Regex search details

DiapDealer · 02-22-2014, 08:59 AM

I see that unicode characters can be found using the \uFFFF format (was \x{FFFF} coming from Sigil's PCRE), and at least some of the unicode categories work (I can find any letter in any language using \p{L} or \p{Letter}), but many of those unicode categories don't seem to have the granularity I'm accustomed to.

The bulk of the punctuation searches seems to work:
\p{P} all punctuation
\p{Pd} dashes
\p{Pi} opening quotes
\p{Pf} closing quotes
etc...

But \p{Ll} and \p{Lu} (or \p{Lowercase_Letter}, \p{Uppercase_Letter}) both seem to find all letters regardless of case--just like \p{L}.

Is this expected/known behavior?

EDIT: Ooops! Never really expected those classes to be subject to the case-sensitive check-box. My bad. Nothing to see here!

mrmikel · 02-22-2014, 09:34 AM

That may explain some seemingly inconsistent behavior I have noted from time to time and just assumed it was because calibre's regex differed from PCRE in that case!

DiapDealer · 02-22-2014, 10:10 AM

What is the regex engine being used here, out of curiosity? Is it Matthew Barnett's Python regex module (hope, hope, hope)?

kovidgoyal · 02-22-2014, 11:38 AM

Yes, it is, and it was your mention of it in the original sigil thread that got me looking into using it. It had a couple of bugs that I helped find/fix, but otherwise it's been great.

DiapDealer · 02-22-2014, 11:58 AM

Cool. Glad it's working out. I'm a big fan of it.

Variable-width lookbehinds (in addition to the robust unicode support) just tickles me pink!

02-22-2014, 08:59 AM	#1
DiapDealer Grand Sorcerer Posts: 27,550 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Regex search details I see that unicode characters can be found using the \uFFFF format (was \x{FFFF} coming from Sigil's PCRE), and at least some of the unicode categories work (I can find any letter in any language using \p{L} or \p{Letter}), but many of those unicode categories don't seem to have the granularity I'm accustomed to. The bulk of the punctuation searches seems to work: \p{P} all punctuation \p{Pd} dashes \p{Pi} opening quotes \p{Pf} closing quotes etc... But \p{Ll} and \p{Lu} (or \p{Lowercase_Letter}, \p{Uppercase_Letter}) both seem to find all letters regardless of case--just like \p{L}. Is this expected/known behavior? EDIT: Ooops! Never really expected those classes to be subject to the case-sensitive check-box. My bad. Nothing to see here! Last edited by DiapDealer; 02-22-2014 at 09:05 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
search/replace multiline regex	Alt68er	Sigil	4	02-01-2014 09:40 AM
Regex search and replace	dwlamb	Sigil	6	04-12-2013 02:34 PM
regex search/replace	Sharlene	Sigil	10	01-28-2012 04:14 AM
need regex help search and replace	schuster	Calibre	4	01-10-2011 09:00 AM
regex search for roman numerals	Blurr	Calibre	2	12-16-2009 05:55 PM

02-22-2014, 09:34 AM	#2
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	That may explain some seemingly inconsistent behavior I have noted from time to time and just assumed it was because calibre's regex differed from PCRE in that case!

02-22-2014, 10:10 AM	#3
DiapDealer Grand Sorcerer Posts: 27,550 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	What is the regex engine being used here, out of curiosity? Is it Matthew Barnett's Python regex module (hope, hope, hope)?

02-22-2014, 11:38 AM	#4
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	Yes, it is, and it was your mention of it in the original sigil thread that got me looking into using it. It had a couple of bugs that I helped find/fix, but otherwise it's been great.

02-22-2014, 11:58 AM	#5
DiapDealer Grand Sorcerer Posts: 27,550 Karma: 193191846 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Cool. Glad it's working out. I'm a big fan of it. Variable-width lookbehinds (in addition to the robust unicode support) just tickles me pink!

Advert

Advert