Closed kiwifb closed 7 years ago
Upstream: Fixed upstream, but not in a stable release.
All the above are solved by upgrading BRiAl
to git master. I'll prepare a release of BRiAl
.
Author: Dima Pasechnik
I need this for #22679 :-)
New commits:
a3e4c6b | brial updated to commit 63f747626822f5e0fa2bf975f7801fcc988eb530 |
Branch: u/dimpase/brial086
Description changed:
---
+++
@@ -71,3 +71,5 @@
are consequences of segfaults above.
+tarball http://users.ox.ac.uk/~coml0531/sage/brial-0.8.6.tar.bz2 +
I know, sorry I cannot seem to move my ass to make a proper release. On it now.
Changed upstream from Fixed upstream, but not in a stable release. to Fixed upstream, in a later stable release.
Need to fix the checksum to the properly released tarball. I'll put it back in review once it is done.
Description changed:
---
+++
@@ -71,5 +71,5 @@
are consequences of segfaults above.
-tarball http://users.ox.ac.uk/~coml0531/sage/brial-0.8.6.tar.bz2 +tarball https://github.com/BRiAl/BRiAl/releases/download/0.8.6/brial-0.8.6.tar.bz2
Reviewer: François Bissey
I'll put it in positive review but can you double check that the new tarball works for you.
New commits:
0943ae5 | updating checksum for official release |
Changed branch from u/dimpase/brial086 to u/fbissey/brial-0.8.6
ironically, it does not work on gentoo Lunix with gcc 5.4.0.
sage -t --long --warn-long 118.6 src/sage/rings/polynomial/multi_polynomial_sequence.py # Killed due to segmentation fault
and two more brial-related errors. While numerous
warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
are probably OK,
[sagelib-7.6] In file included from /home/dima/Sage/sage-dev/src/sage/libs/polybori/pb_wrap.h:22:0,
from /home/dima/Sage/sage-dev/src/build/cythonized/sage/rings/polynomial/pbori.cpp:517:
/home/dima/Sage/sage-dev/src/sage/ext/ccobject.h: In instantiation of ‘void Delete(T*) [with T = polybori::groebner::ReductionStrategy]’:
/home/dima/Sage/sage-dev/src/build/cythonized/sage/rings/polynomial/pbori.cpp:50398:51: required from here
/home/dima/Sage/sage-dev/src/sage/ext/ccobject.h:77:3: warning:
deleting object of polymorphic class type ‘polybori::groebner::ReductionStrategy’
which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor]
delete mem;
^
might spell trouble. It could also be that we must switch to gcc 6. Can you check in what Linux configuration this works for you?
(meahwhile I'll do distclean and a full rebuild, just to see if this still persists)
I confirm, it breaks things here too with gcc-5.4.0. I think it may be down to https://github.com/BRiAl/BRiAl/issues/11#issuecomment-269957098 I'll have to check on that.
I'll have a look at how it plays together with #22554 (which isn't in this branch yet).
Replying to @dimpase:
I'll have a look at how it plays together with #22554 (which isn't in this branch yet).
no, it does not help.
OK I have a fix. I cut a bit too much stuff. Actually adding one character would have been sufficient to fix gcc-6 compilation. It is still worth cleaning some stuff that wouldn't work anyway. I am not sure when I will have time to push the fix, I am rather busy this week end.
brial-0.8.7
(now linked to in the ticket description) should work for all reasonably recent compilers. I am not sure about keeping -std=c++98
in CXXFLAGS
as it could cause problem in the near future I think. I haven't updated the branch yet, feel free to beat me to it.
Description changed:
---
+++
@@ -71,5 +71,5 @@
are consequences of segfaults above.
-tarball https://github.com/BRiAl/BRiAl/releases/download/0.8.6/brial-0.8.6.tar.bz2 +tarball https://github.com/BRiAl/BRiAl/releases/download/0.8.7/brial-0.8.7.tar.bz2
0.8.7 works well on FreeBSD+clang. I'm happy to give it positive review once we have a branch...
Did you try on gentoo as well? Since 0.8.6 was working on FreeBSD+clang but not on Gentoo+gcc?
checking now on Linux. Will take a while---doing a clean rebuild of sage 8.0.beta...
Changed branch from u/fbissey/brial-0.8.6 to u/dimpase/brial-0.8.6
OK, it works on Linux too. No idea why the branch is red though.
Interesting message: "branch doesn't exist".
oops...
Changed branch from u/dimpase/brial-0.8.6 to u/dimpase/brial086
Changed reviewer from François Bissey to François Bissey, Dima Pasechnik
Changed author from Dima Pasechnik to François Bissey, Dima Pasechnik
Changed branch from u/dimpase/brial086 to 382b25c
Split from #12426. Doctest involving
polybori
segfault when everything is built with clangBacktrace
Similarly
sage/rings/polynomial/pbori.pyx
finally the failures in
are consequences of segfaults above.
tarball https://github.com/BRiAl/BRiAl/releases/download/0.8.7/brial-0.8.7.tar.bz2
Upstream: Fixed upstream, in a later stable release.
Component: porting
Author: François Bissey, Dima Pasechnik
Branch/Commit:
382b25c
Reviewer: François Bissey, Dima Pasechnik
Issue created by migration from https://trac.sagemath.org/ticket/22677