| |
| |
| |
|
Page: 1 2 3 4 5 6 7
Comments:
<0> twkm, could you please just explain how curly brackets functin , or point out what doc to read
<1> time uses the parent shell's stderr, { } is an odd way of gaining access to those fd's.
<0> oh i see
<2> does this look better, for locking? report_and_fail is a function that emails a report and exits. http://pastebin.com/740605
<3> hi everybody
<2> hi
<3> i have a function called runcommand
<3> i now wanna write a wrapper function, which duplicates the output of runcommand and writes it to a logfile
<3> i did
<3> function runcomlogged ()
<3> {
<3> rm $1.log
<3> runcom "$1" "$2" | tee $1.log
<3> }
<3> but it's not writing to the logfile.
<3> whats my fault?
<4> lol
<2> seems like a weird way of doing it, what exactly do you want to log
<1> martoss: set -x is your friend.
<1> pastebin.com still isn't responding for me.
<2> should i paste somewhere else? its also slow for me
<5> http://rafb.net/paste is good too
<2> ok, going there
<2> http://rafb.net/paste/results/L7Dxrl25.html
<1> fbtab: don't remove the lock until the things that must not be done more than once at a time are complete.
<2> if it gets a signal, it will be exiting, so should it not remove the lock before that, incase of leaving dead/stale locks behind
<2> :w
<2> oops, sorry
<3> i dont't understand what set -x has to do with that.
<1> it should help you to find what went wrong.
<3> set -x seems to be for debugging purposes
<3> ok
<2> twkm: the removing of the lockfile is being set in the trap, so im thinking it should delete the lockfile before exiting..
<2> what do you think
<1> what if you exit successfully, no signals?
<2> im not trapping 0, but manually 'rm -f' the files at the end of the script
<1> ahh.
<4> faq updated.
<6> hi
<4> that long wiki page is stupid... we need some solution that can break it down into single question entries
<4> at least for editing
<2> GDG: hi
<1> most wiki's allow multiple pages. doesn't greycat's?
<6> I'm defining the PS1 variable in this way: PS1='[\e[33m\t\e[0m] \u@\H:\e[31m\w\e[0m \$ '... when it print path before the prompt it can take many column due to deep nesting in the filesystem... can I set a max length for the printed path adding something like [...] before ? for example [...]/deep/long/path/
<4> and this wiki markup junk makes me sick
<4> GDG: use a script to create the prompt and stick it in PROMPT_COMMAND
<7> i think he wants something like mediawiki, where each header etc creates an editable section, whereas the wiki of choice for the bash faq only permits editing the whole page
<4> that way you can do whatever
<4> Fatal: i want a faq-o-matic that uses html bodies :P
<1> !faq prompt
<8> http://wooledge.org/mywiki/BashFaq#faq53 -- I have a fancy prompt with colors, and now bash doesn't seem to know how wide my terminal is. Lines wrap around incorrectly.
<9> twkm: 53. I have a fancy prompt with colors, and now bash doesn't seem to know how wide my terminal is. Lines wrap around incorrectly.
<9> twkm: http://shelldorado.dyndns.org/wiki/index.php/BashFaq#faq53
<1> greycat: doesn't your wiki allow multiple pages?
<1> r00t|vaio: yeah, except that they all seem to get raped for xss flaws every other day.
<10> if you mean editing parts of pages, no, unfortunately.
<11> twkm: make them read the faq entry on eval
<1> i mean, multiple pages, so that the "main" faq page would be shorter, perhaps with just the q's.
<7> imho it's good to have a _big_ faq, i fetch solutions from the faq every now and then, but rather i search for part of the solution, i almost never come up with the same question as it's stated in the faq
<10> sure, it could be split into multiple pages. but someone would actually have to do it.
<1> heh.
<11> greycat: why "someone"?
<1> yup, ##c's wiki suffers from organizational apathy -- mine, mostly -- too.
<11> i'd say... "someone's computer"
<10> r00t: because FAQs are notoriously unwilling to change themselves spontaneously
<1> speaking of ##c's wiki, i guess i need to get that second host going.
<12> Hello all :)
<12> I have a quick bash scripting question:
<12> Say I have a string: ZENO="c:\foo\bar"
<7> and i'd rather load the site with one request and a bigger chunk of data, rather than clicking around visiting maybe 5-6 pages before actually finding what i was searching for.. but that's maybe just me
<1> meZeno: ewww!
<12> No I want to replace the \ by / (backward to forward slash)
<12> ghaha ;)
<1> meZeno: parameter expansion.
<11> Fatal: the concept of optionally having both formats is just too hard for wikitards to understand
<10> string=${string//\\/\/} I think
<10> or something else equally ugly
<12> I tried this : echot ${ZENO///\//}
<1> i looked around a little for useful wiki's or faq engines for ##c -- nothing really smelled any good.
<10> meZeno: how did you come up with *that*?
<1> too many initial /'s
<12> well {ZENO//X/A} -> X replaced by A
<12> well {ZENO//X//} -> X replaced by /
<12> well {ZENO//\//} -> \ replaced by / (doesn't work ofcourse)
<12> so i tought:
<12> well {ZENO///\//} -> \ replaced by / (escaping it) ;)
<1> you need to escape with \
<10> / is not an escape-character. \ is.
<1> ${var//\\//}
<10> 08:19 greycat> string=${string//\\/\/} I think
<12> ${ZENO//\\/\/} doesnt work ..
<12> it doesn't give errors, but I get the original string back
<10> try twkm's, then. he's usually smarter than I am.
<1> quoted the //\\/\/ does odd things. //\\// works quoted or unquoted.
<12> hmm.. doesn't seem to work :(
<10> imadev:~$ echo "$string"
<10> c:\foo\bar
<10> imadev:~$ echo ${string//\\/\/}
<10> c:/foo/bar
<10> imadev:~$ echo "${string//\\/\/}"
<10> c:\/foo\/bar
<12> bash-2.02$ echo ${zeno//\\//}
<12> C:\da\la\blaaaa
<12> bash-2.02$
<10> what's the original string?
<1> zeno vs ZENO?
<1> $ foo="c:\foo\bar" ; echo "${foo//\\//}" ${foo//\\//}
<1> c:/foo/bar c:/foo/bar
<10> also note he's running a horribly old version of bash
<10> nobody should be running anything other than 2.05b or 3.x without a good reason
<12> bash-2.02$ echo $zeno
<12> C:\da\la\blaaaa
<12> I have a good reason :)
<13> meZeno, just use sed, will ya :p
<10> imadev:~$ zeno='C:\da\la\blaaaa'; echo ${zeno//\\//}
<10> C:/da/la/blaaaa
<10> try it on a newer version of bash. if it works for you on the newer version, then it's time to upgrade.
<12> sed, isn't that like, only for files?
<13> not really
<12> I can't upgrade :D
<10> no, it's, like, for streams. that's why it's, y'know, a stream editor.
<10> meZeno: why not?
<1> yeah, prehistoric bashen aren't worth worrying about, but for meZeno being stuck there points at sed.
<12> If I tell you why.. you *will* kill me :D
<12> I am running bash from windows :)
<12> :$
<12> Why>
<12> well,
<10> oh. that's not a ban-hammering or ignoring offense.
<10> but you can upgrade. you're just unwilling.
<12> have some scripts from linux, which I am running on a windows machine for a short wile ..
<10> I'm quite certain cygwin's version of bash is later than 2.02.
<12> and, I couldn't really find any bash newer then this one, without breaking stuff..
<10> you "found" it? gah... there's your first problem.
<12> =)
<12> ghaha
<12> Cygnus :D
<10> where did you "find" it? a moldy trash heap?
<12> Byt, ehm.. any ideas on how to do this thing with sed?
<13> $] var="\\" ; echo "$var" | sed 's,\\,/,g'
<13> /
<12> I found it @ the junkjard ;)
<1> yes, same way, more or less.
<10> you either install cygwin, or you build bash yourself using whatever toolkit is required for your environment.
<10> you don't just use a random binary that you found lying around in the seediest corner of the internet
<13> that's a good advice :P
<11> btw: it's a great feature how in any non-css browser the edit box on the wiki comes up about 16 characters wide...
<11> (just that it's too long to edit in a browser anyway, so i just use one that can spawn an external editor for textboxes
<10> that's why I use mozilla/firefox with the mozex extension
Return to
#bash
or
Go to some related
logs:
ubuntu jstest
mplayer devfb flicker
hp scanjet 4670 ubuntu
#debian
#osdev
+Kmix +OSD +disable +suse
#php
backuppc Xlib: connection to
(ext2ifs OR ext2dfs) smbfs
#web
|
|