 2009 at 10:16 AM, Stefan Kost <en...@ho...> wrote: >> Am 27.12.2009 16:29, schrieb Bart Van Assche: >>> On Sat, Dec 26, 2009 at 5:52 PM, Stefan Kost <en...@ho... >>> <mailto:en...@ho...>> wrote: >>> >>> I am quite interested in a coverage tool that does not need >>> recompilation. There >>> seems to be a patched qemu somewhere, but then I found >>> valgrind-vcov. I had some >>> issue building the branch. >>> >>> That is because at the time a Valgrind branch is created, only Valgrind >>> is branched but not VEX. If you look up the last modification date of >>> the Valgrind code on the Valgrind branch and update the VEX source code >>> to that date, then the build should succeed. >> >> Thanks that helps, would be nice to put this to README_DEVELOPERS. > > VCov is on a branch, I haven't done anything with it for a while and > probably won't any time soon. > > Nick Should I still submit my patch to bugzilla? If vcov is dead I'll use my time on hacking on bcov then (bcov.sf.net) - it an interesting alternative. Stefan |