Help with reg expressions

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
(Ovid) Re: Help with reg expressions by Ovid (Cardinal) on May 03, 2001 at 20:27 UTC
You'll need to be more specific. First of all, you stated that the common name was in the form "(something)-disclose.gif". Do you really have parentheses in the common name? If so, you'll need to escape those in the regular expression. Further, what's allowed for the "something"? Is it all letters, are numbers allowed? Is punctuation allowed? First advice: ignore what the Anonymous Monk said about using the dot star combination. It's slow and virtually guaranteed to break your code. See Death to Dot Star! for an explanation. Here's a rough stab at your regular expression, assuming that the "something" cannot contain a dash: `# also assumes that the parentheses are not in the filename /<img # match the '<img' \s+ # must have one or more whitespace characters src # match the 'src' \s* # may have zero or more whitespace characters = # match the = \s* # may have zero or more whitespace characters [^-]+ # one or more non-dash characters -disclose.gif # rest of .gif name [^>]* # zero or more non-closing angle brackets >/ix # closing angle bracket` [download] Cheers, Ovid Update: To whomever the Anonymous Monk is who criticized my regex: sign on and join us! That was a great catch and we could use more programmers like you! Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.	[reply] [d/l]
Re: (Ovid) Re: Help with reg expressions by srawls (Friar) on May 04, 2001 at 00:57 UTC
>>To whomever the Anonymous Monk is who criticized my regex: sign on and join us! That was a great catch and we could use more programmers like you! Thank you, I very much appreciate that. As you can see, I have now signed on. The 15 year old, freshman programmer, Stephen Rawls	[reply]
Re: Help with reg expressions by chipmunk (Parson) on May 03, 2001 at 20:21 UTC
This will probably do what you want, but I'm making assumptions that your HTML doesn't contain unusual things like >s in the attribute values: `s/<img [^>]src=[^>]+-disclose.gif[^>]>//ig;` You should be careful about using regular expressions to parse HTML, however. Using a proper parser, such as HTML::Parser, will make your script much more robust.	[reply] [d/l]
Re: Help with reg expressions by astanley (Beadle) on May 03, 2001 at 20:23 UTC
Well if (something) is all the pattern you can provide for us then really the only regex match I can think of is .. However based on your description I can't tell if you have parantheses in the filenames or not. I'll give you a regex for each situation. `if (/\<img src=$(.)$-disclose\.gif.\>/i) { $_ = "" }` [download] That will work if the filenames have parantheses. The following will work if they do not: `if (/\<img src=$.$-disclose\.gif.*\>/i) { $_ = "" }` [download] After the regular expression the variable $1 will contain the name between (something). (ie: `print "$1-diclose.gif\n";` will give you the filename) WARNING: the regexp's are untested but should give you the idea! (in a regex putting a match between () assigns the variables $1,$2,$3...) -Adam Stanley Nethosters, Inc.	[reply] [d/l] [select]
Re: Help with reg expressions by Anonymous Monk on May 03, 2001 at 23:47 UTC
ovid offered some good advice, all but for one thing. Please, DO NOT USE THIS LINE FROM OVID'S REGEX: `[^-]+` [download] take the folowing line for example, `<img src="this.gif"><img src="something-disclose.gif">` [download] what do you think his regex will match ... think carefully--it matches the whole line. I recomend reading the book by Jeff Freidel (sp?) called Mastering Regular Expressions. In it, he stresses to say what you really mean. You should replace this line: `[^-]+` [download] with this: `[^->]` [download] The reason is that ^- means anything but a dash, well the > sign isn't a dash, so his regex will match that, all the way to the first -, witch could be in another image tag on the same line. The 15 year old, freshman programmer, Stephen Rawls	[reply] [d/l] [select]
Re: Help with reg expressions by DeusVult (Scribe) on May 04, 2001 at 01:23 UTC
Ok, folks, I think some people are getting a little paranoid with the questions. Now, nearly every post asked for a clarification of the "(something)-disclose.gif". Now, we all know what happens when we assume, but really...this is a filename! I don't know what the hell kind of crazy operating systems you people are using, but have you ever seen any that allowed parentheses in a filename? Methinks y'all are just getting a wee bit persnickety. In any case, I think we can assume that the (???) can be safely replaced by `/\<img src=[a-zA-z0-9\.\_\-]...` [download] Of course, anything can happen. One guy at work somehow ended up with a file named * on his Solaris machine. Yes, that's right, the file is named *. Of course, its a useless file, but he sure as hell doesn't want to delete it... And somehow another guy ended up with a file named backspace. No, not the word backspace, the name of the file is actually the single character backspace. He hasn't tried to get rid of that one either. But if you have any files like that, I suggest you don't trust it to an automated script to get rid of. If you have any trouble sounding condescending, find a Unix user to show you how it's done. - Scott Adams	[reply] [d/l]
Re: Re: Help with reg expressions by chipmunk (Parson) on May 04, 2001 at 02:20 UTC
I don't know what the hell kind of crazy operating systems you people are using, but have you ever seen any that allowed parentheses in a filename? If you have any trouble sounding condescending, find a Unix user to show you how it's done. - Scott Adams Does that answer your question? ;) Oh, and under MacOS, it's even easier to use filenames with parentheses in them, because you don't have to worry about escaping them.	[reply]
Re (tilly) 2: Help with reg expressions by tilly (Archbishop) on May 04, 2001 at 04:35 UTC
Actually if you have a filename like that, I would use Perl to figure out how to name and delete it. BTW you can also use ANSI control characters in file names. That is a fun one...	[reply]
Re: Re: Help with reg expressions by elwarren (Priest) on May 04, 2001 at 04:25 UTC
Agreed. Just look for any tag containing disclose.gif and remove the entire tag. If you want to get more indepth then write a grab anything between src= and disclose.gif. Sheesh, really not that bad.	[reply]
Re: Help with reg expressions by srawls (Friar) on May 04, 2001 at 00:31 UTC
Hi, I replied earlier with the ^-> negated character class. I just noticed your code at top, you might want to consider replacing this: `if (/\<img src=(???)-disclose.gif(.)\>/i) { $_ = "";` [download] with this: (note: the improved regex is inserted below) `lc; #Much more efficient than having the regex do it s/\<img\ssrc\s=[^->]-disclose.gif[^>]>//g; #more effecient to use a substitution` [download] Please note that I used the lc function, because having the regex perform case insensitive searching is a lot of overhead, also I changed your $_ = "" to a substition, which is also more effecient. Also, if you had more than one img tag on a line before, the whole line would have been deleted if even one of the tags had the -disclose.gif text in them. Also I changed the end of the regex. Instead of matching the rest of the string with `(.)\>` [download] I used the negated character class ^>. Again, the . is greedy, so it could have matched other img tags on the same line, for that mater, it could have matced anything on the same line. The 15 year old, freshman programmer Stephen Rawls	[reply] [d/l] [select]
Re: Re: Help with reg expressions by Anonymous Monk on May 04, 2001 at 00:36 UTC
Whoops, major type. In this line I submited: `s/\<img\ssrc\s=[^->]-disclose.gif[^>]*>//g;` [download] You need to change this: `[^->]` [download] to this: `[^->]+` [download] Sory About that. The 15 year old, freshman programmer Stephen Rawls	[reply] [d/l] [select]
A reply falls below the community's threshold of quality. You may see it by logging in.