Update pari to version 2.8.0

[videlec]

There is a critical bug in the current snapshot of PARI (see #21000).

Later snapshots from PARI have problems with polredbest(), see http://pari.math.u-bordeaux.fr/cgi-bin/bugreport.cgi?bug=1836

Two macros appeared in the doc (\Hom and \Frob) see http://pari.math.u-bordeaux.fr/archives/pari-dev-1607/msg00020.html

Upstream tarball: http://pari.math.u-bordeaux.fr/pub/pari/unstable/pari-2.8.0.alpha.tar.gz

Upstream: Fixed upstream, in a later stable release.

CC: @jdemeyer @vbraun @defeo @kiwifb @jpflori

Component: packages: standard

Author: Vincent Delecroix, Jeroen Demeyer

Branch/Commit: 9cf5776

Reviewer: Volker Braun, Jeroen Demeyer, Vincent Delecroix

Issue created by migration from https://trac.sagemath.org/ticket/21005

[videlec]

New commits:

94b4b67	Trac 19779: Hadamard difference sets
7928cc1	Trac 19779: input blocks
85485d3	Trac 19779: review 1
0e13f24	Trac 19779: alternative if/elif flow
9bb3a46	Trac 19779: some comment and a better doctest
d785f91	Trac 19779: remove cartesian product flattening
9139f67	removed extra ':'
f1bd95d	Trac 19779: new style print statement
a40e6cf	Trac 19779: fix references
b7511d7	Trac 19779: new pari release

[videlec]

Branch: public/21005

[videlec]

Commit: b7511d7

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

de2a6c5

Trac 21005: new pari release

[sagetrac-git]

Changed commit from b7511d7 to de2a6c5

[sagetrac-git]

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

cf2383a

Trac 21005: update paridecl.pxd

[sagetrac-git]

Changed commit from de2a6c5 to cf2383a

[vbraun]

comment:5

Positive review if you fill in your name...

[vbraun]

Reviewer: Volker Braun

[videlec]

Author: Vincent Delecroix

[videlec]

comment:6

I have some trouble with auto generated functions. Some of them disappeared and break doctests (like padicprec).

[sagetrac-git]

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

ee42261

Trac 21005: more pari declaration

[sagetrac-git]

Changed commit from cf2383a to ee42261

[jdemeyer]

comment:10

positive_review from my part if this passes make ptestlong.

[jdemeyer]

Changed reviewer from Volker Braun to Volker Braun, Jeroen Demeyer

[vbraun]

comment:11

PDF build fails because of pound sign

src/sage/libs/pari/auto_gen.pxi:        defined by the equation :math:`\#E(\mathbb{F}_q) = q+1 - t`.

then

[docpdf] Overfull \hbox (623.6548pt too wide) in paragraph at lines 21949--21949
[docpdf] [] 
[docpdf] [278] [279] [280] [281]
[docpdf] ! You can't use `macro parameter character #' in math mode.
[docpdf] l.22319 dividing \(#
[docpdf]                     E(\mathbb{F}_q)\).
[docpdf] ? 
[docpdf] ! Emergency stop.
[docpdf] l.22319 dividing \(#
[docpdf]                     E(\mathbb{F}_q)\).
[docpdf] !  ==> Fatal error occurred, no output PDF file produced!
[docpdf] Transcript written on libs.log.
[docpdf] Makefile:68: recipe for target 'libs.pdf' failed
[docpdf] make[2]: *** [libs.pdf] Error 1
[docpdf] make[2]: Leaving directory '/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/share/doc/sage/latex/en/reference/libs'

[videlec]

comment:64

Ok. For the pdfdoc a naive substitution "@[startlword]Hom@[endlword]" -> "\\text{Hom}" and "@[startlword]Frob@[endlword]" -> "\text{Frob}" was enough. I guess it is better that way than the approach from my previous commits (​1ef03c4, ​b7a9144, ​5b88019, ​4a866cc).

[videlec]

comment:65

At least the html doc and the pdf doc build and tests pass in sage/libs/pari/!

[jdemeyer]

comment:66

Upstream is preparing a release(!) of PARI version 2.8.0.

Given that this ticket is no longer a blocker, I would actually suggest to wait for that release and package that.

Edit: never mind, I see you already posted this above

[jdemeyer]

comment:67

Replying to @kiwifb:

I am expecting Jeroen to move to the next pari development branch not long after the release unless something spectacular happens :)

Exactly. The last non-bugfix release of PARI (version 2.7.0) was 21 march 2014. More than 2 years between releases is a really long time, especially because PARI is quite actively developed.

[jdemeyer]

Changed upstream from Fixed upstream, but not in a stable release. to Fixed upstream, in a later stable release.

[jdemeyer]

comment:69

Did anybody run full doctests already with this new PARI? I did not do that yet.

[jdemeyer]

comment:70

OK. I will run make ptestlong.

[jdemeyer]

comment:71

Works fine for me. Let's wait until 2.8.0 is really released and then this is good to go I hope...

[jdemeyer]

Changed branch from u/vdelecroix/21005 to u/jdemeyer/21005

[jdemeyer]

Changed commit from 512df36 to 9cf5776

[jdemeyer]

New commits:

45e9053	Trac 21005: new pari release
36fdef4	Trac 21005: update paridecl.pxd
9cf5776	Trac 21005: fix pari doc

[jdemeyer]

comment:76

Passes all tests.

[kiwifb]

comment:77

Does it pass make bench properly for you? I have annoying formatting issues like

*** ../src/test/32/polyser      2016-04-09 02:36:45.000000000 +1200
--- gp.out      2016-08-15 09:17:49.995201272 +1200
***************
*** 19,25 ****
  xx
  ? factorpadic(apol,7,8)
  
! [(1 + O(7^8))*y + (6 + 2*7^2 + 2*7^3 + 3*7^4 + 2*7^5 + 6*7^6 + O(7^8)) 1]
  
  [(1 + O(7^8))*y^2 + (1 + 6*7 + 4*7^2 + 4*7^3 + 3*7^4 + 4*7^5 + 6*7^7 + O(7^8
  ))*y + (6 + 5*7 + 3*7^2 + 6*7^3 + 7^4 + 3*7^5 + 2*7^6 + 5*7^7 + O(7^8)) 1]
--- 19,26 ----
  xx
  ? factorpadic(apol,7,8)
  
! [                                                                           
!   (1 + O(7^8))*y + (6 + 2*7^2 + 2*7^3 + 3*7^4 + 2*7^5 + 6*7^6 + O(7^8)) 1]
  
  [(1 + O(7^8))*y^2 + (1 + 6*7 + 4*7^2 + 4*7^3 + 3*7^4 + 4*7^5 + 6*7^7 + O(7^8
  ))*y + (6 + 5*7 + 3*7^2 + 6*7^3 + 7^4 + 3*7^5 + 2*7^6 + 5*7^7 + O(7^8)) 1]
***************
*** 150,153 ****
  name
  ? if(getheap()!=HEAP,getheap())
  ? print("Total time spent: ",gettime);
! Total time spent: 0
--- 151,154 ----
  name
  ? if(getheap()!=HEAP,getheap())
  ? print("Total time spent: ",gettime);
! Total time spent: 10

[jdemeyer]

comment:78

I tested in Sage with SAGE_CHECK=yes which runs the full testsuite and it passed.

[jdemeyer]

comment:79

Apart from that, it also passes make bench outside of Sage.

Maybe report your problem upstream?

[jdemeyer]

comment:80

For me, this ticket is good. If Vincent (or anybody else) agrees, we can set this ticket to positive_review.

[kiwifb]

comment:81

I always have problems of formatting with the tests in sage-on-gentoo. But make bench has been ok before, The full testsuite usually reports failure because my output is formatted in 132 columns (guess) while pari seems to expect it to be formatted in 80 columns.
Those bizarre line jump in make bench are new.

Otherwise I am ok with the ticket, but I thought we were waiting for 2.8.0 final.

[videlec]

comment:82

the doc building is slightly worse... (but at least it fixes some bug)

[vbraun]

Changed branch from u/jdemeyer/21005 to 9cf5776