On Sun, Aug 31, 2008 at 10:23:36PM +0200, Jakub Narebski wrote: > On Sub, 31 August 2008, Avery Pennarun wrote: > > I have a proper implementation in the 'runlock' script in gitbuilder: > > > > http://github.com/apenwarr/gitbuilder/tree/master/runlock > > > > In that particular case, I wanted to handle signals carefully, so I > > needed the manual fork thing even in perl 5.8. You can safely remove > > the signal handling stuff (and of course the lockfile stuff) if you > > just want a minimal safe fork-exec-wait implementation in perl. > > But if we go this way, i.e. fork+exec (perhaps implicit fork), why do > not simply use appropriate commands from Git.pm (Git::Repo doesn't > have it yet, IIRC). As far as I remember Git.pm was created initially > to unify all different "safe_pipe" and "safe_cmd" implementations among > different Perl scripts in Git (Petr "Pasky" Baudis CC-ed). Can anyone give a concrete justification for Perl 5.6 support? Who is needing it and why do they have to use Perl 5.6? Does it offset the time spent discussing and reviewing this and the maintenance burden of extra complicated code? -- Petr "Pasky" Baudis The next generation of interesting software will be done on the Macintosh, not the IBM PC. -- Bill Gates -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
