Regex problem with brackets and returning a value from memory

[Chang]

I was doing a page list for my accessible EPUB 3 ebook according to these instructions (https://gold-chartreuse-795z.squares...ng-a-page-list). In one part there is Regex used:

Create the correct code for the page breaks. This RegEx string works in oXygen XML editor:
<span class="com-rorohiko-pagestaker-style">(\d+?)</span>
<span epub:type="pagebreak" role="doc-pagebreak" id="page$1" aria-label=" Page $1\. " />

I tried this Regex find and replace search in Sigil and it doesn't seem to work. When I use this same Regex string in Sigil's Regex search, all I get is this:

<span epub:type="pagebreak" role="doc-pagebreak" id="page$1" aria-label=" Page $1\. " />

Seems that in Sigil nothing is saved in memory when using brackets. Is there some other way to do this in Sigil? I know in the instructions Sigil is not used for this part but I would like to know why this regex string is not working in Sigil.

[Chang]

I just found a solution for my problem and it seems that I'm not able to delete my post anymore.

Instead of $1 I need to use \1 in Sigil to return the value from the memory.

Sorry for posting too hastingly!

[DNSB]

You may want to practice RegEx on RegEx 101. Use the PCRE2 flavour of RegEx for Sigil practice. The $1 suggests an ECMAscript/Javascript flavour of RegEx.

[Capricorn]

Hmmmm.....I am confused now. I did a little test:
I put this regex into the regex101 machine, using PCRE2: <span>([^<>]*)</span>
Regex101 then tells me there is a mistake and I need to change it to:
<span>([^<>]*)<\/span> notice the backslash in the closing span

However, when I use <span>([^<>]*)</span> - without the backslash, in Sigil v2.3.1 (running on win11 pro, build 22631.4169), it works perfectly. Hence, my question: Is the regex of Sigil a variant of PCRE2 or...?
As I am not an expert on regex, but still like to use it, I am trying to understand what is going on.

And on this page https://sigil-ebook.com/sigil/ it says it uses PCRE

And when I use python on regex101 instead of PCRE or PCRE2, it does not need a backslash either.

Hmm.....yes, definitely confusing............

[theducks]

the backslash is the 'Escape' character
That just says take this literally.
[ ] is normally a Set
\[ \] is text enclosed in square brackets

[DiapDealer]

Sigil is definitely not using any variant of pcre2.

I know of no special meaning for the forward slash in pcre2. I've never needed to escape it. If there's an issue, it would seem to me it is with Regex101 and not Sigil, PCRE2, or Python. It certainly won't hurt to escape them, but it seems a waste of time.

NOTE: the only pcre flavors I see on regex101 seem to be php specific. Am I missing something? Forward slashes would definitely need special treatment if used in php.

My guess is that regex101 being a regex analyzer served by a webserver; the forward slash would most definitely need to be escaped. But to my knowledge, that requirement is not inherent to pcre2 itself.

EDIT: Regex101's PCRE/PCRE2 flavors definitely seem to be predicated on the notion that they're going to be used in PHP. That's a mistake in my opinion, but the extra escapes won't cause any harm if you include them.

[Capricorn]

Good to know that the problem seems to lie with the regex101 engine, and not sigil. So, no need to adapt any of my regexes. Cool.
Thank you.