What regular expression system does Calibre use?

[ownedbycats]

Mostly for the purposes of finding an online regex tester that returns the same results as Calibre. Occasionally a regex that seems to verify okay returns unexpected results in Calibre (good thing I backup)

[DaltonST]

Python.

Try Pythex

[ownedbycats]

Thank you!

[davidfor]

I like https://regex101.com/. The explanation of the parsing helps a lot when debugging a complex expression.

[Quoth]

Originally Posted by davidfor

I like https://regex101.com/. The explanation of the parsing helps a lot when debugging a complex expression.

That's interesting.
I use KATE to test any regex before destroying something with a Pearl program. A language that seems to exist purely as a wrapper for regex and I find head wrecking to do "ordinary" things I've done in VB6, C, C++, Modula-2, Java etc.

There is a Pearl site that's good for regex help.

When I first moved completely to Linux from Windows I ran Notepad++ in WINE, as it's a good multdocument tabbed editor with regex and syntax highlighting that can be customised. Then I discovered KATE.

Pearl and a regex in general seems close to a "write-only" language. I find it hard to see what a regex might do.

I found that while LO Writer has good regex Search you can't use parameters in the Replace field. Also it has some non-standard search symbols? It does seem to use a python library for it, but I could be wrong. I do use it to find misformed dialogue punctuation, spaces where there should be none and incorrectly terminated paragraphs, though that finds titles, headings and preambles as they don't punctuate like body text. So I've rarely used search & replace in Calibre. If a book from Gutenberg or wherever needs a lot of work I'd export it converted to RTF, edit and save in ODT, and a final extra Save As in DOCX for Calibre to import.

[georgemk]

Originally Posted by Quoth

That's interesting.
I use KATE to test any regex before destroying something with a Pearl program.

KATE is now available from the Microsoft Store on Windows 10 along with a few other KDE applications.

[davidfor]

Originally Posted by Quoth

That's interesting.
I use KATE to test any regex before destroying something with a Pearl program. A language that seems to exist purely as a wrapper for regex and I find head wrecking to do "ordinary" things I've done in VB6, C, C++, Modula-2, Java etc.

As my day job is Perl programmer, I have to disagree with that statement. But, I can see it.

There is a Pearl site that's good for regex help.

Someone recently pointed me to https://regexr.com/. That is Perl and Javascript. It has some of the function that the first I mentioned does, but, I haven't used it much.

When I first moved completely to Linux from Windows I ran Notepad++ in WINE, as it's a good multdocument tabbed editor with regex and syntax highlighting that can be customised. Then I discovered KATE.

I settled on geany as the closest the Notepad++. Though I use either Eclipse or VSC when working on projects.

Pearl and a regex in general seems close to a "write-only" language. I find it hard to see what a regex might do.

I frequently take the regex from our code and use the site above to check it.

I found that while LO Writer has good regex Search you can't use parameters in the Replace field. Also it has some non-standard search symbols? It does seem to use a python library for it, but I could be wrong. I do use it to find misformed dialogue punctuation, spaces where there should be none and incorrectly terminated paragraphs, though that finds titles, headings and preambles as they don't punctuate like body text. So I've rarely used search & replace in Calibre. If a book from Gutenberg or wherever needs a lot of work I'd export it converted to RTF, edit and save in ODT, and a final extra Save As in DOCX for Calibre to import.

I do all my book editing in the calibre editor. I've probably done as much regex there as anywhere else.

[ownedbycats]

Originally Posted by davidfor

I like https://regex101.com/. The explanation of the parsing helps a lot when debugging a complex expression.

I tried that site. It's very useful.

[Thasaidon]

I Like the the freeware 'Expresso" program. It will decode existing Regexes and gives you the facility to create, test and store your own Regexes.

It is worth a look at and can be used offline.

[Quoth]

Originally Posted by davidfor

As my day job is Perl programmer, I have to disagree with that statement. But, I can see it.
I do all my book editing in the calibre editor. I've probably done as much regex there as anywhere else.

It's an extreme view from someone that's not seriously used Pearl. You might feel the same about Forth or Prolog.

Since I'm getting source in docx or odt, it wouldn't make sense to edit in Calibre. If I get a PD book that just needs some CSS/HTML changes I might edit in Calibre. If it's been very badly proofed from a scan of an ancient copy, or really stupid formatting/lack of format of headings, images, index etc I might export as RTF and Save As odt, then actually create/edit styles based on what I know works.
If I was only "fixing" epubs I'd probably edit in Calibre.

[PeterT]

Originally Posted by davidfor

As my day job is Perl programmer,

I have fond memories from many years ago when I was first involved with Tivoli tools for systems management. The Distributed Monitoring component relied heavily on perl scripts deployed to servers, and the Enterprise Console used Prolog. Those were the days

[DNSB]

Originally Posted by Quoth

It's an extreme view from someone that's not seriously used Pearl. You might feel the same about Forth or Prolog.

Very fond memories here of using Forth. It was a fun language despite making it so easy to create write-only code. I must admit that I also loved my HP RPN calculators.