#11959: Remove the necessity to set SAGE_PARALLEL_SPKG_BUILD
------------------------------+---------------------------------------------
   Reporter:  jhpalmieri      |          Owner:  GeorgSWeber                    
        
       Type:  enhancement     |         Status:  needs_work                     
        
   Priority:  minor           |      Milestone:  sage-4.7.3                     
        
  Component:  build           |       Keywords:  SAGE_PARALLEL_SPKG_BUILD MAKE 
-j --jobs
Work_issues:                  |       Upstream:  N/A                            
        
   Reviewer:  Leif Leonhardy  |         Author:  John Palmieri                  
        
     Merged:                  |   Dependencies:                                 
        
------------------------------+---------------------------------------------

Comment(by leif):

 P.S.: Using a maximal (avg) sysload of 2.5 to me seems a bit funny [even]
 as an example; I'd use ''at least'' the number of cores, which today is
 usually 4 or more.

 [[BR]]

 In `README.txt`, grepping for "An error" is sufficient and best used with
 `-l` (which only prints the filenames of matches), or perhaps `-c`, which
 prints the number of matches in each file (unfortunately also including
 zero, i.e., one should postprocess that output further), since we always
 append to the logs and hence previous errors never vanish unless one
 deletes (or edits) the log.

 Since the error message ("An error ...") already contains the package's
 name, using
 {{{
 #!sh
 $ grep -h "An error" spkg/logs/*
 }}}
 would also give a more readable output, as an alternative to `-l`, with
 the advantage that one immediately sees whether a package failed to
 install multiple times.

 To get the actual error messages, (e)grepping for "Error" and "Failed" at
 the beginning of the line (`egrep "^(Error|Failed)" spkg/logs/...`^1^),
 optionally with a (non-POSIX) before-context (`-B`), is usually more
 informative.  When case-insensitively ( `-i`) grepping for any  "`error`",
 to suppress frequent errors ignored by `make`, one can pipe the whole
 through `grep -iv ignored`.  `-n` also prepends the line numbers to the
 output, which is useful if you're going to open the log in an editor.

 [[BR]]

 Do we mention `LC_ALL=C` (or e.g. `LC_MESSAGES=en_US.utf8`) for submitting
 error logs?

 ----
 ^1^ `"^Failed"` usually also gives non-fatal errors for the Python spkg
 (''"Failed to find the necessary bits..."'').  In the long run, all of
 ''Sage's'' error messages should really start with "Error" rather than
 "Failed".  (The new ATLAS spkg with the Python install script in contrast
 raises arbitrary exceptions.)

 We should really in addition create a fresh temporary file at the start of
 each build to keep track of which packages failed to build, to be dumped
 at the end (in case of errors).

-- 
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/11959#comment:20>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica, 
and MATLAB

-- 
You received this message because you are subscribed to the Google Groups 
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/sage-trac?hl=en.

Reply via email to