Bug 118372
| Summary: | devel/STLport 5.1.4 port. | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Václav Haisman <v.haisman> | ||||||||||||||||||
| Component: | Individual Port(s) | Assignee: | Thierry Thomas <thierry> | ||||||||||||||||||
| Status: | Closed FIXED | ||||||||||||||||||||
| Severity: | Affects Only Me | ||||||||||||||||||||
| Priority: | Normal | ||||||||||||||||||||
| Version: | Latest | ||||||||||||||||||||
| Hardware: | Any | ||||||||||||||||||||
| OS: | Any | ||||||||||||||||||||
| Attachments: |
|
||||||||||||||||||||
|
Description
Václav Haisman
2007-12-01 15:40:01 UTC
This is a multi-part message in MIME format. Hello, I have been working on such an upgrade, with the following patch: <http://people.freebsd.org/~thierry/ports/stlport.diff> It includes a regression-test: target, which is running OK on i386, but is eating all the available memory on amd64. Did you run these tests? Regards, -- Th. Thomas. Thierry Thomas wrote, On 2.12.2007 21:33:
> Hello,
>
> I have been working on such an upgrade, with the following patch:
>
> <http://people.freebsd.org/~thierry/ports/stlport.diff>
>
> It includes a regression-test: target, which is running OK on i386, but
> is eating all the available memory on amd64. Did you run these tests?
>
> Regards,
Hi,
I did run the tests but I have only i386 box to test with and I have not
noticed the behaviour you describe.
--
VH
On Sun 2 dec 07 at 22:32:43 +0100, Václav Haisman <V.Haisman@sh.cvut.cz> wrote: > Hi, > I did run the tests but I have only i386 box to test with and I have not > noticed the behaviour you describe. Adding a regression-test: target to the port will let the tests run on the packaging cluster for every platform. Regards, -- Th. Thomas.
Thierry Thomas wrote, On 2.12.2007 23:35:
> On Sun 2 dec 07 at 22:32:43 +0100, Václav Haisman <V.Haisman@sh.cvut.cz>
> wrote:
>
>> Hi,
>> I did run the tests but I have only i386 box to test with and I have not
>> noticed the behaviour you describe.
>
> Adding a regression-test: target to the port will let the tests run on
> the packaging cluster for every platform.
>
> Regards,
I have merged regression-test and some other parts of your patch with mine.
Attached is the result.
--
VH
Hi, since the updated STLport would not work on AMD64 and there does not seem to be anybody with the platform interested enough who could fix it, what about create a new port devel/stlport5 instead and marking it broken on that platform? Would this be feasible? -- VH On Wed 12 dec 07 at 8:50:25 +0100, Václav Haisman <V.Haisman@sh.cvut.cz> wrote: > Hi, Hello, > since the updated STLport would not work on AMD64 and there does not seem to > be anybody with the platform interested enough who could fix it, what about > create a new port devel/stlport5 instead and marking it broken on that > platform? Would this be feasible? Actually I have a problem with it even on i386: it breaks cad/opencascade. It needs some more investigations... I'm not sure that a repocopy to devel/stlport5 would be a good solution, because our STLport-4.6.2 is obsolete and unmaintained. Best regards, -- Th. Thomas.
Thierry Thomas wrote, On 12.12.2007 21:04:
> On Wed 12 dec 07 at 8:50:25 +0100, Václav Haisman <V.Haisman@sh.cvut.cz>
> wrote:
>> Hi,
>
> Hello,
>
>> since the updated STLport would not work on AMD64 and there does not seem to
>> be anybody with the platform interested enough who could fix it, what about
>> create a new port devel/stlport5 instead and marking it broken on that
>> platform? Would this be feasible?
>
> Actually I have a problem with it even on i386: it breaks
> cad/opencascade.
>
> It needs some more investigations... I'm not sure that a repocopy to
> devel/stlport5 would be a good solution, because our STLport-4.6.2 is
> obsolete and unmaintained.
>
> Best regards,
Hi,
I investigated the Opencascade build problem and here are some patches that
get me rather far during its build. Unfortunately the build still fails, now
with a linker error. I have investigated it, too, and it seems that at some
point of compilation the _PTHREAD symbol is not defined (probably missing
-pthread on command line for GCC) and thus STLport headers use/define
stlpmtx_std namespace (STLport as not thread safe) instead of stlp_std. This
later shows as the link error because the STLport port is compiled as thread
safe.
I probably will try to fix this issue too but I am absolutely unfamiliar with
the Opencascade build process, so I would appreciate if you could look at the
issue, too.
--
VH
On Sat 8 dec 07 at 12:13:30 +0100, Vaclav Haisman <V.Haisman@sh.cvut.cz> wrote: > I have merged regression-test and some other parts of your patch with mine. > Attached is the result. Just a minor remark: the line ${TAR} -c -C "${PATCH_WRKSRC}/lib" -f - . | ${TAR} -xv -C "${PREFIX}/lib" -f - set the ownership of ${PREFIX}/lib to the owner of the build. Then you get the message: ===> Running ldconfig /sbin/ldconfig -m /usr/local/lib ldconfig: /usr/local/lib: ignoring directory not owned by root I propose the attched patch to fix this problem. (Still working on OpenCascade...) Best regards, -- Th. Thomas. On Sat 15 dec 07 at 19:58:47 +0100, Thierry Thomas <thierry@FreeBSD.org> wrote: > Many thanks for these patches! > > I have already prepared some patches, and I'm going to merge your > patches with mine, in order to solve this problem. Well, finally I have a set of patches allowing to build cad/opencascade with STLport 5.1.4 and on FreeBSD-7. I have to add "-D_REENTRANT -D__unix" to build it against STLport 5.1.4; since I guess that the same problem could occur with other programs, I propose to add the following patch: --- patch-stlport_stl_config_features.h begins here --- --- ./stlport/stl/config/features.h.orig 2007-08-18 22:45:42.000000000 +0200 +++ ./stlport/stl/config/features.h 2007-12-23 11:18:22.000000000 +0100 @@ -218,7 +218,7 @@ #endif /* Operating system recognition (basic) */ -#if defined (__unix) || defined (__linux__) || defined (__QNX__) || defined (_AIX) || defined (__NetBSD__) || defined(__OpenBSD__) || defined (__Lynx__) +#if defined (__unix) || defined (__linux__) || defined (__QNX__) || defined (_AIX) || defined (__NetBSD__) || defined(__OpenBSD__) || defined (__Lynx__) || defined(__FreeBSD__) # define _STLP_UNIX 1 #elif defined(macintosh) || defined (_MAC) # define _STLP_MAC 1 @@ -305,6 +305,7 @@ #endif #if (defined (_REENTRANT) || defined (_THREAD_SAFE) || \ + (defined (_POSIX_THREADS) && defined (__FreeBSD__)) && \ (defined (_POSIX_THREADS) && defined (__OpenBSD__))) && \ !defined (_STLP_THREADS) # define _STLP_THREADS --- patch-stlport_stl_config_features.h ends here --- Best regards, -- Th. Thomas. thierry 2007-12-23 17:00:37 UTC
FreeBSD ports repository
Modified files:
devel/stlport Makefile distinfo pkg-descr pkg-plist
Added files:
devel/stlport/files
extra-patch-no-short-string-optim-user_config.h.diff
extra-patch-ptr_spec-user_config.h.diff
extra-patch-templ-expr-user_config.h.diff
patch-stlport_stl_config_features.h
patch-stlport_stl_config_host.h
Removed files:
devel/stlport/files patch-src::common_macros.mak
patch-src::gcc-freebsd.mak
patch-stlport::c_locale.h
patch-stlport::config::stl_gcc.h
patch-stlport::stdexcept
patch-stlport::stl::_bvector.h
patch-stlport::stl::_cwchar.h
patch-stlport::stlcomp.h
patch-test::eh::gcc-freebsd.mak
src::icc-freebsd.mak stlport::stl_icc.h
test::eh::icc-freebsd.mak
Log:
Upgrade to 5.1.4 and unbreak.
Changelog at
<https://sourceforge.net/project/shownotes.php?group_id=146814&release_id=544939>
PR: ports/118372
Submitted by: Vaclav Haisman <v.haisman (at) sh.cvut.cz>
Revision Changes Path
1.43 +49 -55 ports/devel/stlport/Makefile
1.9 +3 -3 ports/devel/stlport/distinfo
1.1 +14 -0 ports/devel/stlport/files/extra-patch-no-short-string-optim-user_config.h.diff (new)
1.1 +14 -0 ports/devel/stlport/files/extra-patch-ptr_spec-user_config.h.diff (new)
1.1 +14 -0 ports/devel/stlport/files/extra-patch-templ-expr-user_config.h.diff (new)
1.3 +0 -11 ports/devel/stlport/files/patch-src::common_macros.mak (dead)
1.6 +0 -33 ports/devel/stlport/files/patch-src::gcc-freebsd.mak (dead)
1.3 +0 -11 ports/devel/stlport/files/patch-stlport::c_locale.h (dead)
1.6 +0 -43 ports/devel/stlport/files/patch-stlport::config::stl_gcc.h (dead)
1.3 +0 -14 ports/devel/stlport/files/patch-stlport::stdexcept (dead)
1.3 +0 -12 ports/devel/stlport/files/patch-stlport::stl::_bvector.h (dead)
1.2 +0 -11 ports/devel/stlport/files/patch-stlport::stl::_cwchar.h (dead)
1.3 +0 -12 ports/devel/stlport/files/patch-stlport::stlcomp.h (dead)
1.1 +19 -0 ports/devel/stlport/files/patch-stlport_stl_config_features.h (new)
1.1 +14 -0 ports/devel/stlport/files/patch-stlport_stl_config_host.h (new)
1.6 +0 -27 ports/devel/stlport/files/patch-test::eh::gcc-freebsd.mak (dead)
1.4 +0 -57 ports/devel/stlport/files/src::icc-freebsd.mak (dead)
1.3 +0 -38 ports/devel/stlport/files/stlport::stl_icc.h (dead)
1.4 +0 -76 ports/devel/stlport/files/test::eh::icc-freebsd.mak (dead)
1.5 +5 -12 ports/devel/stlport/pkg-descr
1.8 +307 -352 ports/devel/stlport/pkg-plist
_______________________________________________
cvs-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/cvs-all
To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"
thierry 2007-12-23 17:06:55 UTC
FreeBSD doc repository (ports committer)
Modified files:
en_US.ISO8859-1/articles/contributors contrib.additional.sgml
Log:
Add an entry for Vaclav Haisman.
PR: ports/118372
Revision Changes Path
1.746 +5 -0 doc/en_US.ISO8859-1/articles/contributors/contrib.additional.sgml
_______________________________________________
cvs-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/cvs-all
To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"
State Changed From-To: open->feedback I have uploaded a recapitulative patch at <http://people.freebsd.org/~thierry/ports/stlport.diff>. Do you agree with it? Responsible Changed From-To: freebsd-ports-bugs->thierry Take it. State Changed From-To: feedback->closed Committed, thanks! Note: don't forget to submit the patches upstream. |
