Opened 3 years ago

Closed 3 years ago

Last modified 3 years ago

#27367 closed defect (fixed)

Upgrade to mpfrcx 0.5

Reported by: dimpase Owned by:
Priority: major Milestone: sage-8.7
Component: packages: optional Keywords: upgrade, mpfrcx
Cc: slelievre, embray, fbissey Merged in:
Authors: Dima Pasechnik Reviewers: Travis Scrimshaw
Report Upstream: N/A Work issues:
Branch: 7ac6cf0 (Commits, GitHub, GitLab) Commit:
Dependencies: Stopgaps:

Status badges

Description (last modified by slelievre)

This ticket is to upgrade to mpfrcx 0.5.

Tarball:

The upgrade fixes a problem with #27212, which does not work with mpfrcx 0.4.2.

We have not upgraded since introducing mpfrcx 0.4.2 as an optional package in #11806.

Change History (15)

comment:1 Changed 3 years ago by dimpase

  • Branch set to u/dimpase/packages/mpfrcx05
  • Commit set to 7ac6cf02adc32aeb41979f28210ebce0c1e19b0a

needs testing with Sage-supplied GMP/MPIR (works with #27212 - system GMP)


New commits:

7ac6cf0update mpfrcx to version 0.5

comment:2 Changed 3 years ago by slelievre

  • Cc slelievre added
  • Description modified (diff)
  • Keywords upgrade mpfrcx added
  • Summary changed from update mpfrcx to version 0.5 to Upgrade to mpfrcx 0.5

Adding keyword "upgrade" so the release manager knows to upload the tarball.

comment:3 Changed 3 years ago by dimpase

  • Status changed from new to needs_review

comment:4 Changed 3 years ago by dimpase

I tested this a bit with SAGE_CHECK=yes, looks good. (tested on OSX 10.13 with clang, on Ubuntus, on Fedora...)

Last edited 3 years ago by dimpase (previous) (diff)

comment:5 Changed 3 years ago by dimpase

  • Cc embray added

comment:6 Changed 3 years ago by dimpase

  • Cc fbissey added

comment:7 Changed 3 years ago by tscrim

  • Reviewers set to Travis Scrimshaw
  • Status changed from needs_review to positive_review

LGTM. May the buildbots smile upon you. :P

comment:8 Changed 3 years ago by vbraun

  • Branch changed from u/dimpase/packages/mpfrcx05 to 7ac6cf02adc32aeb41979f28210ebce0c1e19b0a
  • Resolution set to fixed
  • Status changed from positive_review to closed

comment:9 Changed 3 years ago by embray

  • Commit 7ac6cf02adc32aeb41979f28210ebce0c1e19b0a deleted

I would really like to get the Cygwin buildbot working again but I'll never get there if we keep merging package upgrades without giving me an opportunity to test them...

comment:10 Changed 3 years ago by dimpase

tell me how to set up a Windows buildbot. I tried using GCE running Windows, but it's just hopeless...

comment:11 follow-up: Changed 3 years ago by embray

We already have one, but it was previously broken due to something (probably openblas) causing the docbuild to hang, and it's just never been working since then: http://build.sagemath.org/#/builders/37

comment:12 Changed 3 years ago by embray

Seems like the Cygwin buildbot is in slightly better shape than I thought? It is building successfully, but then when it gets to running the doctests, it bizarrely exits with error code 1 immediately:

http://build.sagemath.org/#/builders/37/builds/15/steps/8/logs/stdio

This despite being perfectly able to run the tests manually. When run from buildbot it just fails immediately.

comment:13 Changed 3 years ago by dimpase

a usual Cygwin/PTY related headache? Random googling only finds netbeans-related stuff, and I don't even know what netbeans is. https://stackoverflow.com/questions/27284742/unable-to-start-pty-process

comment:14 in reply to: ↑ 11 Changed 3 years ago by tscrim

Replying to embray:

We already have one, but it was previously broken due to something (probably openblas) causing the docbuild to hang, and it's just never been working since then: http://build.sagemath.org/#/builders/37

Sorry Eric, I thought that buildbot was there to catch when upgrades like this break things.

comment:15 Changed 3 years ago by embray

It's ok, it's probably fine. Turns out the Cygwin buildbot has at least been doing builds just not running the tests. Which is not to say something couldn't have still broken with this, but at least we know it builds. I think Volker just isn't checking that buildbot since it still isn't stable.

I'm going to test this locally ASAP but I don't think it's likely to cause any headaches.

Note: See TracTickets for help on using tickets.