GAP 4.7.5

[vbraun]

• ​http://boxen.math.washington.edu/home/vbraun/upstream/libgap-4.7.5.tar.gz
• ​http://boxen.math.washington.edu/home/vbraun/upstream/gap-4.7.5.tar.bz2
• ​http://boxen.math.washington.edu/home/vbraun/upstream/gap_packages-4.7.5.tar.bz2
• ​http://boxen.math.washington.edu/home/vbraun/upstream/database_gap-4.7.5.tar.bz2

[vbraun]

New commits:

​75c08db	update to gap 4.7.5
​2db4551	the osx 10.4 fix is now upstream
​fbddb40	the sparc bugfix is now upstream

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​03ad42a

Increase GAP pool

[dimpase]

does "gap: cannot extend the workspace any more!" stuff really break the docbuild? I remember seeing on sage-develop statements to the contrary.

[vbraun]

Well according to strace it does. Gap has some warnings and some hard errors for low memory, and this one is apparently a hard error. I looked at the source and it means that sbrk failed, so its not unsurprising.

[leif]

Well, the whole idea of importing (=running) code just to build its documentation is a very bad idea anyway (though of course "powerful"), but that's a different story...

I'll give it a try, but I can't imagine I'm actually running out of (virtual) memory in my docbuild-broken upgraded installation.

[vbraun]

You are not necessarily running out of virtual memory, sbrk can fail for other reasons.

[leif]

Replying to leif:

I'll give it a try, but I can't imagine I'm actually running out of (virtual) memory in my docbuild-broken upgraded installation.

Great! That triggers the rebuild of all of Sage for no (real) reason!

Praise the "new workflow" which makes everything so much easier!

[dimpase]

Replying to leif:

Replying to leif:

I'll give it a try, but I can't imagine I'm actually running out of (virtual) memory in my docbuild-broken upgraded installation.

Great! That triggers the rebuild of all of Sage for no (real) reason!

happens to me too...

Praise the "new workflow" which makes everything so much easier!

developing software of the future on the hardware of the past :-)

[leif]

Replying to dimpase:

Replying to leif:

Replying to leif:

I'll give it a try, but I can't imagine I'm actually running out of (virtual) memory in my docbuild-broken upgraded installation.

Great! That triggers the rebuild of all of Sage for no (real) reason!

happens to me too...

At least it fixed the issue for me... (I didn't recall I had gap_packages installed into the installation.)

(Although probably it was really the Maxima upgrade to 5.33 which fixed it, or one of the other 50 tickets that got pulled in by pulling the branch here... ;-) )

[dimpase]

I am running tests with --optional=sage,database_gap,gap_packages, and seeing few errors. E.g.

sage -t --long src/sage/combinat/sloane_functions.py
**********************************************************************
File "src/sage/combinat/sloane_functions.py", line 372, in sage.combinat.sloane_functions.A000001._eval
Failed example:
    sloane.A000001._eval(51) # optional - database_gap
Expected nothing
Got:
    1
**********************************************************************

There are also errors in src/sage/coding/linear_code.py related to guava. Are they for a separate ticket?

[vbraun]

The guava failures are #14979

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​8026003

fix doctest failures except the guava one (see #14979)

[dimpase]

LGTM

[dimpase]

the tarfiles are gone here! In fact, ​http://boxen.math.washington.edu/home/ is almost empty now...