Dominique Pellé <dominique.pe...@gmail.com> wrote: > > Christian Brabandt <cbli...@256bit.org> wrote: > > > >> On Mo, 30 Okt 2017, James McCoy wrote: > >> > >>> I recently uploaded 8.0.1226 to Debian and nearly every build failed[0]. > >>> The most common failure is > >>> > >>> Found errors in Test_popup_and_window_resize(): > >>> function RunTheTest[34]..Test_popup_and_window_resize line 16: > >>> Pattern '^!\\s*$' does not match '' > >>> function RunTheTest[34]..Test_popup_and_window_resize line 26: > >>> Pattern '^!\\s*$' does not match '~ > >>> ' > >>> function RunTheTest[34]..Test_popup_and_window_resize line 28: > >>> Pattern '^!\\s*$' does not match '' > >>> > >>> I investigated some and that test appears to be extremely flaky in those > >>> environments. I haven't had a failure yet locally (amd64 system), but > >>> those architectures fail almost every run. > >>> > >>> Attached is a log from that test and the associated diff showing where I > >>> added ch_log() calls. > >>> > >>> [0]: > >>> https://buildd.debian.org/status/logs.php?pkg=vim&ver=2%3A8.0.1226-1&suite=sid > >> > >> Is it possible to run those tests interactively on those architectures? > >> Or does it succeed when run interactively and only fails when run in the > >> test suite? > >> > >> > >> Christian > > > > I've seen the same test failures on my Linux machine a few days ago. > > It happened several times in a row. > > Somehow it's not happening anymore. Note sure why. > > I'll try to reproduce it. > > > > Dominique > > I can now reproduce the test failure 100% of the times > when Vim-8.0.1238 is configured with: > > $ CC=gcc-7 ./configure --with-features=huge --enable-gui=none > > and when building with asan i.e. with this change: > > diff --git a/src/Makefile b/src/Makefile > index 0b95c6e..50c0fff 100644 > --- a/src/Makefile > +++ b/src/Makefile > @@ -681,7 +681,7 @@ LINT_OPTIONS = -beprxzF > # address sanitizer or with the undefined sanitizer. Works with gcc and > # clang. May make Vim twice as slow. Errors reported on stderr. > # More at: https://code.google.com/p/address-sanitizer/ > -#SANITIZER_CFLAGS = -g -O0 -fsanitize=address -fno-omit-frame-pointer > +SANITIZER_CFLAGS = -g -O0 -fsanitize=address -fno-omit-frame-pointer > #SANITIZER_CFLAGS = -g -O0 -fsanitize=undefined -fno-omit-frame-pointer > SANITIZER_LIBS = $(SANITIZER_CFLAGS) > > Here is how it fails: > > $ cd vim/src > $ make test_popup > ... > Executed 20 tests > 1 FAILED: > Found errors in Test_popup_and_window_resize(): > function RunTheTest[34]..Test_popup_and_window_resize line 16: Pattern > '^!\\s*$' does not match '' > function RunTheTest[34]..Test_popup_and_window_resize line 26: Pattern > '^!\\s*$' does not match '~ > ' > function RunTheTest[34]..Test_popup_and_window_resize line 28: Pattern > '^!\\s*$' does not match '' > > > Increasing the time to wait like this fixes it for me: > > diff --git a/src/testdir/test_popup.vim b/src/testdir/test_popup.vim > index 1be4e6f..cedcd26 100644 > --- a/src/testdir/test_popup.vim > +++ b/src/testdir/test_popup.vim > @@ -639,7 +639,7 @@ func Test_popup_and_window_resize() > endif > let g:buf = term_start([GetVimProg(), '--clean', '-c', 'set > noswapfile'], {'term_rows': h / 3}) > call term_sendkeys(g:buf, (h / 3 - 1)."o\<esc>") > - call term_wait(g:buf, 200) > + call term_wait(g:buf, 400) > call term_sendkeys(g:buf, "Gi\<c-x>") > call term_sendkeys(g:buf, "\<c-v>") > call term_wait(g:buf, 100)
Thanks for digging in to this. Time-dependent tests are bad, but in this case it's hard to do otherwise. I'll make it 500 msec to give it a bit more room. I was also thinking of making the vim instance that's started extra slow, e.g. by running it under valgrind, but didn't get to that yet. I assume ASAN also makes it a bit slower (but less than valgrind). -- Shaw's Principle: Build a system that even a fool can use, and only a fool will want to use it. /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net \\\ /// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\ \\\ an exciting new programming language -- http://www.Zimbu.org /// \\\ help me help AIDS victims -- http://ICCF-Holland.org /// -- -- You received this message from the "vim_dev" maillist. Do not top-post! Type your reply below the text you are replying to. For more information, visit http://www.vim.org/maillist.php --- You received this message because you are subscribed to the Google Groups "vim_dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to vim_dev+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
