[pmwiki-users] Search with non-ascii characters in url q= parameter

[Petko Yotov]

Yes, PmWiki can handle international characters in searches (try on 
pmwiki.org), but you may need to enable UTF-8.

See https://www.pmwiki.org/wiki/PmWiki/UTF-8 as this may not be trivial 
for existing wikis.

In fact, the browser will use the encoding of your page when it submits 
a form. So if your wiki is in the old encoding ISO-8859-1, "ä" will be 
encoded as "%E4", if your wiki is in UTF-8,  "ä" will be encoded as 
"%C3%A4".

The browser's URL bar will show the UTF-8 international characters 
decoded (ä), but not the other encodings (%E4). All browsers, all major 
search engines, Wikipedia and other popular websites use UTF-8.

I have only recently encountered a Microsoft IIS installation that 
refused to correctly decode international characters in pagenames (not 
sure about searches) but somehow it worked with CleanURLs.

Petko
On 24/01/2023 10:50, Hans wrote:

> Hello Pmwiki-users,
> 
> Normal search using a search box works fine, when the query string 
> contains special non-ascii characters.
> For instance searching for pages with "Bär" gets the right results, and 
> "Bär" appears in the search box, and in the URL bar I got the entry:
> http://localhost/softflow/test?n=Site.Search&action=search&q=B%E4r
> so the "ä" character is rendered as "%E4".
> 
> Using a Context Search add-on (for Firefox) I can grab text with 
> non-ascii characters and have that string given to PmWiki as a url q= 
> parameter. But PmWiki does not like that, and garbles it.
> In the example the url parameters will read: 
> ....?n=Site.Search&action=search&q=Bär
> and the search will fail, and the Search page displays with "Bär" in 
> the search box, not the string I want to search for.
> 
> The "Select and Search" Firefox add-on is set up for a PmWiki search 
> like this:
> http://localhost/softflow/test?n=Site.Search&action=search&q=%s
> In a search the %s gets replaced with a highlighted text string using a 
> context menu option to use this add-on with this search option.
> 
> So my question is: could PmWiki handle special non-ascii  characters in 
> url parameters for a search query?
> 
> If I use google search engine, it shows no  problems with special 
> characters being provided via url parameters.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pmichaud.com/pipermail/pmwiki-users/attachments/20230124/397f0e1a/attachment.html>