Regex engine?

troymc · 07-08-2010, 02:07 PM

Anyone know what regex engine Sigil uses?

Most normal regexes seem to work but for some reason I can't get it to like '^' for start of line.

I think I need to go peruse the manual.

Troy

troymc · 07-08-2010, 06:26 PM

Ok, I really like the new manual.

It told me exactly what I wanted to know.

Unfortunately it didn't get me what I wanted.

Sigil's manual basically states the same thing as the QRegExp doc does:

Quote:

In Sigil ^ always signifies the start of the string (apart from within character classes), so carets should be escaped unless used in this way. The same applies to $ which in Sigil always signifies the end of the string.

What does "string" mean in the context of a global S&R in Sigil?

Is it a line? In which case a ^ represents the start of the line.
Is it a file? In which case a ^ represents the start of the file.

Troy

capidamonte · 07-08-2010, 06:59 PM

In regex, it's usually a line. A string is a, well, string of characters.

Can you define better what it is you don't understand?

troymc · 07-08-2010, 08:01 PM

Quote:

Originally Posted by capidamonte

In regex, it's usually a line. A string is a, well, string of characters.

Can you define better what it is you don't understand?

It seems pretty simple. Normally ^ refers the beginning of a line. In Sigil's find/replace dialog '^A' does not find lines beginning with 'A'. (At least not for me, does it work for you?)

So it appears that, in Sigil, ^ does *not* refer to the beginning of the line - so what does it refer to?

The docs say "beginning of string" - so what's a "string"? Is QRegExp parsing a single file as a "string"? or the whole document?

Troy

Hitch · 07-08-2010, 09:39 PM

Well, I'm interested to see how this goes, as I'm struggling with a massive regex problem myself, albeit not yet in Sigil. But, the start of a string and the start of a line aren't the same thing. I think in regex if you only wanted to find the letter A at the beginning of a line you'd have to so indicate, whereas ^A would find any set of characters--string--that started with an A. I think. Maybe. cap will know.

st_albert · 07-08-2010, 11:21 PM

Just a reminder that when we say "regex" there are many variants of "regular expressions" that we may be talking about. I, myself, like to think of Perl Regex, even though "that way lies madness." But at least there are definitions. If you can follow them. See, e.g. http://perldoc.perl.org/perlre.html

It all depends on what you (or your regex variant) means by "string." In PERL it usually means a "line" i.e. whatever is between two lineend characters, but that need not be the case.

Sigil uses "almost" the Perl version of regex, via QT. I'm not sure how close that is to my understanding of ""regex."

Is your brain melting yet?

theducks · 07-09-2010, 02:17 AM

Quote:

Originally Posted by st_albert

Just a reminder that when we say "regex" there are many variants of "regular expressions" that we may be talking about. I, myself, like to think of Perl Regex, even though "that way lies madness." But at least there are definitions. If you can follow them. See, e.g. http://perldoc.perl.org/perlre.html

It all depends on what you (or your regex variant) means by "string." In PERL it usually means a "line" i.e. whatever is between two lineend characters, but that need not be the case.

Sigil uses "almost" the Perl version of regex, via QT. I'm not sure how close that is to my understanding of ""regex."

Is your brain melting yet?

And what looks like a "line" may be part of a "Wrapped paragraph" in code view.

Valloric · 07-09-2010, 04:45 AM

Quote:

Originally Posted by troymc

So it appears that, in Sigil, ^ does *not* refer to the beginning of the line - so what does it refer to?

I think this got broken a while back. I seem to recall someone casually saying how the start of line matching isn't working correctly, but since it never ended up on the tracker, I forgot to check it out. So do add it to the tracker.

And with that I leave for vacation.

capidamonte · 07-09-2010, 05:16 AM

If it's broken...

Hmmm.

Try

Code:

\nA

if you're on Windows, or

Code:

\r\nA

if you're on Linux.

Just make sure to replace the \r\n when you change stuff -- it's selected, so you don't want to overwrite it and remove a carriage return.

troymc · 07-09-2010, 11:43 AM

Quote:

Originally Posted by Valloric

I think this got broken a while back. I seem to recall someone casually saying how the start of line matching isn't working correctly, but since it never ended up on the tracker, I forgot to check it out. So do add it to the tracker.

And with that I leave for vacation.

Issue 498 opened. Have a great vacation!

Quote:

Originally Posted by capidamonte

If it's broken...

Hmmm.

Try

Code:

\nA

if you're on Windows, or

Code:

\r\nA

if you're on Linux.

Just make sure to replace the \r\n when you change stuff -- it's selected, so you don't want to overwrite it and remove a carriage return.

Unfortunately that didn't seem to work either.

But I know this guy Sed, and he's my friend. So I have an adequate work-around in place for what I needed to do.

Thanks, Cap!

Troy

capidamonte · 07-09-2010, 04:52 PM

Sounds like it's fairly-well broken then.

If you know Sed, then you're connected and we should all be asking you...

07-08-2010, 02:07 PM	#1
troymc Groupie Posts: 161 Karma: 608 Join Date: Aug 2008 Location: Plano, TX Device: Sony PRS-505 + B&N Nook + Motion LE1700 + Motorola Xoom Wifi	Regex engine? Anyone know what regex engine Sigil uses? Most normal regexes seem to work but for some reason I can't get it to like '^' for start of line. I think I need to go peruse the manual. Troy

07-08-2010, 11:21 PM	#6
st_albert Guru Posts: 696 Karma: 150000 Join Date: Feb 2010 Device: none	Just a reminder that when we say "regex" there are many variants of "regular expressions" that we may be talking about. I, myself, like to think of Perl Regex, even though "that way lies madness." But at least there are definitions. If you can follow them. See, e.g. http://perldoc.perl.org/perlre.html It all depends on what you (or your regex variant) means by "string." In PERL it usually means a "line" i.e. whatever is between two lineend characters, but that need not be the case. Sigil uses "almost" the Perl version of regex, via QT. I'm not sure how close that is to my understanding of ""regex." Is your brain melting yet? Last edited by st_albert; 07-08-2010 at 11:24 PM. Reason: clarification

07-09-2010, 05:16 AM	#9
capidamonte Not who you think I am... Posts: 374 Karma: 30283 Join Date: Jan 2010 Location: Honolulu Device: PocketBook 360 -- Ivory	If it's broken... Hmmm. Try Code: \nA if you're on Windows, or Code: \r\nA if you're on Linux. Just make sure to replace the \r\n when you change stuff -- it's selected, so you don't want to overwrite it and remove a carriage return.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Finding ebooks. Best engine?	rvdparis	General Discussions	3	08-26-2010 03:20 AM
Ebook Search Engine	Drezin	News	2	01-05-2010 05:16 PM
Possible scripting engine for Sigil	Valloric	Sigil	48	10-17-2009 09:58 AM
Paperback Digital's search engine	Sparrow	Workshop	7	11-26-2007 11:02 AM
Rollyo - create your own search engine	doctorow	Lounge	0	09-30-2005 08:40 PM

07-08-2010, 06:59 PM	#3
capidamonte Not who you think I am... Posts: 374 Karma: 30283 Join Date: Jan 2010 Location: Honolulu Device: PocketBook 360 -- Ivory	In regex, it's usually a line. A string is a, well, string of characters. Can you define better what it is you don't understand?

07-08-2010, 09:39 PM	#5
Hitch Bookmaker & Cat Slave Posts: 11,462 Karma: 158448243 Join Date: Apr 2010 Location: Phoenix, AZ Device: K2, iPad, KFire, PPW, Voyage, NookColor. 2 Droid, Oasis, Boox Note2	Well, I'm interested to see how this goes, as I'm struggling with a massive regex problem myself, albeit not yet in Sigil. But, the start of a string and the start of a line aren't the same thing. I think in regex if you only wanted to find the letter A at the beginning of a line you'd have to so indicate, whereas ^A would find any set of characters--string--that started with an A. I think. Maybe. cap will know.

07-09-2010, 04:52 PM	#11
capidamonte Not who you think I am... Posts: 374 Karma: 30283 Join Date: Jan 2010 Location: Honolulu Device: PocketBook 360 -- Ivory	Sounds like it's fairly-well broken then. If you know Sed, then you're connected and we should all be asking you...

Advert

Advert