| |
| |
| |
|
Comments:
<0> Does GNU Sed support shy * and + (in regexs)?
<1> shy?
<1> You mean "lazy quantifiers" ? No.
<0> Darn.
<1> Instead, you should say what you mean.
<1> don't use . unless you mean "match any character"
<1> If you don't mean *any* character, don't use .
<0> prec: well in this case I want to match <em>([^<]|<[^/]|</[^e]|</e[^m]|</em[^>])+
<1> Aha. Use an HTML/XML parser instead.
<0> That would be overkill.
<0> Apart from which, this is pre-processing BEFORE I apply an SGML parser, because the code is badly formed.
<1> Overkill? Then this should be fine: <em>.*</em>
<0> That would match <em><em>k</em>-term</em>, which is precisely what I'm trying to avoid.
<1> Well, the problem with something like ((x)-y) is that it is non-regular.
<0> Yeah, you're right.
<1> In general, you can't use a regex to parse a non-regular grammar.
<1> There are exceptions (because regexes support back references) but they don't apply here.
<1> Anyway, check the #regex FAQ suggests using XMLStarlet with sed. http://wiki.hypexr.org/wikka.php?wakka=RegexFAQ
<1> preprocessing with Tidy to convert poorly-formed HTML to XHTML.
<0> "converting HTML to XHTML using Tidy and then to PYX using XMLStarlet."
<0> I *am* using tidy; tidy gets it wrong; hence this preprocessing p***.
<1> Well, if you insist on using sed, you can do it. You can't do it in one regex of course, but it's possible.
<1> Since sed is (theoretically) Turing-complete.
<2> nearly anything still in existance is too
<3> hello, i want just print all character between the <table> and </table>, so i use sed -n -e '/<table>.*<\/table>/' to get that, but there are many <table>...</table> pair in my file, this command just can match first token and then exit. how can i let it match all token pair ? thank you.
<3> can anybody understand me ?
<4> how would i go about escaping the character "^@" which, if you can't see, is a caret followed by a commercial '@' symbol. i'm trying to replace it with the unicode for each of those characters.
<4> s/\^@/\@\^/g <num>
<4> .. is what i'm trying.
<5> I think you'll need to escape the ;
<4> goldfish: no i still get the "\^@" not found.
<4> similarly if i don't escape it of course.
<6> echo ^@^@ | sed 's/\^@/\@\^/g'
<6> @^@^
<6> or am I missing something here ?
<4> monsieur: hmm works here too, but it's not in vim.
<5> it's probably not a literal ^@
<4> monsieur: goldfish i appears in the document as a single character.
<4> s/i/it
<4> .. which to me implies that it is a literal character.
<4> monsieur: goldfish here's an example of the character in a text file. less will interpret it as a binary file, but it's not: http://www.selectparks.net/~julian/tmp/test.txt
<1> delire: Are you sure you don't have a NUL character? Vim will display this as ^@
<4> prec: ahah.. ****. that'll be it ;)
<1> od -a input |${PAGER:more} ## see?
<4> yeah it is a nul char. it's unsubstitutable then isn't it ;/
<1> more or less. You can use tr to change it to something else.
<4> cheers.
<4> prec: i can't see how you'd use tr as it's firstly not possible to tag this NUL char.
<1> tag?
<1> tr '\0' '@'
<4> oh
<1> <infile tr '\0@' '@\0' |sed 's/@/^@/g' |tr '\0@' '@\0' >outfile
<1> ***uming your sed is "clean"
<1> :!tr
<1> :)
<4> ;)
<1> You can do this substitution in vim anyway.
<1> type: colon percent s slash C-v C-@ slash caret at slash g enter
<1> :%s/^V^@/^@/g
<4> i'm going to need to break that down. what is the '^V' ?
<4> (i actually want to convert it to unicode chars @^)
<1> A % at the end of your prompt means that your running *csh, right? Right?!
<1> delire: type this characters: colon percent s slash C-v C-@ slash caret at slash g enter
<1> delire: C-v means hold Control while pressing v
<4> yes i realise
<1> delire: C-@ means hold Control while pressing @ (Control-Shift-2)
<1> delire: Or, some terminals let you type C-space for C-@
<5> nice.
<1> Also, you don't have to hold shift; C-2 will work just fine.
<4> it's expanding into ther characters, i'm going to have to play around.
<4> *other
<1> Escape, C-3, C-[
<1> C-\ and C-4 ; C-] and C-5 ; C-7 and C-g ; C-_ and C-/ ; etc
<4> on this host C-V expands to '^[OC'
<1> delire: hrm.
<1> My terminal won't let me type NUL.
<4> prec: hah works now (on another host)
<4> thanks
<1> I wonder what's wrong with my terminal? :(
<4> hehe
Return to
#sed
or
Go to some related
logs:
#perl
+ncurces +select list
can't call method on an undefined value at /usr/bin/cpan
pygstreamer
#gimp
install phpmyadmin to unbuntu
emerging by path is broken
gfxmenu resolution suse
dependency resolution failed packman
#gimp
|
|