| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> But I guess it'll be a good read some day
<1> Dragon Book?
<0> Especially when I'm on that topic at uni again
<0> Teckla: Sure, don't know it? It's about writing compilers
<1> Oh, cool.
<1> Sounds interesting.
<2> Yeah, I've been meaning to get it for a long time
<2> I got it for 27$ used off the amazon marketplace, in great condition, so I'm very happy :D
<1> Does it also discuss interpreters?
<1> Doh. A lot of times when I install FreeBSD packages it ends with "Broken pipe". I wish I knew if that was good or bad.
<2> Teck: it dicusses VM based interpreters to some detail
<1> Humanoid: Very cool.
<3>
<3> heh
<4> broken pipe sounds bad.. it means data was lost. i suppose this is only not bad if you didnt need the data anyway, but in that case, why wasn't it drained to /dev/null?
<5> anyone know where i can lookup CIDR/s by country?
<6> yeah
<6> hrm
<1> AaronWL: I agree.
<1> I do get this "broken pipe" reasonably frequently when running pkg_add.
<5> thats very bad heh
<4> i dont know anything about freebsd.. if its a shell program, you might be able to find it by turning shell tracing on
<4> if not, whatever freebsd's equivilent of strace is, following child processes, should tell you which process it was that got killed
<7> rip: it doesn't really work that way.
<1> I think it might be because it's fetching packages remotely.
<8> Broken pipes spill data all over your file system!
<1> I guess I'll have to install the Brawny Paper Towel extension.
<9> in a makefile, I'd like to use -ggdb3 if my system has one hostname, and not if I use another
<9> or, I'd like to, by testing the hostname, select a certain set of compile-flags
<9> (same thing I'm asking)
<7> man that sounds stupid.
<10> my nick?
<7> the makefile thing.
<10> oh.. I'm doing development and repeatedly copying things over to the server once tested..
<10> Makefile goes too
<10> I'd prefer to have the dev system automatically compile with debug on
<11> bleh.. whatever.
<4> jaggasket, you could use autoconf
<4> with autoconf, you can do interesting things like have a site-local autoconf config file
<4> other than that, you could use an environmental variable only on the dev machine
<4> to actually check hostname, you just need a bit of ugly sh script in your makefile
<4> `test [
<4> CFLAGS = -Wall `if [ x"\`hostname`" == xyourhost ]; then echo -g3; fi`
<4> theres probably a slightly less ugly gmake extension that does similar.
<11> and that sets the CFLAGS environment var... and that's "seen" by make? didn't know that'd work :)
<4> welll,
<4> depending on the make, it will usually p*** the entire thing to the shell at compile time, and the shellw ill evaluate it before gcc is executed.
<4> which makes for really ugly make output.
<4> to make it prettier, you need recursive make: <tab>$(MAKE) -C whatever CFLAGS=`whatever`
<11> ohh
<11> interesting
<4> recursive make is useful, but unfortunate in that it has a performance penalty
<4> eg, if nothing needs to be done, it will still recurse.
<11> nod
<11> interesting though
<11> say, I do love printf("%.*s", len, pointer);
<12> gotta love very, very compiler dependent machine dependent code...
<12> http://lxr.xensource.com/lxr/source/linux-2.6.16-xen/arch/mips/mm/c-r4k.c#L1002
<12> kernel code is so much fun
<5> __asm__ __volatile__("nop" : : "r" (*p)); /* whee... */
<5> erm sorry abut that
<4> thats weird
<5> wasnt mean for this window ;]
<4> what purpose does that *p clobber have???
<5> ask the kernel guys
<4> wait, thats not a clobber.. its an input.
<13> I am having problems with sscanf
<4> it basically says "load *p into any register, any one at all, and then call 'nop'"
<13> http://rafb.net/paste/results/FTtDRb11.html
<13> basicly, given the code shown, I get the mentioned output
<13> anyone know what I could be doing wrong?
<4> it was probably to work around a gcc bug
<13> compiler is Visual Studio .net 2003 if that makes a difference
<4> jonwil, for starters, examine sscanf's return value
<14> jonwil: %d expects an int*. You're p***ing unsigned char*.
<13> I didnt even know sscanf HAD a return value :)
<13> what can I use for unsigned char then?
<4> Xirtam > *
<14> Check your documentation for sscanf(), I'd just be guessing.
<13> thanks, for whatever reason, changing from %d to %c for the 4 char variables worked
<12> whoa. jonwil came in here?!
<12> I didn't even know he was on EFnet
<5> in soviet russia, the efnet is on you.
<12> and pumping hard
<5> i guess
<5> ;]
<5> how good is kqueue?
<15> the best
<12> AaronWL: apparently what it does it fetch data from that pointer into something or other, which gets it into the cache as well
<12> that routine is a cache size detection system for one of the MIPS processor series
<12> which is why that comment up at the top is interesting... any pushes or pops are going to get into the cache while it's trying to detect stuff about it, and will foul up the results
<11> ttyl bubs..
<11> bye aaron!
<11> and twkm!
<16> With changing IPs of computers, what's best solution for p***word-less ssh ?
<17> Aeon_: p***word-less ssh? is your goal to get hacked?
<17> Aeon_: though there's dynamic dns services. although i don't know if there's one you can find that's free
<18> ssh-agent
<5> ssh-agent, or just using keyubased auth
<16> i use expect-based scripts for that
<16> i am not paranoid about 'being hacked
<17> it is fairly common for bots to randomly test ip addresses for ssh access :-)
<16> they casn randomly try it till end of eternity
<16> Alipha: i don't think you understood my question, ok ?
<4> i use a rsa/dsa key ...
<4> that means the remote host won't ask for a p***word. is that good enough?
<16> AaronWL: it won't ask even if IP changes ?
<4> nope
<4> that is, on each host you want to log into with rsa, you create a .ssh folder, and put your public key in it
<4> sshd takes care of the rest.
<4> your local ssh just needs a .ssh in your ~, that has your private key, and you're good to go
<4> the file you need the server to have, specifically, is ~/.ssh/authorized_keys
<16> okthanks
<16> and i need to put server's public key into some file in client's ~/ssh ?
<16> ~/.ssh ?
<4> you need to have a public key and a private key
<4> you generate these with ssh-keygen or whatever
<16> that is part of ssh installation, no ?
<4> for dsa keys, these are ~/.ssh/id_dsa and ~/.ssh/id_dsa.pub
<4> no, you need to generate them
<16> ah, ok
<4> run ssh-keygen
<4> you want an id_dsa and id_dsa.pub (private and puiblic, respectively), in your localhost's ~/.ssh
<16> the per-user keys ruins my mood
<16> I want user-agnostic authentication on bith-sides
<16> I user at least 2 users on each sides
<4> well you just need to copy the key to wheerever you want to log in from
<4> and you put a copy of the public key, manually, on each host you want to log in to.
<16> Right now, I use expect-based scripts which feed p***word automatically and it works perfectly for me
<16> I see
<4> if you want to log into multiple users, you need a .ssh authorized_keys for each user you want to log into.
<4> (it is exactly the same as the old-style rsh)
<16> Oh that's good idea
<16> SHall I have p***phrase for the user key ?
<4> i don't.
<16> YOu don't use p***phrase ?
<4> no.
<4> as long as they don't get your id_dsa file, they cannot do anything.
<16> okthanks
<4> the p***word is for extra encryption on id_dsa, so someone else on your local computer cannot pretend to be you.
<4> if you want to use a local p***word, you can still avoid typing it by using ssh-agent, but gets more complicated.
<16> is it ok to use same user key of different computers ?
<4> yes
<16> do you use it ?
<4> i use the same public key for rasterburn, rm-f, my school, sourceforge, etc
<4> thats id_dsa.pub .. anyone can see it. id_dsa is the secret one.
<4> or id_rsa, if you're using rsa
<16> and on server-side again, i put it into ~/.ssh/autorozed_keys ?
<4> right. you copy the line from your id_dsa to put into .ssh/authorized_keys
<4> this is mine.. should look like this:
Return to
#c
or
Go to some related
logs:
#computers
#stocks
xequtors mp3
#freebsd
dell resize partition 9400
9000 SUX robocop
#computers
#computers
it should be a crime to be that stupid
burn jello
|
|