Re: [boost] [regression] Tests for 'develop' not showing
On Wed, Sep 23, 2015 at 8:16 AM, Tom Kent <lists@teeks99.com> wrote:
On Tue, Sep 22, 2015 at 9:53 PM, Rene Rivera <grafikrobot@gmail.com> wrote:
On Wed, Sep 9, 2015 at 4:35 PM Edward Diener <eldiener@tropicsoft.com>
wrote:
Trying
http://www.boost.org/development/tests/develop/developer/summary.html I
get:
"Error extracting file: The specified zipfile was not found."
And they are messed up again. Tom can you look into your side of this? When I generated the results I got the ZIP files to be 1.5GiB and 2.0GiB, approximately, for develop and master, respectively. Yet what I see on the web site now is 1.5GiB and 200MiB. Which is clearly incorrect. And it's precisely the shorter ZIP, i.e. master.zip, that now doesn't work. I.e browsing develop results works, but not browsing master results.
Over the last week I've seen both the master and develop results showing the zip error message.
I just noticed that in the last master run, I had a segfault in the processing:
tomkent@eddie:~/tmp/boost_reports/boost-reports⟫ tail master.log Reading failures markup Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/BP x86_64 C++11.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a-hard.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a.xml Merging expected results Segmentation fault (core dumped)
I'll look into it more tonight. Tom
It looks like running the reports (esp master) is taking much longer...up to 2hrs. I had a cron job running every hour, which corrupted things when it was kicked off a second time while the first was still running. I've made a change in the script to setup a lock file, so that only one instance will run and submitted a pull request against the regression repo. It's already running on my report generator, so hopefully the problem won't show up again. If anyone sees it please let me know. Tom
On 24.09.2015 01:52, Tom Kent wrote:
On Wed, Sep 23, 2015 at 8:16 AM, Tom Kent <lists@teeks99.com> wrote:
On Tue, Sep 22, 2015 at 9:53 PM, Rene Rivera <grafikrobot@gmail.com> wrote:
On Wed, Sep 9, 2015 at 4:35 PM Edward Diener <eldiener@tropicsoft.com>
wrote:
> Trying > http://www.boost.org/development/tests/develop/developer/summary.html I > get: > > "Error extracting file: The specified zipfile was not found." >
And they are messed up again. Tom can you look into your side of this? When I generated the results I got the ZIP files to be 1.5GiB and 2.0GiB, approximately, for develop and master, respectively. Yet what I see on the web site now is 1.5GiB and 200MiB. Which is clearly incorrect. And it's precisely the shorter ZIP, i.e. master.zip, that now doesn't work. I.e browsing develop results works, but not browsing master results.
Over the last week I've seen both the master and develop results showing the zip error message.
I just noticed that in the last master run, I had a segfault in the processing:
tomkent@eddie:~/tmp/boost_reports/boost-reports⟫ tail master.log Reading failures markup Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/BP x86_64 C++11.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a-hard.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a.xml Merging expected results Segmentation fault (core dumped)
I'll look into it more tonight. Tom
It looks like running the reports (esp master) is taking much longer...up to 2hrs. I had a cron job running every hour, which corrupted things when it was kicked off a second time while the first was still running.
I've made a change in the script to setup a lock file, so that only one instance will run and submitted a pull request against the regression repo. It's already running on my report generator, so hopefully the problem won't show up again. If anyone sees it please let me know.
Is there any progress with this problem? The test results are not showing up for me now.
It's supposed to be fixed. But, yes, I see develop results are broken :-( On Mon, Sep 28, 2015 at 12:58 PM, Andrey Semashev <andrey.semashev@gmail.com
wrote:
On 24.09.2015 01:52, Tom Kent wrote:
On Wed, Sep 23, 2015 at 8:16 AM, Tom Kent <lists@teeks99.com> wrote:
On Tue, Sep 22, 2015 at 9:53 PM, Rene Rivera <grafikrobot@gmail.com> wrote:
On Wed, Sep 9, 2015 at 4:35 PM Edward Diener <eldiener@tropicsoft.com>
wrote: > > Trying >> >> http://www.boost.org/development/tests/develop/developer/summary.html I
> get: >> >> "Error extracting file: The specified zipfile was not found." >> >> And they are messed up again. Tom can you look into your side of this?
When I generated the results I got the ZIP files to be 1.5GiB and
2.0GiB, approximately, for develop and master, respectively. Yet what I see on the web site now is 1.5GiB and 200MiB. Which is clearly incorrect. And it's precisely the shorter ZIP, i.e. master.zip, that now doesn't work. I.e browsing develop results works, but not browsing master results.
Over the last week I've seen both the master and develop results showing the zip error message.
I just noticed that in the last master run, I had a segfault in the processing:
tomkent@eddie:~/tmp/boost_reports/boost-reports⟫ tail master.log Reading failures markup Reading
/home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/BP x86_64 C++11.xml Merging expected results Generating links pages Reading
/home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a-hard.xml Merging expected results Generating links pages Reading
/home/tomkent/tmp/boost_reports/boost-reports/master/incoming/master/processed/CrystaX.NET-apilevel-19-armeabi-v7a.xml Merging expected results Segmentation fault (core dumped)
I'll look into it more tonight. Tom
It looks like running the reports (esp master) is taking much longer...up to 2hrs. I had a cron job running every hour, which corrupted things when it was kicked off a second time while the first was still running.
I've made a change in the script to setup a lock file, so that only one instance will run and submitted a pull request against the regression repo. It's already running on my report generator, so hopefully the problem won't show up again. If anyone sees it please let me know.
Is there any progress with this problem? The test results are not showing up for me now.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
-- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail
On Mon, Sep 28, 2015 at 1:01 PM, Rene Rivera <grafikrobot@gmail.com> wrote:
It's supposed to be fixed. But, yes, I see develop results are broken :-(
So I'm getting segfaults in the report processing for develop.
Here's what I'm seeing: tomkent@eddie:/mnt/fs1/tmp/boost_reports⟫ tail boost-reports/develop.log Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-19-x86.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-21-armeabi-v7a.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-21-armeabi.xml Merging expected results Segmentation fault (core dumped) I tried deleting the whole boost-reports/develop*, but that didn't seem to have any affect. Any ideas how to track this down? Anyone know where the dumped core goes on ubuntu? Tom
On Tue, Sep 29, 2015 at 8:30 PM, Tom Kent <lists@teeks99.com> wrote:
On Mon, Sep 28, 2015 at 1:01 PM, Rene Rivera <grafikrobot@gmail.com> wrote:
It's supposed to be fixed. But, yes, I see develop results are broken :-(
So I'm getting segfaults in the report processing for develop.
Here's what I'm seeing:
tomkent@eddie:/mnt/fs1/tmp/boost_reports⟫ tail boost-reports/develop.log Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-19-x86.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-21-armeabi-v7a.xml Merging expected results Generating links pages Reading /home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-21-armeabi.xml Merging expected results Segmentation fault (core dumped)
I tried deleting the whole boost-reports/develop*, but that didn't seem to have any affect. Any ideas how to track this down? Anyone know where the dumped core goes on ubuntu?
I got it to produce the core file, here's what I was able to get out of it:
tomkent@eddie:/mnt/fs1/tmp/boost_reports⟫ gdb boost-reports/boost_regression/stage/bin/boost_report boost-reports/develop/core GNU gdb (Ubuntu 7.7.1-0ubuntu5~14.04.2) 7.7.1 Copyright (C) 2014 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. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-linux-gnu". Type "show configuration" for configuration details. For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from boost-reports/boost_regression/stage/bin/boost_report...(no debugging symbols found)...done. [New LWP 23489] Core was generated by `/mnt/fs1/tmp/boost_reports/boost-reports/boost_regression/stage/bin/boost_repor'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00000000004472b5 in boost::re_detail::perl_matcher<char const*, std::allocator<boost::sub_match<char const*> >, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::find_restart_any() () Traceback (most recent call last): File "/usr/share/gdb/auto-load/usr/lib/x86_64-linux-gnu/libstdc++. so.6.0.19-gdb.py", line 63, in <module> from libstdcxx.v6.printers import register_libstdcxx_printers ImportError: No module named 'libstdcxx' (gdb) quit
Tom Kent wrote:
On Tue, Sep 29, 2015 at 8:30 PM, Tom Kent<lists@teeks99.com> wrote: <snip>
/home/tomkent/tmp/boost_reports/boost-reports/develop/incoming/develop/processed/CrystaX.NET-apilevel-21-armeabi.xml Merging expected results Segmentation fault (core dumped)
I tried deleting the whole boost-reports/develop*, but that didn't seem to have any affect. Any ideas how to track this down? Anyone know where the dumped core goes on ubuntu?
I got it to produce the core file, here's what I was able to get out of it: tomkent@eddie:/mnt/fs1/tmp/boost_reports⟫ gdb boost-reports/boost_regression/stage/bin/boost_report boost-reports/develop/core GNU gdb (Ubuntu 7.7.1-0ubuntu5~14.04.2) 7.7.1 Copyright (C) 2014 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html
In "normal" circumstances an XML file containing the results has around 10 to 50 MB top, this file has 3 GB. <snip>
Core was generated by `/mnt/fs1/tmp/boost_reports/boost-reports/boost_regression/stage/bin/boost_repor'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00000000004472b5 in boost::re_detail::perl_matcher<char const*, std::allocator<boost::sub_match<char const*> >, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::find_restart_any() () Traceback (most recent call last): File "/usr/share/gdb/auto-load/usr/lib/x86_64-linux-gnu/libstdc++. so.6.0.19-gdb.py", line 63, in <module> from libstdcxx.v6.printers import register_libstdcxx_printers ImportError: No module named 'libstdcxx' (gdb) quit
Regarding the above segfault. AFAIR the only place where regex matching is used is here: https://github.com/boostorg/regression/blob/develop/reports/src/add_expected... calling: https://github.com/boostorg/regression/blob/develop/reports/src/add_expected... I've added it some time ago in order to find text phrases indicating a type of an error. This could be done without Regex so we could drop this dependency but I don't know it this would fix the problem. The whole call stack would be helpful. For now I've tested the report tool on a different platform (win7, msvc-10) so my results may be not very reliable. In x84 it throws bad_alloc here: https://github.com/boostorg/regression/blob/develop/reports/src/xml.cpp#L352 where the whole file is copied into an in-memory storage (vector). In x64 it correctly generates the files for me, but it takes a few GB of memory in the process. Regards, Adam
On Wed, Sep 30, 2015 at 1:00 PM, Adam Wulkiewicz <adam.wulkiewicz@gmail.com> wrote:
The whole call stack would be helpful.
tomkent@eddie:/mnt/fs1/tmp/boost_reports⟫ gdb boost-reports/boost_regression/stage/bin/boost_report boost-reports/dep/core GNU gdb (Ubuntu 7.7.1-0ubuntu5~14.04.2) 7.7.1 Copyright (C) 2014 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. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-linux-gnu". Type "show configuration" for configuration details. For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>. Find the GDB manual and other documentation resources online at: <http://www.gnu.org/software/gdb/documentation/>. For help, type "help". Type "apropos word" to search for commands related to "word"... Reading symbols from boost-reports/boost_regression/stage/bin/boost_report...(no debugging symbols found)...done. warning: exec file is newer than core file. [New LWP 23489] Core was generated by `/mnt/fs1/tmp/boost_reports/boost-reports/boost_regression/stage/bin/boost_repor'. Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00000000004472b5 in boost::re_detail::perl_matcher<char const*, std::allocator<boost::sub_match<char const*> >, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::find_restart_any() () Traceback (most recent call last): File "/usr/share/gdb/auto-load/usr/lib/x86_64-linux-gnu/libstdc++. so.6.0.19-gdb.py", line 63, in <module> from libstdcxx.v6.printers import register_libstdcxx_printers ImportError: No module named 'libstdcxx' (gdb) bt full #0 0x00000000004472b5 in boost::re_detail::perl_matcher<char const*, std::allocator<boost::sub_match<char const*> >, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::find_restart_any() () No symbol table info available. #1 0x0000000000445d6d in boost::re_detail::perl_matcher<char const*, std::allocator<boost::sub_match<char const*> >, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::find_imp() () No symbol table info available. #2 0x0000000000446318 in bool find_regex<boost::basic_string_ref<char, std::char_traits<char> > >(boost::basic_string_ref<char, std::char_traits<char> > const&, char const*) () No symbol table info available. #3 0x0000000000441fb1 in process_test_log(boost::regression::test_structure_t::test_log_t&, boost::regression::failures_markup_t const&, boost::regression::expected_results_t const&, std::string const&) () No symbol table info available. #4 0x0000000000442778 in boost::regression::add_expected_results(boost::regression::test_structure_t::run_t&, boost::regression::failures_markup_t const&, boost::regression::expected_results_t const&, std::string const&) () No symbol table info available. #5 0x00000000004084e3 in main () No symbol table info available.
participants (4)
-
Adam Wulkiewicz -
Andrey Semashev -
Rene Rivera -
Tom Kent