Opened 9 years ago

Closed 9 years ago

#13210 closed defect (duplicate)

ATLAS: flush output before os.system()

Reported by: jdemeyer Owned by: tbd
Priority: minor Milestone: sage-duplicate/invalid/wontfix
Component: packages: standard Keywords:
Cc: Merged in:
Authors: Reviewers: Jeroen Demeyer
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description (last modified by jdemeyer)

When doing something like

./sage -i atlas &>my_log_file

the output to my_log_file will be badly ordered because we're not flushing stdout/stderr before calling system() in spkg-install.

New spkg: http://boxen.math.washington.edu/home/jdemeyer/spkg/atlas-3.8.4.p2.spkg (for review, changes not yet committed)

atlas-3.8.4.p2 (Jeroen Demeyer, 6 July 2012)

  • Trac #13210: Flush stdout and stderr before calling system().
  • Detect the Apple linker (which is still unsupported).
  • Remove unneeded .orig files in patches/
  • Remove mmsearch-with-temp-Solaris-fix.c patch (was needed for gcc-4.4.0 on Solaris).
  • Remove copying of patches/*tgz since there are no such files.
  • Use 'patch' for patching.

Attachments (1)

atlas-3.8.4.p1-p2-nodelete.diff (13.8 KB) - added by jdemeyer 9 years ago.
Diff for the atlas spkg (without deleted files). For reference / review only.

Download all attachments as: .zip

Change History (11)

comment:1 Changed 9 years ago by leif

AFAIK the "error messages" are suboptimal as well, i.e., at least in some cases exceptions get raised (e.g. KeyError IIRC) and tracebacks printed instead of "Error: <informative text>".

comment:2 Changed 9 years ago by vbraun

I think Python tracebacks are more useful than manual error messages, not less. Of course os.system() doesn't raise exceptions so some amount of manual error messages is unavoidable.

comment:3 Changed 9 years ago by jdemeyer

  • Description modified (diff)

comment:4 Changed 9 years ago by jdemeyer

  • Description modified (diff)

comment:5 follow-up: Changed 9 years ago by vbraun

I'd rather keep working on packaging Atlas 3.9.x, any further frobnicating the old package is a waste of time imho.

comment:6 in reply to: ↑ 5 Changed 9 years ago by jdemeyer

Replying to vbraun:

I'd rather keep working on packaging Atlas 3.9.x, any further frobnicating the old package is a waste of time imho.

I don't think "I'm working on a new version of feature X" is ever a good excuse to not fix the old feature X.

Changed 9 years ago by jdemeyer

Diff for the atlas spkg (without deleted files). For reference / review only.

comment:7 Changed 9 years ago by jdemeyer

  • Status changed from new to needs_review

Besides, I think the changes here are likely good for ATLAS-3.9.x also.

Needs review.

comment:8 Changed 9 years ago by vbraun

ATLAS-3.9.x is sufficiently different that it requires major changes. The shared library stuff is completely different for starters. I'm happy to apply the changes that still apply in the next version but I'm not going to spend time reviewing this ticket.

comment:9 Changed 9 years ago by vbraun

Superseded by #10508

comment:10 Changed 9 years ago by jdemeyer

  • Authors Jeroen Demeyer deleted
  • Milestone changed from sage-5.4 to sage-duplicate/invalid/wontfix
  • Resolution set to duplicate
  • Reviewers set to Jeroen Demeyer
  • Status changed from needs_review to closed
Note: See TracTickets for help on using tickets.