#perl
29 September 2007
Total 19 pages. You are browsing page 2/19.
First :: Prev :: [1] [2] [3] [4] [5] [...] :: Next :: Last
01:05 <****> search.cpan.org has been updated, neat
01:06 <****> it looks the same!
01:07 <****> gullible is misspelled on merriam webster website?
01:07 <****> Yeah
01:07 <****> I just corrected it.
01:07 <****> as long as it is official
01:09 <****> /close
01:10 <****> GumbyBRAIN: confuzzled?
01:10 <****> udderly confuzzled.
01:11 <****> GumbyBRAIN: moo!
01:11 <****> Khisanth: nothing major... they just added photos of authors and a little box with the number of bugs and a link to RT
01:11 <****> oh that, already saw it!
01:12 <****> :)
01:12 <****> People we abandon you regrettably we ðåàë áóõíóëè and we smoke go since we from Russia
01:13 <****> bugs? can somebody help me fix a bug in net::ping? i never made a update to a cpan modul. but i know the error
01:13 <****> Not ðåàë we from ZHukovskogo there where max passed
01:13 <****> International aerospace salon so íåòóïèòå))
01:14 <****> bugs? can somebody help me fix a bug in net::ping? i never made a update to a cpan modul. but i know the error
01:14 <****> sorry
01:15 <****> BoaThor99: if you have a patch you can submit it to rt
01:16 <****> Caelum Patch this are this diff-file stuff?
01:16 <****> Caelum: at best practical?
01:17 <****> kspath: rt.cpan.org
01:18 <****> Caelum: I figured. Just my baaad sense of humour
01:18 <****> Caelum: sorry
01:19 <****> your.patch
01:23 <****> Timing/race condition question. I keep an ordered queue of timers, based on when they'll expire. I enter a poll(), with a timeout based on "next time" - now, and when poll returns, see if the next timer event can go. But sometimes, it fails; poll() returns 0 but very very slightly early, so there's just a few microseconds between "now" and when it should expire. What's the best strategy to deal with this..?
01:25 <****> LeoNerd: do not use microseconds, round?
01:26 <****> LeoNerd: you will likely always have that issue no matter
01:27 <****> LeoNerd: sorry, did not think it through before saying anything.
01:27 * kspath back to sleep
01:27 <****> LeoNerd: do you really need microsecond resolution for your timers?
01:27 <****> leonerd time::hires?
01:28 <****> here, there is girls?
01:28 <****> I'm already using Time::HiRes
01:28 * CPAN upload: Perl-Configure-0.05 by MSCHILLI
01:29 <****> leonerd time handling is not very precisly it depends on many things ... for example kernel load... icp states
01:29 <****> Yes yes, I'm aware of this :)
01:31 <****> LeoNerd: no matter your criteria it will always be a question of the impact of how greedy you want it to be.
01:31 <****> LeoNerd: is that vague enough to be an answer?
01:31 <****> GumbyBRAIN: Am I vague enough?
01:31 <****> is that vague enough to give a variable i can read.
01:31 <****> LeoNerd it´s a pain...
01:31 <****> anyone been unfortunate enough to use Perl with IIS/ASP ?
01:32 <****> sort of curious if it's worth it or not
01:33 <****> LeoNerd: why not expire it next time and check next again
01:33 <****> kspath: It's in a linear test script
01:33 <****> Real program code sits in a while(1); loop... that's fine
01:33 <****> But this is a linear test script, starts at the top, does things, finishes.
01:34 <****> I think a loop in my test script is the only sane way forward
01:34 <****> That way, the test script most closely resembles what a real program would look like
01:36 <****> LeoNerd i use epochen as hash key and lokk every round if the timestap matches timeor is older...
01:36 <****> LeoNerd: each timer isn't aware of other timers?
01:36 <****> quatrix_: I think he said an ordered queue of timers
01:37 <****> quatrix_: I read emphasis on the word queue
01:37 <****> Yes, it's a queue
01:37 <****> (well, actually it's a heap, but never mind)
01:38 <****> GumbyBRAIN: heap my queue over to the core dump
01:38 <****> I mean, how, if it were a superstar html monkey, and you don't get my perl dump core.
01:38 <****> BoaThor99: Yes, this is my problem.
01:38 <****> in the heap of the night... great song
01:38 <****> Example: I want to exprire in 2000ms time
01:38 <****> So, I store now+2000ms in the queue. Then poll() for 2000ms.
01:38 <****> OS isn't exact.
01:38 <****> Maybe, when poll() returns 0, and I gettimeofday(), I'll only get 1999ms more than when I started.
01:39 <****> It's an ms behind. So I don't run the event.
01:39 <****> LeoNerd: what happens if ntpd updates -512ms ?
01:40 <****> Then most programs fail
01:40 <****> (Not just mine ;) )
01:40 <****> LeoNerd: expired ones are gone so they do not come back?
01:40 <****> Anyway, ntpd doesn't make big jumps
01:40 <****> Any timer whose fire time is in the past, gets run
01:41 <****> omg
01:41 <****> scheduledtime)
01:41 <****> GumbyBRAIN: holy timers batman!
01:41 <****> You have multiple timers for a batman vs predator movie.
01:41 <****> BoaThor99: Yes, that's what I'm doing
01:41 <****>run; }
01:41 <****> (basically)
01:41 <****> Problem is that sometimes, just sometimes, OS's timer isn't perfect
01:42 <****> i'm still trying to get the problem, the problem is that if a timer ran, and finished too soon, it will run again?
01:42 <****> instead of the next timer?
01:42 <****> No
01:42 <****> The timer runs 0 times, instead of the 1 time I was expecting
01:42 <****> Example: now is 1400.000
01:42 <****> Leonerd check time twice^^
01:42 <****> I want it to expire at 1500.000
01:42 <****> So, I poll() for 100.000
01:42 <****> poll() returns 0 - timeout
01:42 <****> I gettimeofday, and find 1499.999
01:42 <****> Because OS isn't perfect.
01:43 <****> So my queue says "oh, nothing ready yet, I'll sleep again"
01:43 <****> At which point my test script fails.
01:43 <****> LeoNerd: ChuckNorris runs ntpd but it never updates out of fear of contradicting ChuckNorris view of static ChuckNorris Universal Time time
01:43 * CPAN upload: Class-Root-0.01 by NIF
01:44 <****> leonerds THEN do if (scheduledtime == time or scheduledtme == time-$x or scheduledtime == time+$x )
01:44 <****> Hrm... espilon
01:44 <****> I hadn't pondered that idea
01:44 <****> HA I RULE AGAIN!! HAHAHAHHAHA
01:44 <****> ok
Total 19 pages. You are browsing page 2/19.
First :: Prev :: [1] [2] [3] [4] [5] [...] :: Next :: Last