regex behaving greedier?

DrChiper · 04-16-2014, 03:42 AM

I try to remove unwanted "span's" of the form:

Code:

<span class="font1 font2 something">interesting text</span>

and used to use the following regex in find:

Code:

find: <span(.*)>(.*)</span>
replace: \2

However, this no longer works as now *everything* is found up to the terminating <crlf> or until a second terminating when this is found earlier just before the terminating <crlf>.

It seems that regex behaves now greedier (while dotall is *not* enabled) then before. What am I missing?

jbacelar · 04-16-2014, 03:59 AM

find: <span.*?>(.*?)
replace: \1

DrChiper · 04-16-2014, 08:31 AM

@jbacelar: Thanks, but no, that is not the solution. A combination of '*' (zero or more) and '?' (zero or one) together does not make sense here.
I just tried my original regex within calibre 1.31 and it works as intended. However, it does not seem to work in calibre v1.32, which is odd.

Skeeve · 04-16-2014, 08:37 AM

I don't get what you want to expres... Could you post some examples?

kovidgoyal · 04-16-2014, 08:46 AM

* - means match zero or more matching as many as possible. i.e. be greedy

*? - means match zero or more matching as few as necessary, i.e. dont be greedy

Skeeve · 04-16-2014, 12:04 PM

If I understood correct, he's saying that

Code:

<span(.*)>(.*)</span>

would also match

Code:

<span>something<p><crlf>

DrChiper · 04-16-2014, 12:08 PM

I checked and it appears the text is/was very bad formatted in which is sometimes missing, thus causing the weird behavior by selecting unexpectedly large sections. And that my original regex did worked in the past for other texts may just be pure luck

Oh well, at least I did learned a new regex trick and refreshed my knowledge about "Lazy quantifiers" versus "Greedy quantifiers". Thanks all!

DiapDealer · 04-16-2014, 01:24 PM

Just remember; using regex to remove only certain span|div tags will usually blow up in your face wherever spans|divs are nested.

eschwartz · 04-17-2014, 04:14 AM

The modify epub plugin is going in interesting places regarding junk spans. Might be worh a look...

Using a negative lookahead you can avoid nesting issues, there is an example by me in that thread.

04-16-2014, 03:42 AM	#1
DrChiper Bookish Posts: 1,049 Karma: 2006208 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	regex behaving greedier? I try to remove unwanted "span's" of the form: Code: <span class="font1 font2 something">interesting text</span> and used to use the following regex in find: Code: find: <span(.)>(.)</span> replace: \2 However, this no longer works as now everything is found up to the terminating <p><crlf> or until a second terminating </span> when this is found earlier just before the terminating <p><crlf>. It seems that regex behaves now greedier (while dotall is not enabled) then before. What am I missing?

04-16-2014, 03:59 AM	#2
jbacelar Interested in the matter Posts: 421 Karma: 426094 Join Date: Dec 2011 Location: Spain, south coast Device: Pocketbook InkPad 3	find: <span.?>(.?)</span> replace: \1

04-16-2014, 12:04 PM	#6
Skeeve Zealot Posts: 142 Karma: 669192 Join Date: Nov 2013 Device: Kindle 4.1.1 no touch	If I understood correct, he's saying that Code: <span(.)>(.)</span> would also match Code: <span>something<p><crlf>

04-16-2014, 12:08 PM	#7
DrChiper Bookish Posts: 1,049 Karma: 2006208 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	I checked and it appears the text is/was very bad formatted in which </span> is sometimes missing, thus causing the weird behavior by selecting unexpectedly large sections. And that my original regex did worked in the past for other texts may just be pure luck Oh well, at least I did learned a new regex trick and refreshed my knowledge about "Lazy quantifiers" versus "Greedy quantifiers". Thanks all! Last edited by DrChiper; 04-16-2014 at 12:11 PM.

04-17-2014, 04:14 AM	#9
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85400180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	The modify epub plugin is going in interesting places regarding junk spans. Might be worh a look... Using a negative lookahead you can avoid nesting issues, there is an example by me in that thread. Last edited by eschwartz; 04-17-2014 at 04:17 AM.

04-16-2014, 08:31 AM	#3
DrChiper Bookish Posts: 1,049 Karma: 2006208 Join Date: Jun 2011 Device: PC, t1, t2, t3, Clara BW, Clara HD, Libra 2, Libra Color, Nxtpaper 11	@jbacelar: Thanks, but no, that is not the solution. A combination of '*' (zero or more) and '?' (zero or one) together does not make sense here. I just tried my original regex within calibre 1.31 and it works as intended. However, it does not seem to work in calibre v1.32, which is odd.

04-16-2014, 08:37 AM	#4
Skeeve Zealot Posts: 142 Karma: 669192 Join Date: Nov 2013 Device: Kindle 4.1.1 no touch	I don't get what you want to expres... Could you post some examples?

04-16-2014, 08:46 AM	#5
kovidgoyal creator of calibre Posts: 45,609 Karma: 28549044 Join Date: Oct 2006 Location: Mumbai, India Device: Various	* - means match zero or more matching as many as possible. i.e. be greedy *? - means match zero or more matching as few as necessary, i.e. dont be greedy

04-16-2014, 01:24 PM	#8
DiapDealer Grand Sorcerer Posts: 28,892 Karma: 207182180 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Just remember; using regex to remove only certain span\|div tags will usually blow up in your face wherever spans\|divs are nested.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Paperwhite behaving strangely	route66	Amazon Kindle	6	03-25-2013 01:42 AM
Authors Behaving Badly	mr ploppy	Writers' Corner	2	08-27-2012 06:55 PM
PRS-300 Reader behaving badly!	docusk	Sony Reader	14	04-03-2012 07:46 PM
PRS-300 PRS-300 behaving badly	docusk	Sony Reader	2	03-22-2012 12:46 PM
Kindle behaving oddly after reset	ficbot	Amazon Kindle	2	09-11-2010 03:10 PM

Advert

Advert