|
Regina Henschel |
|
|
Hi all,
build breaks, I have attached the build_error.log. I have run "make build" twice to make sure, that it not a simple timing problem. I use MSVC 2008 Express with cygwin on WinXP. Kind regards Regina _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Matúš Kukan |
|
|
On 29 March 2012 22:56, Regina Henschel <[hidden email]> wrote:
> Hi all, > > build breaks, I have attached the build_error.log. I have run "make build" > twice to make sure, that it not a simple timing problem. make[1]: *** No rule to make target `/c', needed by `/cygdrive/c/git/LO36APR/workdir/wntmsci12/CxxObject/extensions/source/propctrlr/browserlistbox.o'. Stop. does not look good. I can't imagine where /c came from. Sorry, no other help then generic: make extensions.clean && make extensions of maybe: make extensions.clean && gb_FULLDEPS= make extensions There is non-zero chance it will help. Btw, what is output of: make cmd cmd="\${GNUMAKE} -v" Best, Matúš _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Regina Henschel |
|
|
Hi Matúš,
Matúš Kukan schrieb: > On 29 March 2012 22:56, Regina Henschel<[hidden email]> wrote: >> Hi all, >> >> build breaks, I have attached the build_error.log. I have run "make build" >> twice to make sure, that it not a simple timing problem. > > make[1]: *** No rule to make target `/c', needed by > `/cygdrive/c/git/LO36APR/workdir/wntmsci12/CxxObject/extensions/source/propctrlr/browserlistbox.o'. > Stop. > does not look good. I can't imagine where /c came from. > > Sorry, no other help then generic: > make extensions.clean&& make extensions > of maybe: make extensions.clean&& gb_FULLDEPS= make extensions > There is non-zero chance it will help. Will try it tomorrow (go sleeping now), if nobody else has an idea. > Btw, what is output of: > make cmd cmd="\${GNUMAKE} -v" $ /opt/lo/bin/make cmd cmd="\${GNUMAKE} -v" echo "custom cmd" && ( /opt/lo/bin/make -v ) custom cmd GNU Make 3.82 Built for i686-pc-cygwin Copyright (C) 2010 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. LibreOffice patched version Kind regards Regina _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Meeks-2 |
|
|
In reply to this post by Matúš Kukan
On Fri, 2012-03-30 at 01:11 +0200, Matúš Kukan wrote: > On 29 March 2012 22:56, Regina Henschel <[hidden email]> wrote: > > build breaks, I have attached the build_error.log. I have run "make build" > > twice to make sure, that it not a simple timing problem. > > make[1]: *** No rule to make target `/c', needed by > `/cygdrive/c/git/LO36APR/workdir/wntmsci12/CxxObject/extensions/source/propctrlr/browserlistbox.o'. > Stop. > does not look good. I can't imagine where /c came from. So - I -think- /c is a truncated /cygdrive/... probably from a generated .d dependency file. Quite possibly either the gcc generation of these, or the concatenation of them is not as robust as it could be to being aborted mid-flow (you have enough free disk-space right ? :-). Anyhow - if we can chase that down: grep -R browserlistbox.o core/workdir might give some a handful of hits, and by running 'tail' on them all you could see whether they are truncated perhaps ? if that is the case we can dig in the gnumake dependency rules and/or the deps simplifier / concatenator to see if that is the problem. Thanks ! Michael. -- [hidden email] <><, Pseudo Engineer, itinerant idiot _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Tor Lillqvist-2 |
|
|
> Quite possibly either the gcc generation of these, or the concatenation
> of them is not as robust as it could be Yeah, I used to see lots of these screwed up .d files, too. (Either oddly truncated, or looking as if several simultaneous processes had written to the same .d file.) Never understood the root cause. For some reason I have not been seeing them lately, that could be because I am now using our own make 3.82, or because I don't really build that often on Windows any more. Or then I have just been lucky. --tml _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Meeks-2 |
|
|
In reply to this post by Michael Meeks-2
Hi guys,
Regina's problem comes down to a corrupt .d file. Any ideas how that could happen ? are these generated by a pipeline ? are we failing to delete them if a compile fails mid-flow ? Then again, Regina didn't abort the build manually, so the ctrl-c-in-mid-flow case is not at issue. Thoughts, Michael. On Fri, 2012-03-30 at 11:30 +0200, Regina Henschel wrote: > workdir/wntmsci12/Dep/CxxObject/extensions/source/propctrlr/browserlistbox.d > > looks bad, I see > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/stdexcept \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/exception \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/xstddef \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/cstddef \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/eh.h \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/malloc.h \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/xstring \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/xmemory \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/new \ > /c > > workdir/wntmsci12/Dep/LinkTarget/Library/ipcr.lib.d > looks good, I see > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/ymath.h \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/cfloat \ > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/cmath > > /cygdrive/c/git/LO36APR/solver/wntmsci12/inc/offapi/com/sun/star/xsd/WhiteSpaceTreatment.hdl: > > /cygdrive/c/git/LO36APR/solver/wntmsci12/inc/offapi/com/sun/star/xsd/WhiteSpaceTreatment.hpp: > > /cygdrive/c/git/lo36apr/extensions/source/propctrlr/xsdvalidationpropertyhandler.hxx: -- [hidden email] <><, Pseudo Engineer, itinerant idiot _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Meeks-2 |
|
|
On Fri, 2012-03-30 at 10:36 +0100, Michael Meeks wrote: > Regina's problem comes down to a corrupt .d file. One more data-point, that file was exactly 4096 bytes large: > > /cygdrive/c/PROGRA~1/MICROS~1.0/VC/include/new \ > > /c [EOF] :-) Michael. -- [hidden email] <><, Pseudo Engineer, itinerant idiot _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Lubos Lunak |
|
|
In reply to this post by Michael Meeks-2
On Friday 30 of March 2012, Michael Meeks wrote:
> Hi guys, > > Regina's problem comes down to a corrupt .d file. > > Any ideas how that could happen ? are these generated by a pipeline ? > are we failing to delete them if a compile fails mid-flow ? > > Then again, Regina didn't abort the build manually, so the > ctrl-c-in-mid-flow case is not at issue. > > Thoughts, Is the problem reproducible? And if yes, does using our own make from contrib help? I don't know if it's related, but I had the stock cygwin make generating broken .d files (although IIRC the problem was generating Windows paths rather than Unix paths). -- Lubos Lunak [hidden email] _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Regina Henschel |
|
|
Hi Lubos,
Lubos Lunak schrieb: > On Friday 30 of March 2012, Michael Meeks wrote: >> Hi guys, >> >> Regina's problem comes down to a corrupt .d file. >> >> Any ideas how that could happen ? are these generated by a pipeline ? >> are we failing to delete them if a compile fails mid-flow ? >> >> Then again, Regina didn't abort the build manually, so the >> ctrl-c-in-mid-flow case is not at issue. >> >> Thoughts, > > Is the problem reproducible? And if yes, does using our own make from contrib > help? I don't know if it's related, but I had the stock cygwin make > generating broken .d files (although IIRC the problem was generating Windows > paths rather than Unix paths). > I'm using the LibreOffice patched version of GNU Make 3.82 already. Kind regards Regina _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Matúš Kukan |
|
|
In reply to this post by Michael Meeks-2
On 30 March 2012 11:36, Michael Meeks <[hidden email]> wrote:
> Regina's problem comes down to a corrupt .d file. > > Any ideas how that could happen ? are these generated by a pipeline ? > are we failing to delete them if a compile fails mid-flow ? They are generated by a pipeline: $(CXX) ... | filter-showIncludes.pl <dep-file> <object> <source>; exit ${PIPESTATUS[0]} > One more data-point, that file was exactly 4096 bytes large: What does that mean ? Best, Matúš _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Christian Lohmaier-2 |
|
|
On Fri, Mar 30, 2012 at 1:05 PM, Matúš Kukan <[hidden email]> wrote:
> On 30 March 2012 11:36, Michael Meeks <[hidden email]> wrote: >> Regina's problem comes down to a corrupt .d file. >> [...] >> One more data-point, that file was exactly 4096 bytes large: > > What does that mean ? Well - it is either a coincidence, or not.. A multiple of 1024 i.e exactly 4KB in size.. ciao Christian _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Kevin Hunter |
|
|
At 7:56am -0400 Fri, 30 Mar 2012, Christian Lohmaier wrote:
> On Fri, Mar 30, 2012 at 1:05 PM, Matúš Kukan wrote: >> On 30 March 2012 11:36, Michael Meeks wrote: >>> Regina's problem comes down to a corrupt .d file. >>> [...] >>> One more data-point, that file was exactly 4096 bytes large: >> >> What does that mean ? > > Well - it is either a coincidence, or not.. A multiple of 1024 i.e > exactly 4KB in size.. Perhaps dumb question: Is she out of space on that drive/partition? Kevin _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Norbert Thiebaud |
|
|
In reply to this post by Christian Lohmaier-2
On Fri, Mar 30, 2012 at 6:56 AM, Christian Lohmaier
<[hidden email]> wrote: > On Fri, Mar 30, 2012 at 1:05 PM, Matúš Kukan <[hidden email]> wrote: >> On 30 March 2012 11:36, Michael Meeks <[hidden email]> wrote: >>> Regina's problem comes down to a corrupt .d file. >>> [...] >>> One more data-point, that file was exactly 4096 bytes large: >> >> What does that mean ? > > Well - it is either a coincidence, or not.. A multiple of 1024 i.e > exactly 4KB in size.. and 4K is the typical memory page, and the kidn of unit you would see for partial write. the final write to device. this behavior can be seen if you have a running process that write to a log and look at the log file while it is running the log will grow by 4K increment until you ffllush() of fclose() the file. So that what one would expect to see if the filter-showincludes.pl was interrupted abruptly. Norbert _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Regina Henschel |
|
|
In reply to this post by Kevin Hunter
Hi Kervin,
Kevin Hunter schrieb: > At 7:56am -0400 Fri, 30 Mar 2012, Christian Lohmaier wrote: >> On Fri, Mar 30, 2012 at 1:05 PM, Matúš Kukan wrote: >>> On 30 March 2012 11:36, Michael Meeks wrote: >>>> Regina's problem comes down to a corrupt .d file. >>>> [...] >>>> One more data-point, that file was exactly 4096 bytes large: >>> >>> What does that mean ? >> >> Well - it is either a coincidence, or not.. A multiple of 1024 i.e >> exactly 4KB in size.. > > Perhaps dumb question: Is she out of space on that drive/partition? No, I'm not. There are about 149 GB free. Kind regards Regina _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Regina Henschel |
|
|
Should be "Hi Kevin". Please excuse.
_______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Kevin Hunter |
|
|
At 1:21pm -0400 Mon, 02 Apr 2012, Regina Henschel wrote:
> Should be "Hi Kevin". Please excuse. :-) No worries. I am often not sure whether to say "Hi" when responding back and forth in email discussions on a list. Is it a "formal" interaction? A friendly interaction? An engineer's "optimal" interaction? Regardless, "Hello." Kevin _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Meeks-2 |
|
|
On Mon, 2012-04-02 at 13:32 -0400, Kevin Hunter wrote: > :-) No worries. I am often not sure whether to say "Hi" when responding > back and forth in email discussions on a list. Is it a "formal" > interaction? A friendly interaction? An engineer's "optimal" interaction? Hopefully friendly and informal :-) ATB, Michael. -- [hidden email] <><, Pseudo Engineer, itinerant idiot _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Stahl-2 |
|
|
In reply to this post by Matúš Kukan
On 30/03/12 13:05, Matúš Kukan wrote:
> On 30 March 2012 11:36, Michael Meeks <[hidden email]> wrote: >> Regina's problem comes down to a corrupt .d file. >> >> Any ideas how that could happen ? are these generated by a pipeline ? >> are we failing to delete them if a compile fails mid-flow ? > > They are generated by a pipeline: > $(CXX) ... | filter-showIncludes.pl <dep-file> <object> <source>; exit > ${PIPESTATUS[0]} Kendy told me on IRC that he has fixed the filter-showIncludes.pl on master so that it writes to a temp file and when that is complete, moves it over the target .d dep-file, which should hopefully prevent this kind of problem in the future. i've just pushed that to libreoffice-3-5 as well. http://cgit.freedesktop.org/libreoffice/core/commit/?h=libreoffice-3-5&id=9d8662264abfad3da87e1662abda4556333889f0 _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
|
Michael Meeks-2 |
|
|
On Mon, 2012-04-02 at 22:12 +0200, Michael Stahl wrote: > Kendy told me on IRC that he has fixed the filter-showIncludes.pl on > master so that it writes to a temp file and when that is complete, moves > it over the target .d dep-file, which should hopefully prevent this kind > of problem in the future. Thanks Kendy & Michael :-) Michael. -- [hidden email] <><, Pseudo Engineer, itinerant idiot _______________________________________________ LibreOffice mailing list [hidden email] http://lists.freedesktop.org/mailman/listinfo/libreoffice |
| Powered by Nabble | Edit this page |