pkgsrc-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: PR/47769 CVS commit: pkgsrc/textproc/ispell-ro



The following reply was made to PR pkg/47769; it has been noted by GNATS.

From: rodent%NetBSD.org@localhost
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: PR/47769 CVS commit: pkgsrc/textproc/ispell-ro
Date: Fri, 26 Apr 2013 15:51:58 -0400

 On Fri, Apr 26, 2013 at 05:30:12PM +0000, Robert Elz wrote:
 >  
 >  I don't think that was the right solution - the DIST_SUBDIR was there
 >  (I believe) because the distfile has no version info in its name.
 
 Hmm, no. DIST_SUBDIR was added, because the distfile changed and was not
 versioned. That is the function of DIST_SUBDIR. It's not pkgsrc's position to
 care what upstream names distfiles, as long as versioning can be determined in
 some way. Ex. it is not important if the file were named foo (version 1.0) and
 the subsequent version (1.1) was named bar. DIST_SUBDIR would not be used in
 that case unless 1.1 was also named foo. See commit message for 1.2 of
 distfile.
   
 >  Not having version info anywhere in a distfile name is a recipe for
 >  (eventual) madness - putting the file in a DIST_SUBDIR that contains
 >  a version identifier (it was a date) seems like a sane solution.
 
 It will only be a problem in this package if upstream regenerates the distfile
 for the next version and doesn't choose to name it something sensible. Will
 fix it if it breaks.
 
 >  Of course, since it appears that the distfile hasn't changed in the past
 >  10 years, it may be that no-one will ever be bitten by this particular
 >  case, but still, in general, distfiles without any kind of version ID
 >  are a very very bad idea.
 
 Ideally, this would have been communicated to upstream so that the next
 version will not have the same issue.
 


Home | Main Index | Thread Index | Old Index