Re: Issues on AIX when sorting revision summaries (pass 5)

HI.

I resolve the problem with the folowing custom sort:
It is a derty perl script, but it work ( I do not know in wich branch of the
if ... but it work)
--------------------------------
#!/usr/bin/perl
my $inf = $ARGV[0];
my $outf = $ARGV[2];
my $xarg = $#ARGV;
if ( $xarg == 0 ) {
  open (ORIGFILE, $inf);
  print sort(<ORIGFILE>);
} else {
  open (ORIGFILE, $inf);
  open (FINALFILE, "> ".$outf);
  close (FINALFILE);
  close (ORIGFILE);
}

regards

mario

mhagger wrote:
> 
> Greg Mundy wrote:
>> I currently have cvs2svn running on an AIX box and am getting hung up on
>> an issue during pass 5. From my understanding, pass 5 is where CVS
>> revision summaries are sorted. I'm not 100% sure, but I believe that the
>> tool is using the native AIX sort operation to handle this.
> 
> As explained in the FAQ, cvs2svn currently requires GNU sort.  If you
> have GNU sort installed elsewhere on your computer, you can tell cvs2svn
> the path of the executable using the "--sort=PATH" option.  Another
> alternative is to run the conversion on another computer then copy the
> dumpfile to the destination server and "svnadmin load" it there.
> 
> There has been a suggestion to implement a Python sort to remove this
> dependency on GNU sort [2], but so far nobody has stepped forward to
> either do the work or sponsor it.
> 
> Michael
> 
> [1] http://cvs2svn.tigris.org/cvs2svn.html#reqs
> [2] http://cvs2svn.tigris.org/issues/show_bug.cgi?id=123
> 
> ------------------------------------------------------
> http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=2390430
> 
> To unsubscribe from this discussion, e-mail:
> [[email protected]].
> 
> 

-- 
View this message in context: http://old.nabble.com/Issues-on-AIX-when-sorting-revision-summaries-%28pass-5%29-tp25264951p27657799.html
Sent from the cvs2svn - users mailing list archive at Nabble.com.

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=2449155

To unsubscribe from this discussion, e-mail: [[email protected]].