Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Updates: Status: Fixed Labels: fixed_2_15_33 Comment #38 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 fixed. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Updates: Status: Verified Comment #39 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 (No comment was entered for this change.) ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #37 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Any idea whether the fix for issue 2356 causes a difference here? ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #35 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Actually, since the segfault is not being all that helpful right now: What are the warnings in the versions where they occur? It might be easier to rein in an area for proofreading based on them if they happen to stem from the same problem. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #36 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 They are not relevant I believe (possibly some articulate bugs). http://thread.gmane.org/gmane.comp.gnu.lilypond.bugs/33663/focus=33693 And as you've seen already in comment 20, they do not appear in stripped version of the file. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Issue 2335 in lilypond: 64 bit LilyPond segfaults
Status: Accepted Owner: Labels: Type-Crash OpSys-Linux New issue 2335 by philehol...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Reported by Przemysław 'Przemoc' Pawełczyk With the attached LilyPond source, lilypond downloaded 64 bit binaries segfault. There are lots of ways of stopping this, but it remains a bug. Lots more information here: http://old.nabble.com/Abrupt-end-of-interpreting-music-%28segmentation-fault%29-td33347885.html Attachments: 2335.ly 2.2 KB ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Updates: Labels: -Type-Crash Type-Critical Comment #1 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 according to his information about the history, this did not crash in 2.14.2. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #2 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 woah, that ly is huge! I would have rejected it for not having a Tiny example. Could you reduce that to a Tiny example? I particular want to know about articulate and resetRelativeOctave. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #3 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Can you crash it in the debugger? If that does not work, can you switch on core dumps (ulimit -c unlimited) and then use the debugger to give a backtrace? ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #4 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 I would simplify it if I knew how to do it preserving the segfault. :) Summary of tests performed using 64-bit LilyPonds in Debian Wheezy x64 (as it should be here too for convenience). GNU LilyPond 2.14.2 - warnings GNU LilyPond 2.15.0 - warnings GNU LilyPond 2.15.1 - warnings GNU LilyPond 2.15.2 - warnings GNU LilyPond 2.15.3 - warnings GNU LilyPond 2.15.4 - warnings GNU LilyPond 2.15.5 - warnings GNU LilyPond 2.15.6 - warnings GNU LilyPond 2.15.7 - warnings GNU LilyPond 2.15.8 - warnings GNU LilyPond 2.15.9 - warnings GNU LilyPond 2.15.10 - warnings GNU LilyPond 2.15.11 - ok GNU LilyPond 2.15.12 - ok GNU LilyPond 2.15.13 - ok GNU LilyPond 2.15.14 - ok GNU LilyPond 2.15.15 - ok GNU LilyPond 2.15.16 - ok GNU LilyPond 2.15.17 - ok GNU LilyPond 2.15.18 - ok GNU LilyPond 2.15.19 - ok GNU LilyPond 2.15.20 - ok GNU LilyPond 2.15.21 - ok GNU LilyPond 2.15.22 - ok GNU LilyPond 2.15.23 - segfault GNU LilyPond 2.15.24 - segfault GNU LilyPond 2.15.25 - segfault GNU LilyPond 2.15.26 - segfault GNU LilyPond 2.15.27 - segfault GNU LilyPond 2.15.28 - ok GNU LilyPond 2.15.29 - ok GNU LilyPond 2.15.30 - segfault Actually removing resetRelativeOctave was tested and it still segfaulted. Removing articulate was too much though and there was no segfault then. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #5 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 David, my first mail had result from gdb, but it's useless: Program received signal SIGSEGV, Segmentation fault. 0x00618a31 in ?? () I provided ltrace dump: http://paste.przemoc.net/lilypond/abrupt-end-of-interpreting-music/lilypond-2.15.30.ltrace-last-10334-lines.bz2 ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #6 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 My bad, there is some backtrace available after all: #0 0x00618a31 in ?? () #1 0x0049321d in ?? () #2 0x0048b977 in ?? () #3 0x0066d7aa in ?? () #4 0x006d5d99 in ?? () #5 0x004b2245 in ?? () #6 0x004b2204 in ?? () #7 0x00619789 in ?? () #8 0x0049fe6a in ?? () #9 0x00492d0d in ?? () #10 0x004df844 in ?? () #11 0x004dd664 in ?? () #12 0x004de44e in ?? () #13 0x0061f974 in ?? () #14 0x0045f66e in ?? () #15 0x00460af8 in ?? () #16 0x0045e5fd in ?? () #17 0x7792e85c in scm_dapply () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #18 0x77931e1e in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #19 0x005d1c3e in ?? () #20 0x006fd305 in ?? () #21 0x0070ad67 in ?? () #22 0x0052a247 in ?? () #23 0x00524d24 in ?? () #24 0x77930d69 in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #25 0x7792ebd0 in scm_dapply () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #26 0x779838c9 in scm_c_catch () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #27 0x779839a7 in scm_catch_with_pre_unwind_handler () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #28 0x7792e85c in scm_dapply () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #29 0x77931e1e in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #30 0x77930a17 in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #31 0x7792ebd0 in scm_dapply () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #32 0x7336f251 in scm_srfi1_for_each () from /usr/lib/libguile-srfi-srfi-1-v-3.so #33 0x77931d21 in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #34 0x77930a17 in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #35 0x779325c8 in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #36 0x7792ebd0 in scm_dapply () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #37 0x00542a12 in ?? () #38 0x77948d1f in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #39 0x77920aea in ?? () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #40 0x779838c9 in scm_c_catch () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #41 0x77920e66 in scm_i_with_continuation_barrier () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #42 0x77920f00 in scm_c_with_continuation_barrier () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #43 0x77982af3 in scm_i_with_guile_and_parent () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #44 0x77948ce5 in scm_boot_guile () from /home/przemoc/lilypond/2.15.30/usr/bin/../lib/libguile.so.17 #45 0x00540af4 in ?? () #46 0x75b09ead in __libc_start_main (main=optimized out, argc=optimized out, ubp_av=optimized out, init=optimized out, fini=optimized out, rtld_fini=optimized out, stack_end=0x7fffe468) at libc-start.c:228 #47 0x004098aa in ?? () #48 0x7fffe468 in ?? () #49 0x001c in ?? () #50 0x0002 in ?? () #51 0x7fffe6f1 in ?? () #52 0x7fffe721 in ?? () #53 0x in ?? () ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #7 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 That's basically useless: looks like you have binaries without debug symbols. Do have a way to compile the code yourself? If that were possible, one could at least try to figure out the responsible commit. If we look at git log --oneline release/2.15.22-1..release/2.15.23-1 and cut out everything improbable (according to my personal metrics) we get to 7ce94ab Prevents cross-staff Stems from colliding with articulations. 968e472 Looser spacing for ragged-last-bottom; issue 1377 ff72533 hairpin.cc: transition between spanners aligns on note; issue 1216 99ef1a8 Updates version string in span-bar-articulation. f417a60 Implements padding at right-broken hairpins that come up against span ba 5543eb5 Prevents scripts in cross-staff contexts from colliding with span bars. 3bcfd69 Prevents dynamics from colliding with cross-staff stems. The first one would likely be related to articulate.ly, but then articulate.ly is good for a lot of spacing changes. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #8 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 As stated in the mails and in this issue description, problem regards Linux 64-bit builds linked from LilyPond site. My own build doesn't segfault. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #9 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 @comment 4: guidelines for making a Tiny example are here: http://lilypond.org/tiny-examples.html if you can't be bothered to follow them, then I can't be bothered to look at this bug report. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #10 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 I understand you, but it may be a bug that just doesn't manifest when the code is short. I'll try to simplify it as much as I can preserving the segfault. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #11 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Here is possibly the limit of my squeezing. Removing _any_ non empty line here makes segfault disappear. Heck, change first 'dis' to 'd' and it works. Magic! Attachments: strange-bug-strip-2.ly 1.8 KB ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #12 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 ... seriously? You absolutely *must* have the \time signature? Remove a single tie ~ will make it work? Removing the left or right hand will make it work. Hmm. If you're correct, then this is valuable information; it suggests there's a weird bit of memory corruption which only occurs in very rare circumstances. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #13 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Any differences in behavior or output with -ddebug-gc ? ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #14 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 I missed ties and slurs, they can be removed, but beside that, yes, almost anything (I may missed something) changed here make segfault gone (and definitely removing whole lines works like that). Like changing \minor to \major (but changing also a to c make it segfaults again). I also thing it is some extremely hard to catch (apparently somewhat layout dependent) memory-related bug, but if it happens only in provided builds, maybe building environment has some tools that spits wrong output? (like old gcc?). Attachments: strange-bug-strip-3.ly 1.7 KB ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #16 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 As it happens extremely rare, only on Linux 64-bit and only on provided builds, maybe it doesn't need to be marked as critical? I mean, it could needlessly delay LilyPond 2.16.0 release and I definitely wouldn't want to be the one guilty of that (by finding this peculiar bug in provided LilyPond builds). If custom builds were corrupted too, then it would be a dead serious thing indeed. But they are not. So it is possibly only a building-process-related problem. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #17 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 I am afraid that particularly with a stable release, people tend to use the provided binaries. I hate this cross compilation multiplatform business. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #18 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 It segfaults on my machine too, so don't feel that it's you personally guilty (unless you were lying about it working on 2.14.2). I'm not going to check that version. Input that used to work no longer works. We did not deliberately add a segfault. Therefore it is a Critical bug, and the release candidate is cancelled. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #19 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond
segfaults
http://code.google.com/p/lilypond/issues/detail?id=2335
holy mao. I owe you an apology, przemoc.
I have verified that commenting out useless lines will eliminate the
segfault. strange-bug-strip-3.ly segfaults. If I comment out any ONE of:
global = {
%\key c \major
%\time 4/4
%\tempo 4 = 132
}
rightHand = \relative c' {
%\clef treble
then it works. I would *never* believe anybody who told me this was the
case, and even after seeing it myself a few times, I scarcely believe it.
___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #20 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 I never thought I would say it, but I regret that I was not lying (and the input works flawlessly in 2.14.2, obviously). $ /home/przemoc/lilypond/2.14.02/usr/bin/lilypond ./strange-bug-strip-3.ly GNU LilyPond 2.14.2 Processing `./strange-bug-strip-3.ly' Parsing... Interpreting music... [8][16] Preprocessing graphical objects... Interpreting music... MIDI output to `strange-bug-strip-3.midi'... Finding the ideal number of pages... Fitting music on 1 or 2 pages... Drawing systems... Layout output to `strange-bug-strip-3.ps'... Converting to `./strange-bug-strip-3.pdf'... success: Compilation successfully completed ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #21 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Even useless lines produce music events. As long as it is not purely in the grob juggling part that the problem is occuring, that is not all that spectacular. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #22 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Sorry for off-topic comment, but I feel I have to add this. This piece is a fragment of my WIP work (well, lately not in progress at all) that I named early Tempt fate, even before I hit this bug. Quoting Linus: the perversity of the universe is endless. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #23 on issue 2335 by mts...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 The articulate script has tons of (cool) hacks, many of which may be doing funny things w/ smobs and memory management. I'd recommend going into the script and commenting out stuff until the problem disappears. Then, let us know what, when commented out, makes this go away. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #24 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 From Scheme, you can't really do much of hacking smobs and memory management. It is likely that anything you can do to make the problem disappear is more related to garbage collection running at a different time rather than a direct consequence of the outcommented stuff. Does not harm to try, but I would not assign all too much importance to the results. Except that it is conceivable that some code in LilyPond might have been written on fragile assumptions regarding when objects will no longer be accessed. But I would not hold my breath for that. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #25 on issue 2335 by mts...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 This is what I meant. I can make LilyPond segfault w/ many a Scheme hack. Any function that uses unsmob_spanner, for example, will likely segfault if a grob is passed into it. There's tons of code that avoids segfaults via assumptions, so I think that the commenting out exercise will help expose where that code may lie. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #26 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Mike: the bug reporter claimed that he had a tiny example. I called him on it; he produced another two versions. His final version still looks large-ish for a segfault. BUT I HAVE CHECKED THIS. It is sitting in my ~/tmp right now, and after 15 minutes the only thing I have discovered is that I can comment out the \set PianoStaff.midiInstrument in pianoStaffMidi and still keep the segfault. Comment out other \set commands? no segfault. Comment out some music? no segfault. Comment out the \time 4/4 ? no segfault. As I said, I would never believe anybody who told me this was happening, but I've seen it myself. As you know, I'm an absolute bugger when it comes to proving people wrong, but the *only* think I have found that I can comment out is that one \set command. This bug really is as strange as he suggests. I have 8 years of calling users idiots over lilypond, but I can't do that here. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #27 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 It's likely that if you crank some knob concerning garbage collection that things will be totally different. Unfortunately I see no such knob. Try putting some #(gc) in the music before the segfault occurs. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #28 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Beside the line mentioned by Graham (damn, I overlooked it), \unfoldRepeats can be removed too. David, the problem is that changing this file in almost any way makes segfault vanish. Not only deleting lines. Add r1 at the beginning of rightHand - it works! (adding in leftHand, doesn't thought). Add #(gc) - it works! And so on... Attachments: strange-bug-strip-4.ly 1.6 KB ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #29 on issue 2335 by d...@gnu.org: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 #(gc) does not add music events in a music list. So unless it is an input alignment problem (does adding empty lines change anything?), this is a rather sure pointer to memory protection problems. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #30 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 valgrind reports a ton of memory invalid reads... but it also reports a ton of jumps depending on uninitialized variables. I mean, not just running valgrind on lilypond processing this example; running valgrind on lilypond with no input at all produces tons of those warnings. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #31 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Adding empty lines does not help (I tried it almost immediately at the beginning, even before reporting the issue). Just in case someone overlooked it, I repeat what I wrote in my original mail. Redirecting output to a file or piping it to other process (stdout alone is enough, yes, the one that is not used...), makes segfault go away too. Graham, would it be possible to build LilyPond in the same way as provided binaries are built, but w/o stripping debug symbols? It could possibly help a bit in tracking it down (or not, who knows...). ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Comment #32 on issue 2335 by gra...@percival-music.ca: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Graham, would it be possible to build LilyPond in the same way as provided binaries are built, but w/o stripping debug symbols? No. Or rather: it is certainly possible; here's the code: https://github.com/gperciva/gub but I'm not going to do it. I'm already about 10 hours over my limit for this week. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
Updates: Labels: -Type-Critical Type-Crash Comment #34 on issue 2335 by philehol...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Keeps head down... The reason I didn't mark this as a critical regression is that I spent some time looking at it and discussing it with the reporter before raising it. I concluded that it's highly likely that it's been in the build for a long time, and appears and disappears semi-randomly based on factors we don't understand. We define a regression as something that used to work deliberately in a previous stable release. My contention would be that whether this input crashed a 64 bit linux build is pure happenstance - there's no evidence that any change has made this happen. I'm not suggesting not trying to fix this. Just that, by the rigorous defintition of critical regression - this isn't one. ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
Re: Issue 2335 in lilypond: 64 bit LilyPond segfaults
lilyp...@googlecode.com writes: Comment #31 on issue 2335 by prze...@gmail.com: 64 bit LilyPond segfaults http://code.google.com/p/lilypond/issues/detail?id=2335 Adding empty lines does not help (I tried it almost immediately at the beginning, even before reporting the issue). Just in case someone overlooked it, I repeat what I wrote in my original mail. Redirecting output to a file or piping it to other process (stdout alone is enough, yes, the one that is not used...), makes segfault go away too. Turns unbuffered output into buffered output, and buffers use memory allocation. -- David Kastrup ___ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
