| |
| |
| |
|
Page: 1 2
Comments:
<0> i believe the term you're looking for is "Space Ranger"
<1> heh
<0> ;-)
<2> Power Rangers > *
<0> heh
<3> is there a simple way to convert every first letter of every word to a capital?
<2> yes.
<3> oh sweet. what is it?
<2> Well, it is, if you have GNU sed.
<3> yeah I do
<3> screw that ancient crap, I'm running Gnu/Linux!
<4> shad0w1e: echo "abcd efg. hij klmno" |sed -e 's/^.\| ./\U&/g' #Abcd Efg. Hij Klmno
<3> gnubien, hey thanks
<4> shad0w1e: sure, its not perfect, its only a working kludge
<3> looks pretty damn perfect to me. I'll change whatever necessary manually. it's not that much stuff. thank you
<4> shad0w1e: have fun ;)
<5> i want to forward the output of ls to sed and find all the files that begin with a.how can i do it with sed?
<6> ls | sed -n '/^a.how/p'
<6> bye - gotta go
<6> l8r
<7> this may be google-able but i haven't figured out a good way to google it
<7> i want to be able to use sed to match only things that are inside or outside of dialogue in an ebook
<7> the theory is that the number of " should be even
<7> so one search I might want to only search outside of ""... I think I could figure out how to use sed to delete such sections
<7> but i'd prefer it if i could somehow mark them for later searching
<7> the problem being multilines? although i'm not sure if this is a problem
<8> lydgate, you might be better off writing your own little parser for this.
<7> Rado: really? would this be difficult?
<7> hmm ok
<7> so what's easier, figuring it out in sed or learning enough perl to do it
<7> ***uming no perl, only python and c as a hobbyist...
<8> lydgate, sed has only 2 "variables", not much for complex tasks like "counting".
<7> hmm ok
<8> But ... for learning purpose you could/should try it in sed.
<7> yeah, i'm going to at least mess with it
<8> At least I could learn from it when you find out :)
<7> i want to see if i can first eliminate all quoted p***ages from a text
<7> i'm sure sed can do that
<7> yeah so, turns out 46% of the brothers karamazov is dialogue :P
<5> Rado, what did you meant when you said sed has only 2 variables?
<8> To program stuff you need variables to store states, sed has only 2: pattern space and hold space.
<8> And both of them are just stupid strings, no arithmetics.
<7> hmm this doesn't work quite as well as hoped, apparently grepping for 2000 lines in 32000 lines is a slow operation
<7> grepping for 2000 patterns rather
<8> Ehh, why would you want to do _that_?
<0> sounds like it's time for a database with a join command
<7> yeah
<7> i'm being inefficient, still working out what the best way to do this is
<7> given an arbitrary novel which is good about using "" to indicate dialogue, I want to separate the narration from the dialogue
<7> then be able to search either half with regex, then find it again in the main text
<0> you could try sed '/"/,/"/d'
<7> yes that seems to work... although wouldn't that delete the whole line, not just what's between the ""?
<7> yes, and if there are two instances of a single word like this: "word" it will delete all the lines between them
<0> hrm, that's a good question. try it on a test file and see
<7> i think i got it, it's a bit more complicated
<7> from googling around, adapting a script that removes <> from html
<7> but it seems to work
<7> it's 9 lines but i can paste it if you're interested
<7> actually it works kind of beautifully
<0> nah, i'm into something else right now. thanks
<7> yeah no problem, thanks for your suggestion
<0> ;-)
<8> lydgate, /topic -> paste
<7> http://rafb.net/p/6aLdGN82.html
<5> lydgate, another approach a little bit more mathematic could be to work with cut
<7> marinosi: but how would that work with multiline quotes?
<5> for example cat _file_ | grep -c
<5> for example cat _file_ | grep -c '"'
<5> and then cut -f 1,3,..., `cat _file_ | grep -c '"' ` -1 -d '"'
<7> that would tell you how many quotes there were in the file...
<7> ok let me test it out
<5> i can make it for you if you wnat
<7> yeah i'm interested, i don't see how it would work
<5> i just need a little to test it
<7> ok i'll paste you some text
<7> marinosi: http://rafb.net/p/dhsyIK14.nln.html
<5> the problem is that grep counts lines that have the wanted char and not how many chars..ahmmm
<7> this is what the sed script gives: http://rafb.net/p/DDJJRl16.html
<7> which is actually pretty impressive
<5> lydgate i have to go now but i will prepare it for you later..Could you please give me your email to send it to you?
<7> sure
<5> it is challenge for me and it worths a try :D
<7> sed@lydgate.e4ward.com
<5> thanks
<5> :D
<5> when i come back i ll make it
<7> ok cool
<9> If I want to remove everything inclusively between <style.* and </style>, so I want sed '/<style.*/,/<\/style>/d' file > tmp && mv tmp file ?
<9> seems to work... Any suggestions for a better way?
<2> "Your sed might have -i."
<9> :P
<9> Does sed exit with nonzero exit status if it doesn't find a match in a file?
<8> kojiro, your "man sed" should tell you all exit codes.
<9> Rado: I tried searching for the word 'exit' or 'nonzero' or 'zero' but came up with nothing
<8> search no-case
<9> I also checked info sed, but didn't see a topic that applied. OK
<9> a case-insensitive search had two hits on 'exit'; --help and --version
<0> search for "exit
<0> "
<9> er...
<9> I am
<9> now I'm searching info sed
<9> I guess you have to explicitly use q and say which exit code you want?
<9> which makes sed like an inversible grep
<9> since it only accepts one address
<0> you are using gnu sed?
<9> yes
<0> ok, standard sed has a line on "exit" but i don't see anything in gnu sed yet
<9> beebum: in info sed I found 'q [EXIT-CODE]'
<0> oh, that isn't what you are looking for
<9> Maybe I should just state what I really want
<9> :)
<9> but it's kinda complex so I'll pastebin it
<9> http://rafb.net/p/cUMkgM56.html
<9> When you boil it down, that's what I need.
<0> that "...isn't the last InstanceEndEditable in the given document." part is beyond me but i can probably do the rest of it. :-/
<9> hmm
<9> if you nuke from <!-- InstanceEndEditable --> to the end of the file that's a problem :/
<0> no, it wouldn't do that
<0> it would just nuke every Begin/End range in the document
<9> err, no, I need to nuke from End to Begin
<9> I.e. closing up the gaps
<0> ok
<9> but only the ones inside the <body> -- the <head> is already finished
<0> ok, i think i can do that
<9> If that's easy for you, great, but I would just as soon learn it myself if you can show me... So far I've been doing it with pipes
<9> i.e. sed -n '/<body>/,/<\/body>/p' | sed ...
<0> use -e
<2> or ;
<9> right, but one starts with -n and the other doesn't
<2> kojiro: if what begins with <body> ?
<9> goldfish: I'm trying to close up editable regions with the <body>...</body> part of an html template
<0> kojiro, if you use -n you will not get all the other stuff you want, only the stuff you are trying to delete.
<2> I want to inclusively remove everything from '<!-- InstanceEndEditable -->'
<2> to the next '<!-- InstanceBeginEditable name="[^"]*" -->', but only when it
<2> begins within '<body>'
<2> so like <body> wold be the line above, <!-- Instance .... ?
<9> goldfish: right, only when <!-- InstanceBegin/End --> fall within <body> and </body>
<2> k
<0> brb
<9> goldfish: <body> could be many lines about <!-- Instance...
<9> s/about/above
<2> sed -n '/<body>/,/<\/body>/{/<!-- InstanceEndEditable -->/,/<!-- InstanceBeginEditable name="[^"]*" -->/d;p}'
<2> Won't work, for your "only if it isn't the last blah in file" eh, criteria, or whatever.
<2> And may not work for any other parts of it :) I didn't test it much.
<9> wow
<9> I actually understand that, now that you point it out
<0> i think it will work for the "last" part
<0> since you are going from end to begin
Return to
#sed
or
Go to some related
logs:
#suse
Unknown tag: %kmdl madwifi
VarStructor
#web
#php
Ubuntu mouse not workin
disable-pcmcia debian
sed literal new line
phprox
#math
|
|