Topic on Project:Support desk

Search indexing affected by less-than sign?

7
Open4D (talkcontribs)

I am an end user of a wiki that is not on the Internet. Details:

MediaWiki 1.21.1
PHP 5.3.3
MySQL 5.1.47

The list of extensions doesn't seem to include anything for search. (Not Lucene or Sphinx, for example)


I was just wondering whether anyone else using MediaWiki (without any search extension) sees the same behaviour as me ...

Edit a normal wiki page. Place the following sentence (if I can call it that) on a line by itself:

One < two butundeniably seven > six.

And then try using MediaWiki's search functionality to search for 'butundeniably'. I get: "There were no results matching the query". I suspect the less-than sign is relevant in this behaviour.

Then change the word "two" to be the corresponding digit "2", resulting in this:

One < 2 butundeniably seven > six.

Now the search successfully finds the page.


There is not enough detail currently at Help:Searching to determine whether this behaviour is to be expected. Does anyone else encounter it?

Thanks,

88.130.122.136 (talkcontribs)

Hi!

I have not realized there is such behavour, but I see that your first expression, the one, which won't work, except the spaces looks like an HTML tag. Maybe the MediaWiki parser somehow thinks that "< two" would be a valid HTML tag (one of the self-closing ones) and so it does not take the contents into the search results as it thinks that it's only markup.

Open4D (talkcontribs)

Hi!

Yes, thank you. I think that could be the reason. If it is, I would ideally like to find out the full details of what's happening, and consider modifying "Help:Searching".

Does anyone know of public MediaWiki wikis (without any search extension) that might not mind if I did a quick test of this? I tried testing with the Project:Sandbox here, but had some problems, and anyway this wiki (www.mediawiki.org) does use Lucene (I believe), and some other search extensions listed here.

Or is it very rare not to use a search extension? Does anyone really bother with the built-in search functionality?

Thanks

Open4D (talkcontribs)

Or is anyone aware of any technical documentation about the built-in search functionality? I might use this to supplement the user documentation at Help:Searching.

Thanks

88.130.91.210 (talkcontribs)

Hi again!

I am sorry to tell you, but I do not know of more technical docs cocerning the default search.

However, I do think that this search is used often: You don't have to set up anything, it provides some basic options and it works out of the box. So, yeah, I think it will be used widely.

I have just checked that with this search query: "< range of use". It finds pages with that content. A search for "< range of use >" (should display nearly the same pages as the text on the pages continued with these signs), however, that search incorrectly does not display most of the pages.

So: Yes, this behaviour is also reproducable for me.

Open4D (talkcontribs)

Thanks again 88.130.91.210!


> I do not know of more technical docs cocerning the default search

Does anyone else know?


> this behaviour is also reproducable for me.

And it feels like a bug to me, so I might try Bugzilla next. (Perhaps that way I will also find out some information to let me enhance "Help:Searching".)

Open4D (talkcontribs)

> And it feels like a bug to me, so I might try Bugzilla next.

I've raised bug 60354.

Reply to "Search indexing affected by less-than sign?"