[Debian-science-sagemath] GAP: issue related to compressed manual.six: PATCHES: reproducing issue

[Ximin Luo]

+Bill and the bug back in

Jerome BENOIT:
> [..]
> 
> gap-alnuth gap-atlasrep gap-autpgrp gap-character-tables gap-core gap-dev gap-doc gap-gapdoc gap-libs gap-online-help gap-polycyclic gap-prim-groups gap-small-groups gap-table-of-marks gap-trans-groups
> 
> Please try with this Debain GAP packages installed.
> 

Thanks Jerome, now I can reproduce this issue. In fact the problem is with gap-autogrp. If I install

$ sudo aptitude install gap-atlasrep gap-character-tables gap-prim-groups gap-small-groups gap-table-of-marks gap-trans-groups

I cannot reproduce the bug. As soon as I install gap-autogrp, the problem occurs:

$ ./sage -c "print(gap.help('SymmetricGroup', pager=False)[:100])"
[..]
    raise RuntimeError(message)
RuntimeError: Gap produced error output
Error, no method found! For debugging hints type ?Recovery from NoMethodFound
Error, no 1st choice method found for `+' on 2 arguments

   executing ? SymmetricGroup
1

In fact if you look at it:

$ zcat /usr/share/gap/pkg/AutPGrp/doc/manual.six.gz | head -n3
C autpgrp.tex 1. Introduction
C method.tex 2. The automorphism group method
F 2.0. AutomorphismGroup

I know nothing about GAP but this does not look like the other manual files, which all start with:

$ zcat /usr/share/gap/pkg/ctbllib/doc/manual.six.gz | head -n3
#SIXFORMAT  GapDocGAP
HELPBOOKINFOSIXTMP := rec(
encoding := "UTF-8",

The packages gap-alnuth and gap-polycyclic are dependant on gap-autogrp, so I could not test them separately. However: 

$ zcat $(dpkg -L gap-polycyclic | grep six) | head -n3
#SIXFORMAT  GapDocGAP
HELPBOOKINFOSIXTMP := rec(
encoding := "UTF-8",

$ zcat $(dpkg -L gap-alnuth | grep six) | head -n3
C intro.tex 1. Introduction
S 1.1. Acknowledgements
S 1.2. License

So I would guess that the former is "OK" but the latter would need to be "fixed".

Arguably GAP should not error out the whole "HELP" command when encountering one of the "bad" files. But probably either you or Bill know better here, what the "real" problem is.

BTW, I can confirm that the Python issue discussed previously is a red herring and NOT the cause of this bug. I've added the work around to our Debian SageMath patches, but all it does is silence the "gzip: stdout: Broken pipe" message, it has no other effect on the above things.

X

-- 
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git