Hi there,
as some of you may have noticed, the biannual Boost rebuild has been underway since Saturday! So far about 100 packages have been rebuilt. I'll appreciate any help that I can get with resolving the current failures. Just ping me on IRC (_petr) so that we don't duplicate effort. Currently the following are the failures that I know about (with task numbers in case you wish to inspect):
C or C++ errors:
pcl 5663101 error: 'sqrt' is not a member of 'Eigen::internal' pokerth 5663090 gcrypt.h: No such file or directory minion 5663196 ICE zarafa 5663602 C++ errors openlierox 5663123 'LUA_GLOBALSINDEX' was not declared in this scope
Bugs apparently related to unversioned docfiles change:
CGAL 5663106 Installed (but unpackaged) file(s) found: snapper 5665599 Installed (but unpackaged) file(s) found stdair 5663142 Installed (but unpackaged) file(s) found fatrat 5672534 Installed (but unpackaged) file(s) found
-- Frankly, I'm confused by those. Fatrat and CGAL don't mention version in any way--they simply use %{_docdir} and %doc. stdair does mention version, but the build failure complains about stuff installed by %doc. What gives?
Bugs apparently related to ARM:
spring 5664257 unrecognized argument in option '-mtune=generic' ceph 5663470 ARM unsupported? (gperftools-devel) mongodb 5663581 ARM unsupported? (gperftools-devel) 0ad 5663431 ARM unsupperted? (nvidia-texture-tools or somethig)
Assorted errors:
hugin 5663415 POD document had syntax errors at /usr/bin/pod2man line 69. python-tag 5663153 urllib2.URLError: <urlopen error [Errno -2] Name or service not known> polybori 5663365 SCONS: internal error in regular expression engine
Unresolved deps:
bookkeeper 5663346 Package: xbean-3.13-2.fc20.noarch (build) Requires: eclipse-equinox-osgi airrac 5668958 depends on stdair airsched 5669057 depends on stdair sevmgr 5668947 depends on stdair simfqt 5668967 depends on stdair travelccm 5669008 depends on stdair fawkes 5669050 depends on pcl mrpt 5669093 depends on pcl iwhd 5668878 depends on mongodb condor 5668979 depends on mongodb openscad 5668923 depends on CGAL crrcsim 5669040 depends on CGAL
There's also this:
libreoffice 5668888 BOOST_NOEXCEPT
which is apparently boost-related, but I was unsuccessful in reproducing this locally, and am leaving it aside in favor of lower-hanging fruit ;)
Thanks, PM
On 2013-07-29 19:01, Petr Machata wrote:
Bugs apparently related to unversioned docfiles change:
CGAL 5663106 Installed (but unpackaged) file(s) found: snapper 5665599 Installed (but unpackaged) file(s) found stdair 5663142 Installed (but unpackaged) file(s) found fatrat 5672534 Installed (but unpackaged) file(s) found
-- Frankly, I'm confused by those. Fatrat and CGAL don't mention version in any way--they simply use %{_docdir} and %doc. stdair does mention version, but the build failure complains about stuff installed by %doc. What gives?
The root cause for all these is the same: for various reasons (upstream defaults, whatever) they install files to a versioned dir in /usr/share/doc. Previously when the special %doc used that same versioned dir, it (possibly accidentally) picked up files already in it and they ended up in packages. Special %doc is no longer versioned, so nothing picks up those files any more -> build fail.
I can look into these soon unless others beat me to it.
On 2013-07-29 19:25, Ville Skyttä wrote:
On 2013-07-29 19:01, Petr Machata wrote:
Bugs apparently related to unversioned docfiles change:
CGAL 5663106 Installed (but unpackaged) file(s) found: snapper 5665599 Installed (but unpackaged) file(s) found stdair 5663142 Installed (but unpackaged) file(s) found fatrat 5672534 Installed (but unpackaged) file(s) found
[...]
I can look into these soon unless others beat me to it.
These are now fixed in master.
On Mon, Jul 29, 2013 at 12:01 PM, Petr Machata pmachata@redhat.com wrote:
Hi there,
as some of you may have noticed, the biannual Boost rebuild has been underway since Saturday! So far about 100 packages have been rebuilt. I'll appreciate any help that I can get with resolving the current failures. Just ping me on IRC (_petr) so that we don't duplicate effort. Currently the following are the failures that I know about (with task numbers in case you wish to inspect):
C or C++ errors:
pcl 5663101 error: 'sqrt' is not a member of 'Eigen::internal'
This is most likely related to a recent update of the rawhide eigen3 package (3.1.3 -> 3.2.1) I'll handle fixing it, as there's a new upstream release of PCL I'm working on importing anyway.
Rich
Petr Machata pmachata@redhat.com writes:
as some of you may have noticed, the biannual Boost rebuild has been underway since Saturday! [...] I'll appreciate any help that I can get with resolving the current failures.
I forgot to mention that if you wish to build Boost clients, you should use the following incantation:
fedpkg build --target f20-boost
Also, I'm leaving for my vacation on Wendesday. I will likely request tag merge tomorrow (on Tuesday) midnight-ish UTC. (Hopefully with libreoffice resolved ;) )
Thanks, PM
Il 29/07/2013 18:01, Petr Machata ha scritto:
bookkeeper 5663346 Package: xbean-3.13-2.fc20.noarch (build) Requires: eclipse-equinox-osgi
sorry, i rebuilt without boost 1.54.x support... eclipse-equinox-osgi is available also in arm* repos regards
"puntogil@libero.it" puntogil@libero.it writes:
Il 29/07/2013 18:01, Petr Machata ha scritto:
bookkeeper 5663346 Package: xbean-3.13-2.fc20.noarch (build) Requires: eclipse-equinox-osgi
sorry, i rebuilt without boost 1.54.x support... eclipse-equinox-osgi is available also in arm* repos
No problem, I rebumped and rebuilt it in f20-boost.
Thanks, PM
Il 30/07/2013 02:40, Petr Machata ha scritto:
"puntogil@libero.it" puntogil@libero.it writes:
Il 29/07/2013 18:01, Petr Machata ha scritto:
bookkeeper 5663346 Package: xbean-3.13-2.fc20.noarch (build) Requires: eclipse-equinox-osgi
sorry, i rebuilt without boost 1.54.x support... eclipse-equinox-osgi is available also in arm* repos
No problem, I rebumped and rebuilt it in f20-boost.
Thanks, PM
there is also zookeeper... to be rebuild with the new boost? regards
"puntogil@libero.it" puntogil@libero.it writes:
there is also zookeeper... to be rebuild with the new boost?
Yeah, I only got around to rebuilding those that directly depend on Boost DSO's. I guess I can order builds of the rest of the dependencies today, though originally my plan was to go through DSO deps only. I'll hardly have time to go through failures though, as I'd like to merge in about 12 hours.
Thanks, PM
On Mon, 2013-07-29 at 18:01 +0200, Petr Machata wrote:
pokerth 5663090 gcrypt.h: No such file or directory
Fixed, built. http://koji.fedoraproject.org/koji/taskinfo?taskID=5675562 . Just waiting on ARM to finish.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Mon, 29 Jul 2013 18:01:12 +0200 Petr Machata pmachata@redhat.com wrote:
spring 5664257 unrecognized argument in option '-mtune=generic' ceph 5663470 ARM unsupported? (gperftools-devel) mongodb 5663581 ARM unsupported? (gperftools-devel) 0ad 5663431 ARM unsupperted? (nvidia-texture-tools or somethig)
ive added an arm to the excludearch list of spring its going to need some porting, it seems to have some issues around its dealling with cflags
gperftools is in progress to get fixed.
ill dig into 0ad
Dennis
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Mon, 29 Jul 2013 22:59:02 -0500 Dennis Gilmore dennis@ausil.us wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Mon, 29 Jul 2013 18:01:12 +0200 Petr Machata pmachata@redhat.com wrote:
spring 5664257 unrecognized argument in option '-mtune=generic' ceph 5663470 ARM unsupported? (gperftools-devel) mongodb 5663581 ARM unsupported? (gperftools-devel) 0ad 5663431 ARM unsupperted? (nvidia-texture-tools or somethig)
ive added an arm to the excludearch list of spring its going to need some porting, it seems to have some issues around its dealling with cflags
gperftools is in progress to get fixed.
ill dig into 0ad
i have fixed nvidia-texture-tools 0ad now fails to compile on i686
Dennis
On Mon, Jul 29, 2013 at 10:01 AM, Petr Machata pmachata@redhat.com wrote:
polybori 5663365 SCONS: internal error in regular expression engine
I will try to figure this one out.
2013/7/30 Jerry James loganjerry@gmail.com:
On Mon, Jul 29, 2013 at 10:01 AM, Petr Machata pmachata@redhat.com wrote:
polybori 5663365 SCONS: internal error in regular expression engine
I will try to figure this one out.
This should be 32 bit only. http://bugs.python.org/issue17998
It should be possible to workaround the problem in polybori's SConstruct, just that the trivial solution does not work, but I did not notice any problems after rebuilding python with the patch in the bugs.python.org report...
-- Jerry James http://www.jamezone.org/
-- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Paulo
On Tue, Jul 30, 2013 at 11:15 AM, Paulo César Pereira de Andrade < paulo.cesar.pereira.de.andrade@gmail.com> wrote:
2013/7/30 Jerry James loganjerry@gmail.com:
On Mon, Jul 29, 2013 at 10:01 AM, Petr Machata pmachata@redhat.com
wrote:
polybori 5663365 SCONS: internal error in regular expression engine
I will try to figure this one out.
This should be 32 bit only. http://bugs.python.org/issue17998
It should be possible to workaround the problem in polybori's SConstruct, just that the trivial solution does not work, but I did not notice any problems after rebuilding python with the patch in the bugs.python.org report...
Neither my python-fu nor my scons-fu is great enough for me to figure out how to work around this problem. Can any python masters on this list see a way to work around the bug?
This is https://bugzilla.redhat.com/show_bug.cgi?id=974257 for anybody interested.
On Wed, Jul 31, 2013 at 2:18 PM, Jerry James loganjerry@gmail.com wrote:
Neither my python-fu nor my scons-fu is great enough for me to figure out how to work around this problem. Can any python masters on this list see a way to work around the bug?
This is https://bugzilla.redhat.com/show_bug.cgi?id=974257 for anybody interested.
Well, I "fixed" it by applying the following patch. Close your eyes now, lest they explode:
--- SConstruct.orig 2013-01-29 15:37:01.000000000 -0700 +++ SConstruct 2013-08-01 15:20:08.695760012 -0600 @@ -561,16 +561,14 @@ """Creates soname."""
target = str(env.subst(target)) - import re - soPattern = re.compile('(.*).[0-9]*.[^.]*$', re.I|re.S) - soname = soPattern.findall(path.basename(target)) - + from subprocess import check_output + basename = path.basename(target) + if len(basename) == 0: + return '' + soname = check_output('sed 's/.[0-9]*.[0-9]*$//' <<< ' + basename, shell=True).rstrip() if len(soname) > 0: - return prefix + soname[0] + suffix + return prefix + soname + suffix else: - if env['PLATFORM']=="darwin": - return prefix + path.basename(target) + suffix - return ''
@@ -1103,16 +1101,14 @@ def SymlinkReadableLibname(files): """ Generate symbolik link with more readable library name."""
+ from subprocess import check_output suffix = env.subst('$SHLIBVERSIONSUFFIX') simplesuffix = env.subst('$SHLIBSUFFIX') result = [] - import re - soPattern = re.compile('(.*).[0-9]*.[0-9]*$', re.I|re.S) - sonameversion = soPattern.findall(path.basename(env.subst('$LIBRARY_VERSION')))[0]
for fname in files: fname = str(fname) - soname = soPattern.sub(r'\1', fname) + soname = check_output('sed 's/.[0-9]*.[0-9]*$//' <<< ' + fname, shell=True).rstrip() versionname = fname.replace('.' + env.subst('$LIBRARY_VERSION'), '') simple = fname.replace(suffix, simplesuffix)
But it would be much, much better if the python regular expression engine were fixed. Are any of the python maintainers seriously looking at the upstream patch? If not, could you please?
More failures from another bunch of rebuilds of about 100 packages that have API-only dependence on Boost. I fixed those overtly Boost-related, what remains seems to fail due to something else, so it should be OK to just fix these in Rawhide and ignore Boost.
Note that f20-boost will be merged soon. Right now I'm just waiting for R-bigmemory and stp rebuilds to finish--I don't see any other in-progress builds in f20-boost. Thanks to everyone who helped me move this along!
Unresolved deps:
Macaulay2 5677473 No Package found for {libfac,factory}-{devel,static} gpsdrive 5677699 No Package found for compat-libgda-devel libkni3 5677935 No Package found for texlive-utils gr-air-modes 5677713 GnuRadio Runtime required to compile gr-air-modes openoffice.org-diafilter 5678089 depends on libreoffice-core
Assorted C/C++ errors:
abiword 5677483 #include <gcrypt.h> ardour 5677496 invalid conversion from 'const LilvPort* {aka const LilvPortImpl*}' to 'LilvPort* {aka LilvPortImpl*}' bind10 5677520 error: 'inlinline' was not declared in this scope ^^^ apparent typo???
csound 5677569 error: unknown type name 'TREE' dyninst 5677602 error: redefinition of 'struct ptrace_peeksiginfo_args' valyriatear 5678319 invalid use of incomplete type 'png_info {aka struct png_info_def}' roboptim-core 5678187 typedef 'value_type' locally defined but not used ^^^ Probably safe to drop, but if the typedef is meaningful but unused, then either convert to BOOST_STATIC_ASSERT, or use __attribute__ ((unused)).
stellarium 5678268 POD document had syntax errors at /usr/bin/pod2man line 69 ^^^ The last bunch also had one like this.
fritzing 5677658 no matching function for call to 'qMax(double&, qreal)' kdenetwork 5677818 invalid conversion from 'void (*)(void*, OtrlNotifyLevel, const char*, const char*, const char*, const char*, const char*, const char*)' to 'void (*)(void*)' nodejs-mapnik 5678058 no matching function for call to 'mapnik::vector::processormapnik::vector::backend_pbf::processor(backend_type&, mapnik::Map&, double&, unsigned int&, unsigned int&, unsigned int&)' pathfinder 5678114 cannot allocate an object of abstract type 'xplc_ptr<WvBufUrlStream>::ProtectedPtr'
Apparently LUA was updated:
highlight 5677808 'lua_number2integer' was not declared in this scope monotone 5678043 'LUA_GLOBALSINDEX' was not declared in this scope widelands 5678370 error: 'LUAI_UINT32' does not name a type
Unversioned docs:
cyphesis 5677579 Installed (but unpackaged) file(s) found scribus 5678227 Installed (but unpackaged) file(s) found
Related to ARM:
yoshimi 5678415 ARM: unrecognized command line option '-msse'
Obscure errors:
oyranos 5678093 deutsche Fehler libclaw 5677926 File not found by glob: /builddir/build/BUILDROOT/libclaw-1.7.0-9.fc20.x86_64/usr/lib64/*.so.* zorba bin/zorba (or make?) Segmentation fault xmms2 5678386 {task 25618896: xsubpp XMMSClientResult.xs -> XMMSClientResult.c} ???
Thanks, PM
On Tue, Jul 30, 2013 at 5:02 PM, Petr Machata pmachata@redhat.com wrote:
fritzing 5677658 no matching function for call to 'qMax(double&, qreal)'
This popped up the last time Fritzing was built on ARM; I have a pending update for this package anyway, so I'll catch this at the same time.
2013/7/31 Petr Machata pmachata@redhat.com:
monotone 5678043 'LUA_GLOBALSINDEX' was not declared in this scope
For Lua 5.2 and newer Boost it has to be updated to latest head from MTN. As I am on vacation now, that will likely have to wait until next week.
- Thomas
devel@lists.stg.fedoraproject.org
-
Adam Williamson -
Dennis Gilmore -
Ed Marshall -
gil -
Jerry James -
Paulo César Pereira de Andrade -
Petr Machata -
Rich Mattes -
Thomas Moschny -
Ville Skyttä