GAP 4.14.0-beta1

[fingolfin]

Dear all,

I plan to release GAP 4.14.0 in 1-2 weeks (before December 1st) assuming all goes well. To prepare for that, I've made a prerelease to allow testing.

Version 4.14.0-beta1 is mostly intended for use by people who package GAP for Linux distributions, or similar things (Homebrew, FreeBSD ports, Conda, SageMath, ...). The idea is to iron out issues related to packaging before the final release.

That said, everyone else is of course also welcome to give it a try. I don't expect major changes in the final release, though we'll still merge PRs and package updates if they don't appear "risky".

You can find this pre-release at https://github.com/gap-system/gap/releases/tag/v4.14.0-beta1

If you want all GAP sources and all packages, you need gap-4.14.0-beta1.tar.gz.

Some people who I know package GAP and might be interested in this include @antonio-rojas @isuruf @saraedum @jamesjer @orlitzky @collares @AndreasEnge @yurivict @kiwifb @mkoeppe

[yurivict]

I verified that 4.14.0-beta1 builds and passes all tests from the FreeBSD port's perspective.

[fingolfin]

Thank you!

[jamesjer]

I did some test builds for Fedora and everything seems okay so far. I have not tried building all of the packages yet, but will do so over the next few days.

[fingolfin]

Thank you @jamesjer

[orlitzky]

We have some fresh Gentoo bugs for GCC-15 compatibility,

• https://bugs.gentoo.org/943925
• https://bugs.gentoo.org/944356
• https://bugs.gentoo.org/944036

due to -Werror=incompatible-pointer-types and the switch to C23 as the default standard (which e.g. has its own bool type). I've built my own GCC-15 and can confirm that the problems persist in the latest versions of things, including the beta.

If the development roadmap can be trusted, GCC-15 is in stage three (out of four) of its release cycle, and could be released while GAP 4.14.0 is current (typically April/May). Now might be the time to either fix what it is complaining about, or append the necessary -Wno-error and -std flags to the builds system.

[orlitzky]

With GCC 14.2.1, things are mostly OK.

I somehow have test failures in hap-1.66 even with GAP 4.13.1:

testing: /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testallV11/poin.tst
   ? `intvec`(GAP_intvec) = `bigintvec` is not supported
   ? expected `intvec` = `intvec`
   ? error occurred in or before STDIN line 6948: `, 1, intvec(1,1,1,2,2,3,4) );`
   ? expected intvec-expression. type 'help intvec;'
########> Diff in /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testallV11/poin.tst:3
# Input is:
P:=PoincareSeriesLHS(G);
# Expected output:
(1)/(-x_1^3+3*x_1^2-3*x_1+1)
# But found:
Error, no method found! For debugging hints type ?Recovery from NoMethodFound
Error, no 1st choice method found for `Remove' on 2 arguments
########
########> Diff in /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testallV11/poin.tst:5
# Input is:
RankHomologyPGroup(G,P,1000);
# Expected output:
251000
# But found:
The input must be a rational function follwed by a positive integer. 
Error, List Element: <list> must be a list (not the value 'fail')
########

With 4.14.0-beta1, I still get those two, but also a few others:

testing: /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testall3/3.9.3.tst
   ? `intvec`(GAP_intvec) = `bigintvec` is not supported
   ? expected `intvec` = `intvec`
   ? error occurred in or before STDIN line 3963: `, 1, intvec(1,1,1,2,2,3,4) );`
   ? expected intvec-expression. type 'help intvec;'
########> Diff in /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testall3/3.9.3.tst:4
# Input is:
P:=HilbertPoincareSeries(A);;
# Expected output:
# But found:
Error, no method found! For debugging hints type ?Recovery from NoMethodFound
Error, no 1st choice method found for `Remove' on 2 arguments
########
########> Diff in /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testall3/3.9.3.tst:5
# Input is:
RankHomologyPGroup(G,P,1001);
# Expected output:
251503
# But found:
The input must be a rational function follwed by a positive integer. 
Error, List Element: <list> must be a list (not the value 'fail')
...
testing: /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testallV11/cohodata.tst
   ? `intvec`(GAP_intvec) = `bigintvec` is not supported
   ? expected `intvec` = `intvec`
   ? error occurred in or before STDIN line 3981: `, 1, intvec(1,1,4) );`
   ? expected intvec-expression. type 'help intvec;'
########> Diff in /var/lib/portage/tmp/portage/dev-gap/hap-1.66/temp/pkg/hap-1.66/tst/testallV11/cohodata.tst:2
# Input is:
CohomologicalData(SmallGroup(8,4),20);
# Expected output:
Integer argument is large enough to ensure completeness of cohomology ring presentation.

Group order: 8
Group number: 4
Group description: Q8

Cohomology generators
Degree 1: a, b
Degree 4: c

Cohomology relations
1: a^2+a*b+b^2
2: b^3

Poincare series
(x^2+x+1)/(-x^3+x^2-x+1)

Steenrod squares
Sq^1(c)=0
Sq^2(c)=0

<algebra of dimension 15 over GF(2)>
# But found:
Integer argument is large enough to ensure completeness of cohomology ring presentation.

Error, no method found! For debugging hints type ?Recovery from NoMethodFound
Error, no 1st choice method found for `Remove' on 2 arguments

...

-----------------------------------
total    411750 ms (58485 ms GC) and 13.3GB allocated
              5 failures in 3 of 89 files

#I  Errors detected while testing

fgets() failed with errno 5
I/O error

[fingolfin]

The hap failure you describe suggest an issue with Singular (the messages involving intvec etc. look like they come from Singular). Perhaps you are using a newer Singular version than we use in CI and it once again has a breaking change. Not too much we can do about that right now then.

[jamesjer]

Yes, Singular 4.4 changed the return type of "hilb" from intvec to bigintvec. Patching lib/HapPrime/gradedalgebra.gi to expect the bigintvec return type still leaves a failing test, though. In tst/testall1/1.3.2.txt, the call to BettiNumber(M,0) on line 9 is expected to return 20, but returns 80 with Singular 4.4.