Re: [Tiki-devel] Anchors are screwed up

Volker Wysk <post-hhF2Jplw28UoZk/[email protected]> Tue, 17 Jan 2023 13:39:04 +0100
Newsgroups	gmane.comp.cms.tiki.devel
Message-ID	<[email protected]>
Am Sonntag, dem 15.01.2023 um 14:37 +0000 schrieb Jonny Bradley via
TikiWiki-devel:
> Hi Volker
> 
> That sounds possible, odd that no one's noticed before - i suggest you
> create a wish on https://dev.tiki.org/Make-a-wish and maybe a junior dev
> can be found to look at it :)
> 
> Meanwhile i guess you have to do the underscore replacement yourself...

The underscore replacement is the problem. It allows only ASCII alphanumeric
characters. For Asian or other languages, which don't use latin letters, the
anchors generated from headings are the empty string!

See here for more details: 
https://volker-wysk.de/tiki/Anchors%20in%20Tiki.html

I've got a plan of how the situation could be resolved. Only limited effort
is required.

1. Write a function which does a proper "URL-fragment string" encoding.
(That's a term from the URL standard.) 

2. For ANAME/ALINK plugins and for anchors generated from headings, encode
the anchor name with this function, in the generated HTML. All characters
are allowed and get encoded.

3. Use the old method, which throws away non-ASCII characters, as well. This
means that *two* anchors are generated for each occurrence of an anchor in
the Tiki page.

This way, compatibility with old Tiki pages is ensured.

4. Encode the subsection part of internal Tiki links with the new function,
when generating HTML. Old links, which have the manual underscore
replacement in it, won't get changed by the new encoding. They will match
the old anchors (3.). New links, which don't have an encoding in the
subsection part, will match the new anchors (2.).

Doing it this way wouldn't require any transformation of old Tiki pages'
source code. 


So I'm creating a wish now...

Cheers,
Volker

> > On 15 Jan 2023, at 11:18, Volker Wysk <post-hhF2Jplw28UoZk/[email protected]> wrote:
> > 
> > Hi.
> > 
> > In wikiplugin_aname.php and wikiplugin_alink.php, there's the following:
> > 
> >    // the following replace is necessary to maintain compliance with XHTML
> >    // 1.0 Transitional and the same behavior as tikilib.php. This will 
> >    // change when the world arrives at XHTML 1.0 Strict.
> >    $aname = preg_replace('/[^a-zA-Z0-9]+/', '_', $aname);
> > 
> > This means, only latin letters and digits are allowed in anchor names.
> > Everything else is replaced by an underscore when rendering the HTML page.
> > 
> > This (sometimes) works, as long as only the AName and ALink plugins are
> > used. But it doesn't, when the subsection part in an internal link is used -
> > ((Page|#Subsection|...)). I've tried it with "((Ankertest|#Überschrift-
> > Anker))" and got a subsection part of "#%C3%9Cberschrift-Anker" in the
> > generated HTML.
> > 
> > The subsection part needs to undergo the same treatment as the anchor name,
> > which the AName plugin does, when this is supposed to work.
> > 
> > Anchors which are generated from headings are different as well. I've tried
> > it with a heading called "Überschrift dazwischen", and the resulting anchor
> > name in the HTML page is "berschrift_dazwischen".
> > 
> > And what about languages that don't use a latin alphabet, such as many Asian
> > languages? The complete anchor names will be replaced by an underscore.
> > 
> > Anchor handling is really screwed up.
> > 
> > The world has "arrived at XHTML 1.0 Strict" long ago - in 2002, it says in
> > the https://www.w3.org/TR/xhtml1/ page. So it's time that "this will
> > change".
> > 
> > Regards,
> > Volker
> > _______________________________________________
> > TikiWiki-devel mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel
> 
> 
> 
> _______________________________________________
> TikiWiki-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel

_______________________________________________
TikiWiki-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel
signature.asc (application/pgp-signature, 833 B)