Editor freezes while searching

[zetbox]

I’ve got this Regex for searching chapters

 <p([^>]*>\(<[^/].+?>)*chapter.+?</)p>

It worked fine for days. Today the editor stopped liking it. It will run once and then freezes.
The other saved searches work ok. Removed and reinstalled Calibre.

[eschwartz]

Why are you escaping one half of a parentheses pair?

[zetbox]

I forgot to remove the backslash. I run it actually like this

<p([^>]*>\s*(<[^/].+?>)*Chapter.+?</)p>

[eschwartz]

Well, that looks a lot better.

I'm afraid I don't know why the editor would hang on a well-formed regex though.

[zetbox]

Update:
I removed Calibre. Deleted any files related to it. Reinstalled.
I can run the RE from find/replace but not from saved searches it would run once ok but then stops, no freezing now - even after removing all searches and starting from scratch.
Any idea?
thanks a lot

[zetbox]

Update 2:
Find/Replace stopped working: Error No matches were found and this is what if copied to clipboard:

calibre, version 2.28.0
ERROR: Not found: <p>No matches were found for <pre style="font-style:italic">&lt;p([^&gt;]*&gt;\s*(&lt;[^/].+?&gt;)*Chapter.+?&lt;/)p&gt;</pre>

[kovidgoyal]

No matches being found is not "stopped working". It means your regex did not match anything.

[zetbox]

Originally Posted by kovidgoyal

No matches being found is not "stopped working". It means your regex did not match anything.

My bad I didn't set it to search All Text Files.
It looks like the freezing was caused by corrupt saved search

[zetbox]

Update 3:
It looks like Calibre Editor doesn't like my RE - Sigil runs it with no problem.
The RE engine goes to an endless loop (rather than freezes) if there are more than 2 tags like:

<p id="filepos2207" class="calibre1"><span class="calibre2"><span class="bold">Chapter One</span></span></p>

but it doesn't have a problem with this:

<p class="PCurrentAndTimesNewRomanAnd10ptAndCentreAlignmentAndLeftIndentAndRightIndentAndSpacingAfter"><span><span class="calibre18">Chapter 8</span></span></p>

[kovidgoyal]

I cannot reproduce that. Steps I tried:

1) Paste: <p id="filepos2207" class="calibre1"><span class="calibre2"><span class="bold">Chapter One</span></span></p> into an html file
2) Paste: <p([^>]*>\s*(<[^/].+?>)*Chapter.+?</)p>
into the find box and set mode to regex and search in current file
3) Click Find, the tag from (1) is highlighted

[zetbox]

As I can remember it is always after the second hit of Find or Find then Count All or Replace all

BUT...

Update 4:
Today it looped on <p class="calibre2" id="filepos3012"><span class="calibre3">Chapter 1</span></p>
I tried it a few times with the same result.
I replaced the "bad" entries with "good": <p class="p3"><span class="t2"><b>Chapter Two</b></span></p> and now it loops after first Find hit. I tested the "improved" file on Sigil - no problem.
So it is something else.

[zetbox]

Update 5:
I exported chapter html files from the "improved" epub that Editor had a problem with and made a new epub with Sigil. The Editor RE engine had no problem with the new file.
My guess is Sigil cleaned the html files while importing.

The conclusion: The problem wasn't the line with chapter number but some character/garbage after that.
I may try to run Editor in debug mode to find the problem.