Bill Allombert: > On Fri, Dec 02, 2016 at 05:39:24AM +0000, Jerome BENOIT wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA512 >> >> Hello All, >> >> finally I figured out a way to implement compression (gz) with the help >> of the zlib library rather than through gzip pipes. So snow Sage no more >> emits failures concerning GAP. Please find in attachment a patch that >> replaces >> gzip pipes with zlib high level functions, as well concomitant patches. >> I have also attached an updated d/p/series files: note that the patch >> d/p/fix-compressed-six-files is no more necessary and that some autoconf >> scripts have to be regenerated (I followed the hint given by >> dpkg-builpackage). >> >> The material could not be deposited at Alioth as there is no GAP repository >> there. >> On the other hand, I have just uploaded the source ball, the debian stuff >> and the debballs at the Sage repository deb-sci-sage (see the Debain Sage >> page >> [1] for more inrmation); therefore it can be easily tested within a schroot >> environment. >> >> At last, as the main patch, namely fix-zlib-stringfile, slightly improves >> the GAP kernel itself, it must also be applied to the libGAP package: >> I will do if it is effectively applied to GAP. > > I am sorry, but I cannot apply that to gap. We are already during the > freeze. This is too intrusive and does not solve any issue with gap. > > However you can send this upstream. > > But more importantly, this does not fix the underlying Sage bug that it > messes with signal of subprocesses. There are other uses of pipes in > GAP, and there are other software in Sage that use pipes+gzip (PARI/GP > at least). So this patch is fixing the symptom and not the problem. > Using pipes is perfectly legitimate and has to be supported by Sage. > > This bug should be reassigned to sage. >
Hi all, Sorry I only just briefly scanned through the thread now. However I found this post relating to gzip, pipes and python: https://blog.nelhage.com/2010/02/a-very-subtle-bug/ which summarises the bug report here: https://bugs.python.org/issue1652 I hope this helps. If not, sorry for the noise. I will try to debug this myself on Thursday or Friday. X -- GPG: ed25519/56034877E1F87C35 GPG: rsa4096/1318EFAC5FBBDBCE https://github.com/infinity0/pubkeys.git