| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Comments:
<0> Roderick, yeah that's what is baffling
<1> Can we see the code?
<0> Roderick, yeah i'm in the middle of adding a fork inside the sighup handler to test various things but sure
<0> I shouldn't need the fork but i've seen strange things )
<2> perl need a module called chopsticks that does some kind of forking. it would be funny
<0> http://rafb.net/paste/results/WgeBth10.html
<3> The paste WgeBth10 has been copied to http://erxz.com/pb/2355
<1> First thing: You aren't testing that the exec() worked.
<1> exec $0 or die "Can't exec $0: $!\n";
<0> Roderick, yeah that is the short version w/o the error checking :). I had that in there before
<1> Good news is it doesn't work for me either.
<1> If you do the exec() from outside of the signal handler it works.
<1> Perhaps since you never returned from the signal handler the kernel isn't delivering a duplicat signal to you.
<0> Roderick, yeah it seems very odd. seeing how exec is supposed to _REPLACE_ the rpocess
<0> Roderick, I sort thought the same thing which lead me to believe i'd fork and execute it in the child and exit the parent. giving me a new pid but still same problem
<1> Why not have the signal handler set a variable and the program's main loop notice it?
<0> Roderick, not opposed to that
<4> Hey all, sorry about the m*** spam, this will goto channels that wont be intrested as well (my fault) I wont be online for a while because im fixing some stuff on my bsd boxen, cut short im not opening up 6 server connections and 50+ rooms in irssi :) if you want me catch me on irc.undernet.org #magick
<5> I have a question about mod_perl - does anyone know how to dynamically require a certain user based on the url - for example if someone goes to htt://myhost.com/culkinj3 - I want them to have to auth as user culkinj3
<0> culkinj3, that should be in %ENV
<0> you'll have to parse the URL... atleast that's how i'd do it
<1> I think I know: While a signal is being delivered, it is blocked. Since the handler doesn't return, it doesn't get unblocked. exec() doesn't change the list of blocked signals.
<0> Roderick, hmmm interesting. I'm curious to see if I wrote this is C if it'd work
<1> So you could manually unblock it, or set the var and have the handler return.
<1> If you test it in C be sure to use a signal API which sets the signal to block.
<1> (sigset(), sigaction(), BSD signal())
<0> Roderick, the variable in loop approach definitely works though
<5> signuts, but how would I communicate that info to apaches
<0> culkinj3, perform the authentication yourself
<0> Roderick, if you're curious the C version works as expected :)
<1> Let's see it.
<0> http://rafb.net/paste/results/HbQpIU74.html
<3> The paste HbQpIU74 has been copied to http://erxz.com/pb/2356
<1> What's your OS?
<0> Debian testing/unstable 2.6.16.16
<0> Very odd indeed
<0> I don't care too much because I accomplished the same thing by setting the variable in the loop but it's rather intersting
<1> It isn't working, the execv() fails because you aren't giving it a list. If you use execl() it fails the same as the Perl version.
<1> I mean the 2nd arg to execv() isn't a char **, it's a char *.
<0> Roderick, ahhh
<1> I'm confident in the blocked signal explanation.
<0> Roderick, yeah I agree
<6> how would you trouble shoot an internal server error for a cgi - i tried fatals to browser, but it does not give me an error, and there is nothing in apache error log
<7> nothing in the log? that's pretty odd
<8> qt4x11: Is there anything in the access log?
<5> could anyone provide a quick example of an ldap auth in perl?
<9> hi. anyone familiar with stack based overflows?
<10> b0red: the quicker you get to your real, perl related question, the quicker it's likely to get an answer.
<6> nothing in access_log for that time
<8> qt4x11: So um, your server is all broken. Fix.
<6> genius advice thanks
<10> qt4x11: did you look in error_log? Are you sure you're outputting the proper headers?
<11> Hello pearl :)
<10> qt4x11: and what does this have to do with perl exactly? :)
<8> PerlJam: Apparently he isn't actually hitting the server.
<8> No access logs entries.
<12> Daveman=~s/a//
<13> hmm, it would be nice to have a GUID for each tuple in a database
<9> hi. anyone familiar with stack based overflows?
<8> Die in a fire.
<13> that's unfriendly!
<14> I'm familiar with stacks of flapjacks.
<14> Intimately.
<13> Plumbers fit those wee overflow pipes that dribble down the sides of buildings
<15> I am familiar with Qs.
<13> b0red: Are you taking a survey?
<15> buu: check this out...
<15> Copperbot: mjl69
<16> mjl69: MIKE JULIET LIMA SIX NINE
<15> for spelling things to dumb people on the phone
<12> Copperbot: tybalt89
<16> tybalt89: TANGO YANKEE BRAVO ALPHA LIMA TANGO EIGHT NINE
<17> mjl69: You need a bot for that?
<13> "NINER"
<8> heh
<17> NINER is ICAO.
<15> that's a good idea.
<11> tybalt :(
<17> It doesn't take too much effort to memorize the ICAO alphabet.
<12> Daveman: I was pulling the "a" from "pearl"
<17> At least include the alternate alphabets in the bot.
<15> Yaakov: not really. of course it's just a simple sub running in a simple bot.
<15> alternates?
<18> buu++ # "Die in a fire"
<11> tybalt, I noticed.
<17> mjl69: http://morsecode.scphillips.com/alphabet.html
<11> stealing my a's :(
<8> It seemed appropiate.
<19> he uses that a lot.
<19> 19:07 < buu> DIE IN A FIRE
<15> Yaakov: thanks.
<19> mainly because he is an unimaginative piece of ****.
<19> fire, how unoriginal.
<8> imagination--
<19> you--
<15> (organic memory)--
<8> perlbot karma you
<20> Karma for you: -7
<11> integration--
<8> heh
<17> mjl69: http://www.bckelk.uklinux.net/phon.full.html
<11> (genetic memory)++
<19> G'hualds!!!
<11> perlbot karma )
<20> ) doesn't have any karma
<11> Evan, learn how to speal kthx
<11> perlbot :o
<21> )++
<11> :)
<17> mjl69: for some reason that page doesn't render in Safari, but you can see the content in the source.
<18> now I'm imagining it with a thick southern accent: "Dah in uh far."
<11> perlbot karma $)
<20> $) doesn't have any karma
<11> perlbot karma $(
<20> $( doesn't have any karma
<11> aww
<8> cjeris: Con-air style?
<11> perlbot karma $$
<20> Karma for $$: 1
<11> :D
<21> perlbot karma <!
<20> Karma for <: -3285
<15> this could get pretty complicated. I see a possible new cpan module developing here...
<11> ROFL
<21> perlbot karma <!
<20> Karma for <: -3285
<18> buu: dunno, never seen con air
<11> perlbot karma !
<20> usage: karma (thing)
<21> perlbot, learn only to ignore trailing punctuation OR trailing spaces, and not both.
<8> =[
<11> :o
<11> BUU's a thing!
<8> Conair is awesome.
<11> buu the hairdryer or movie?
<8> Both.
<11> nah
<11> buu ****s
<21> "But I can write something better than infobot!" Sure, and you'll suffer the same interface learning period that infobot is suffering, but she's got a nine year lead.
<19> perlbot karma undef
<20> Karma for undef: -2
<15> Masque: are you programming perlbot in applescript?
<11> perlbot karma apple
<20> Karma for apple: 1
<11> apple--
<21> mjl69: In FUTURESCRIPT
<8> Masque: Been there, past that.
Return to
#perl
or
Go to some related
logs:
dvdcss for ubuntu
ubuntu gnome-foot
phparch gis
gentoo resolution horisync
rsync unkillable process
#suse
#kde
#perl
#centos
podcast downloader for firefox
|
|