#perl
15 October 2007
Total 32 pages. You are browsing page 32/32.
First :: Prev :: [...] [28] [29] [30] [31] [32] :: Next :: Last
23:33 <****> bpalmer, ahh, just put the pids in a @array and wait on each one of them thx
23:34 <****> just what i needed.. just couldn't get my head around it
23:34 <****> er, waiting on an individual pid when you have four others is probably not the best idea
23:34 <****> you need to setup a $SIG{CHLD} handler and handle the dead children as they come in
23:34 <****> then just sleep;
23:35 <****> http://use.perl.org/~Alias/journal/34680
23:35 <****> heh
23:35 <****> Somni, you mean the @pid array would wait for the first one in array independently of when it returns
23:36 <****> cctoide: yow
23:36 <****> davidfetter I think that I would try someting like /==\s+Applied Patches\s+==\s+([^\s].*)(?=== Rejected Patches)/s
23:36 <****> nollan: say '10' is $pid[0], now you're waiting on pid 10, and pid 11 exits
23:36 <****> cctoide: hahaah oopz!
23:37 <****> _abc_, thanks, trying :)
23:37 <****> Somni: what's the big deal?
23:37 <****> Somni, yeah. thats what i meant.. would the 11 one be a zombie then?
23:37 <****> so it'll be zombie for a little while until they're all done ... no problems
23:37 <****> nollan: it would, but zombies are not inherently bad
23:37 <****> the problem is you're sitting there waiting, with only four active processes (or less)
23:38 <****> didn't you want to have five running at a time, constantly?
23:38 <****> that's what he wanted, no?
23:38 <****> Somni, yeah
23:38 <****> so you need to wait() when you know there's something to wait on
23:38 <****> oh. well, all right... but one of those problem statements above made it sound like he wanted to fork 5 and then wait for them all to finish before goign on.
23:38 <****> you can either sleep and wait on a SIGCHLD, or just call wait() and .. wait
23:39 <****> ah, ok i get it thanks
23:39 <****> bpalmer: yea, he was asking about this earlier
23:39 <****> bpalmer, i could do that as well but i'd prefer the other
23:39 <****> ok :)
23:39 <****> Can't you use POE / Event / IO::Async / etc.... for that?
23:43 <****> LeoNerd: no no, we don't ask these sorts of questions
23:49 <****> man, I really need to transition to svn
23:50 <****> also pg
23:50 <****> From..?
23:51 <****> cvs and mysql
23:51 <****> Ah... not much point I'd say
23:51 <****> SVN is just about as bad as CVS is,.. just in different ways
23:52 <****> my biggest problem with cvs are the hacks I've had to do to get it going over ssh remotely and pserver locally
23:52 <****> Somni: you might also think about looking into Mercurial, Git, Bazaar and possibly SVK. they all have some very nice features.
23:52 <****> My money would be on Bazaar, personally..
23:52 <****> But then I'm a bit biased :)
23:52 <****> also, the chroot jail crap, which I haven't looked at in a while
23:52 <****> i tried it for a while, but it felt a bit unpolished
23:53 <****> It probably was, it's approaching a 1.0 release currently.
23:53 <****> bzr sucks!
23:53 <****> but i'll give it another go at some point
23:53 <****> (I have no idea if it sucks or not, I'm just baiting you guys ;-)
23:53 <****> When they declare it 1.0 I'd imagine would be the time.. It's 0.91 or so now...
23:53 <****> PerlJam: you suck! also, Perl!
23:54 <****> PerlJam: ;P
23:54 * LeoNerd sucks...
23:54 <****> PerlJam: and jam!
23:54 <****> ... on a slice of pear
23:54 <****> Heh.. that'll teach me
23:54 <****> LeoNerd: do you suck in other places?
23:54 <****> I almost choked on it :P
23:54 <****> I've been using svn for a few years yet and git for the last several months. Each has its good things and its suckitude
23:55 <****> cvs is showing its age, and subversion is the only one I've looked at seriously
23:55 <****> we've been using SVN for a couple of years here at work. i've been wanting to mess with Mercurial for a while, though. its patch-queue is pretty sweet looking.
23:55 <****> SVN is just CVS that learned one cheap trick (cheap reference copy), and threw away lots of things as a result
23:55 <****> SVN has no branches, no tags.
23:55 * Somni shrugs
23:56 <****> hm, actually, I might start use branches or tags if they were convenient
23:56 <****> Branch/merge is just as painful as CVS.
23:56 <****> In fact, moreso in places
23:56 <****> it became a royal pain in cvs to keep them going
23:56 <****> If you want nice branches/merges, don't use SVN
23:56 <****> LeoNerd: Are you sure you've used svn?
23:56 <****> i used cvs for a long time, and svn is much easier to teach new users and get more people indoctrinated and using source control heavily
23:56 <****> Somni: branching in SVN is *easy*. merging can be a bit of a pain, though not as much as with CVS.
23:56 <****> SVN suffers one key flaw - it forgets a merge
23:56 <****> It has a command called "svn merge". Don't be fooled. It's the same as cvs up -j.
23:56 <****> All it does is applies a delta
23:57 <****> It gets very confused if you ever do more than a CVS-style single branch, edit, merge, forget.
23:57 <****> with ancestry
23:57 <****> that damn svn, always tricking me
23:57 <****> LeoNerd: svn's big problem is that *it* doesn't keep track of the merge info as it should, but that's fixed in 1.5
23:57 <****> the only other big thing is being able to rename files without having to ssh into the cvs server
23:57 <****> and I'm pretty sure everything but cvs does that
23:57 <****> Ya.. but any other current-generation RCS can do that also
23:58 <****> And bzr even does it in a way that doesn't suck like Git's way
23:58 <****> SVN allows for file/dir renaming, though it's kinda broken
23:58 <****> hello all, One thing I just noticed is that perl CPAN is not working...I'm getting the following error when I try to get a new package, reload the local index, etc:
23:58 <****> git just notices an add and a delete with the same MD5sum
23:58 * CPAN upload: Wx-0.78 (+++++) by MBARBON
23:58 <****> This therefore breaks if you rename _and_ edit a file
23:58 <****> A likely case in perl, for example
23:58 <****> jafari_: you're about to use a paste bot, right?
23:58 <****> Bar.pm involves editing the contents of the file as well
23:58 <****> LeoNerd: ugh
23:58 <****> Whereas, bzr tracks that, and Just Knows.. so it's fine to edit and rename at the same time
23:59 <****> And you can even merge a branch that contains edits to a file you locally renamed.. it doesn't care
23:59 <****> i don't know how mercurial does it, but i think it's similar to git in that respect
23:59 <****> yes
--- Log closed Tue Oct 16 00:00:14 2007
Total 32 pages. You are browsing page 32/32.
First :: Prev :: [...] [28] [29] [30] [31] [32] :: Next :: Last