[pmwiki-users] Strange characters, but only in some cases on Markup parse

[Petko Yotov]

On 31/10/2026 23:59, Christopher Cox wrote:
> https://youtu.be/euLOyGadyDs&list=PLZOfSznMOwYenRt25XSChMwxxim5bKVOx&index=104
> 
> I get (notice the 18B19B at the end):
> 
> [31-Oct-2026 17:52:54 America/Chicago] yt site: //youtu.be
> [31-Oct-2026 17:52:54 America/Chicago] yt id and parms: 
> euLOyGadyDs&list=PLZOfSznMOwYenRt25XSChMwxxim5bKVOx&index=10418B19B
> 
> The difference is the m[4] aspect ratio in quotes is missing, But isn't 
> that ok here? (zero or one time)  And I'm still not sure what the 
> 18B19B means?

In your example, there seems to be no aspect ratio in quotes after the 
link.

When I enabled your custom markup, both your links, ending with "4:3", 
got correctly found on the page.

These 18B and 19B may be "keep tokens", some markup or escaped text that 
was passed through Keep() and, in the "B" case, expected to be restored 
just before markup blocks are processed.

If so, it is actually $KeepToken + 18B + $KeepToken; each $KeepToken 
contains 2 control characters that browsers may not display.

Note that now people may [=escape=] URLs with special characters to 
avoid url-encoding them, e.g. spaces or parentheses like this:

https://en.wikipedia.org/wiki/[=Wiki (disambiguation)=]

So the Keep tokens may be part of the URL, to be restored and encoded 
later.

You can call RestoreMarkup() to see what was escaped.

Petko


On 31/10/2026 23:59, Christopher Cox wrote:
> Consider this cookbook recipe snippet (I'm looking for youtube video 
> links):
> 
> Markup('youtu_be_watch','<urllink',
>   
> "/\\b(?>(\\L))(\/\/youtu\.be)\/([^\\s$UrlExcludeChars]+)(\"[0-9]+:[0-9]+\")?/",
>   "YoutubeWatchLink");
> 
> function YoutubeWatchLink($m) {
>   extract($GLOBALS["MarkupToHTML"]); # get $pagename
> 
>   $aspectRatio = '';
>   error_log('yt site: ' . $m[2]);
>   error_log('yt id and parms: ' . $m[3]);
> 
> ... rest not shown.
> 
> 
> If on my page I have:
> 
> https://youtu.be/3kx24YtwPKA"4:3"
> 
> Looking at my debug log messages I see:
> 
> [31-Oct-2026 17:47:33 America/Chicago] yt site: //youtu.be
> [31-Oct-2026 17:47:33 America/Chicago] yt id and parms: 3kx24YtwPKA
> 
> Which I expect.  This is fine.
> 
> But then if I have:
> 
> https://youtu.be/euLOyGadyDs&list=PLZOfSznMOwYenRt25XSChMwxxim5bKVOx&index=104
> 
> I get (notice the 18B19B at the end):
> 
> [31-Oct-2026 17:52:54 America/Chicago] yt site: //youtu.be
> [31-Oct-2026 17:52:54 America/Chicago] yt id and parms: 
> euLOyGadyDs&list=PLZOfSznMOwYenRt25XSChMwxxim5bKVOx&index=10418B19B
> 
> The difference is the m[4] aspect ratio in quotes is missing, But isn't 
> that ok here? (zero or one time)  And I'm still not sure what the 
> 18B19B means?
> 
> I may just be failing regexp school.
> 
> _______________________________________________
> pmwiki-users mailing list
> pmwiki-users at pmichaud.com
> http://www.pmichaud.com/mailman/listinfo/pmwiki-users