| |
| |
| |
|
Page: 1 2
Comments:
<0> how can I replace a bunch of spaces with a single comma?
<1> s/ */,/
<0> prec: thank ya :)
<1> Sure.
<2> hi. anybody know how to reference a substring that's matched by a regexp ?
<2> like perl's $1 $2 stuff
<1> s/\(foo\)\(bar\)/\2\1/
<2> should i be using grep if i'm not doing a substition at all?
<2> i just want to extract a portion of text from a line
<2> but then grep prints lines...
<1> GNU grep has -o
<2> sounds like that would do.
<2> so it's not a job for 'sed' though?
<1> one sec.
<2> actually. what i need to do is. like perl's =~ /VERSION=(.*)/
<1> http://www.rafb.net./paste/results/aFx9Yg76.html
<2> so i get what's after VERSION= as $1
<1> Everything is after VERSION=
<1> sed -e '/VERSION=/!d;s///'
<2> ok. i wasn't being fair.
<2> hold on
<2> export LTSP_KERNEL_VERSION="2.6.16.17-ltsp01"
<2> this is my string
<2> i need the 2.6.16.17-ltsp01 part
<1> sed -e '/^LTSP_KERNEL_VERSION=/!d;s///;s/"//g'
<2> returns nothing
<1> sed -e '/^export LTSP_KERNEL_VERSION=/!d;s///;s/"//g'
<1> oops.
<2> so. it looks like sed uses \(.*\) to match regions. then refers to them as \1 ?
<2> so i would match and replace the whole line with the match only ?
<2> i.e do a substition although i don't need to
<1> Hmm?
<1> sed -ne 's/^export LTSP_KERNEL_VERSION="\([^"]*\)"$/\1/p'
<1> That will work too.
<2> sed -e 's/export LTSP_KERNEL_VERSION=\(.*\)/\1/'
<2> this works. though i don't know why it handles the double quotes
<1> Handles?
<2> first things first
<1> sed doesn't care about double quote characters -- they are just ordinary characters to sed.
<2> sed -ne 's/^export LTSP_KERNEL_VERSION="\([^"]*\)"$/\1/p' didn't work
<2> second.
<2> sed -e 's/export LTSP_KERNEL_VERSION=\(.*\)/\1/'
<2> works. but i was expecting it to return "2.6.16.17-ltsp01"
<2> but it returned 2.6.16.17-ltsp01
<2> which is what i want. but. how does it know that. i don't get.
<1> It doesn't.
<2> oh
<2> it must be a shell substition thing
<1> That'd be my guess.
<2> right
<2> so i'll try your version again without involving the shell
<2> works
<2> this is actually the output of a command but i was just echo'ing a constant string for testing
<2> mine works too when i add the double quotes
<2> can you explain the difference ?
<2> [^"] what does that mean?
<2> at the beginning ?
<2> or negation?
<1> Without invoking the shell? Hmm.
<2> nah.
<1> [^"] matches a single non-quote character.
<2> oh and * matches all single non-quote characters
<2> until the double quotes at the end
<2> ?
<1> No, * does not match.
<2> quantifier no?
<2> 0 or more
<1> Yes. * is the same as \{0,\}
<2> awesome.
<1> Which means "match zero or more of the preceding item"
<2> ok. just like perl. great.
<2> thanks very much. i think i've got enough to work with.
<1> OK. :)
<2> later
<3> Evening.
<3> What I need is a math for a string which is a max of 4 chars long and starts at the line.
<3> Further more it only the number in one line.
<3> I have gotten this.
<3> s/^[0-9]\{1,4\}//p
<3> I just want to print the match
<3> Nothing else.
<4> WoodyWoodpecker: sed -n '/^[0-9]\{1,3\}/p' filename
<3> gnubien: But 1,3 isn't that a maximum of 3 chars?
<4> WoodyWoodpecker: ok, try 1,4
<3> I didn't try, I just wondered.
<3> It it not mine anyway.
<3> But thank you anyway.
<4> have fun ;)
<3> lol, sure I will ^^
<5> HI all.can sed use with environment variables?
<5> for example,sed -e 's/$USER/fog/' ...
<6> yes, you can
<6> by using the right quotes to allow your shell to substitute the variable
<6> echo $USER; echo "$USER"; echo '$USER' - not sed's fault
<7> hum....how can i delete the first line in patten space which have mutil lines content,and don't use "D" ?
<7> thx
<4> giant: post an example line you want to delete
<7> gnubien:i ***ume there are 3 lines in my partten space now
<7> first line:1
<7> second line:2
<7> third line:3
<7> and i wanna del the first line
<7> how to do it ?thx!
<4> giant: line 1,2,3 are all one long line?
<4> giant: line 1,2,3 each ended with a newline char?
<7> each number occupys one line in mpatten space.
<7> yep
<7> how to express the "newline char" if i use s(earch) command to patten space ?
<4> newline == \n like echo -e "1\n2\n3\n"
<8> regex q: how to replace string part after last 'x' occurance wih 'y'? i.e. asxfooxbar -> asxfooxy
<9> s/^\(.*x\)[^x]*$/\1y/
<4> MZM: echo "asxfooxbar" |sed 's/^\(.*x\)[^x]*$/\1y/' #asxfooxy
<9> giant: if you have all three lines in pattern space (N;N) then you delete first like this: s/^[^\n]*\n//
<8> tnx
<9> echo -e "1\n2\n3" |sed 'N;N;s/^[^\n]*\n//'
<9> echo -e "1\n2\n3\n"
<7> helloman:hum...
<9> clear?
<7> yep,but...not work.
<9> what doesn't work?
<7> wait for a moment .
<7> helloman:your script seems delete all lines except the last line.
<7> it will find the first "\n" and delete the line,the continue to find the second "\n" and delete the line also......
<9> what exactly do you want to do with those three lines?
<7> just delete the first line!
<7> delete "1\n"
<9> well, sed '1d'
<7> ...
<9> it is in pattern space, well?
<7> yep,these 3 lines are in the pattern space
<9> then s/^[^\n]*\n// must work
<7> hum...do you konw which site provides the similar service likes pastebin.com?
<9> every free webhosting :)
<4> giant: pastebin.ca
<7> get it
<7> http://cpp.enisoc.com/pastebin/7393
<9> wau
<9> what do you want the result to be?
<7> i paste it just wanna explain that your script search "\n" in lines one by one,instead of delete all lines between "^" and the last "\n" once .
<7> i paste it just wanna explain that your script search "\n" in lines and delete the line one by one,instead of delete all lines between "^" and the last "\n" once .
<9> but if you delete all lines between ^ and last \n you do not delete first line as you wanted, but everything but the last one
<9> if you have "1\n2\n3" in pattern space s/^[^\n]*\n// results in "2\n3"
<7> ok,i didn't discribe it clearly.sorry for my lame english.
<9> mine is alse lame
<7> e,but seems not get the result in "2\n3"
<9> also
<7> the result is "3"
<7> PATT:2
<7> HOLD:
<7> 1
Return to
#sed
or
Go to some related
logs:
totem-exine
#centos
perl use lib relative
Failed Running automake! Failed Running automake!
ldapsearch scandinavian characters
#suse
Can't locate object method finsh
#linux
xorg-7 + tgz
#lisp
|
|