[09:00:25] Hello and a big thank you for the awesome wiki engine! [09:01:55] I have an Apache2 rewrite rule that works very well, but the problem is that I do not understand _why_ it works. In the webroot I have the wiki in /w/ with URLs rewritten to be /wiki/Article_name. I have a Wordpress in /wp/ and a Matomo Analytics in /matomo/ [09:02:11] the rewrite rule, which enabled the VE to work is as follows: [09:02:14] RewriteRule ^(?:(?!rest.php/).)*$ %{DOCUMENT_ROOT}/w/index.php [L] [09:05:32] I can get that the wiki works with this one, but I do not understand how this does not mess up the /wp/ and the /matomo/. It has been a long time since I looked since more complicated regex syntax. If someone could help me understand why the RewriteRule does not mess up the Matomo and the Wordpress, I would very much like to rollout this arrangement to other wikis of mine [09:10:43] I cannot quite read that RewriteRule, but it looks to me a bit like that, everything that does not start with 'rest.php/' should be rewritten to %{DOCUMENT_ROOT}/w/index.php. Sorry, me not the texiest person on the Internets [09:16:32] I'm confused about the dot. I think I recall that signifies exactly one character, so "if there are no instances of 'rest.php/' followed by one character, rewrite the request to /w/index.php"? I am not sure what the colon means. I'll look it up now [09:21:24] could be the '?:' why the /wp/ and /matomo/ work fine. I'll shut up now, unless I get this figured out and would very much appreciate if someone could help me wrap my brain around this one, so I am not flying this thing like "It flies, but I do not know why" [09:43:46] There is asterisk after the parenthesis, so the entire expression (1 character not preceeded by `rest.php/`) is being matched for as many times as possible [09:45:04] There is also `^` at the beginning and `$` at the end, so the match has to start at the start of the string and end at the end of the string [09:46:57] thanks Archimedes1560[m [09:49:42] ahh, yes, the regular expression is not trying to catch what comes after rest.php/, but only to check for its existence and in case it did not find it, rewrite to /w/index.php ? [09:54:25] the remaining question mark for me is what does '?:' stand for. Isn't '?' used to express "exactly one occurrence"? Is the semicolon just a semicolon, I tried altavistaing for it and I'm not sure [09:55:58] `(?:)` is a non capture group [09:56:51] So the same as `()` except it can't be backreferenced with `$N` in replace string [09:56:54] ahhh... ok so the question mark followed by a semicolon will cause it being evaluated for matching but not being included in the $1, $2 etc [09:57:17] thanks, you answered my question already [09:57:42] Yeah, but only if it's inside (and at the start of parenthesis [09:57:49] * start of) parenthesis [09:58:11] Same as `(?!)` [09:58:30] Archimedes1560[m: So as long as no other software in the webroot uses something that starts with rest.php, I'm good to put the Wordpress and Matomo Analytics in the same webroot. Splitting to subdomains is horrible for SEO [10:01:19] ok, I can make my own mind up about that. I thank you human Archimedes1560[m for helping me understand why and even how the RewriteRule works. I used to understand the previous one I used, but that one did not work with VE, or more like VE didn't work with that, but eventually someone saw what was wrong and introduced this new one [10:02:21] that was one giant umbrella ticket about the REST errors in phab, but as Linus says "given enough eyeballs all bugs are shallow" [13:29:14] Hmm, are we not using WAL mode for the main sqlite DB? I wonder if that's intentional, we use it for objectcache when using sqlite [13:32:50] I wonder if there's an easy way to test what the difference would be. Do we have software to simulate normal user load on mediawiki? [17:42:09] finally... [17:43:09] I'm having trouble with gerrit access ... I can log into the webui and update my creds, but git with creds (ssh or http) isn't working. [18:04:57] Is there a way to disable Scribunto's auto documentation page feature? [18:09:22] Which part are you wanting to disable? [18:09:40] If you alter MediaWiki:Scribunto-doc-page-name to - (to disable it)... [18:09:46] https://github.com/wikimedia/mediawiki-extensions-Scribunto/blob/3ef9b40ecd52881adef93677de96ce4107a8cd87/includes/Scribunto.php#L89-L133 [18:09:55] That might have some of the desired result... [18:13:40] why would you want to [18:13:46] so I would create a page called MediaWiki:Scribunto-doc-page-name to set the value? [18:14:51] Izno[m]: Because I imported a lot of Wikipedia stuff to make infoboxes work on my wiki and I don't want all the redlinks, so disabling Scribunto docs will clear some redlinks. And I don't think we will (ever) need to create our own modules as it is a private wiki. [18:15:57] OK, this seems fairly straightforward. Thanks Reedy! [18:58:04] you could just blank all the doc pages instead [19:11:23] Too late :) [19:26:21] I'm dealing with a MW site using SMW where the SMW parser functions are sometimes not parsed when CirrusSearch is installed. Have any of the CirrusSearch devs seen anything like that? [20:03:36] hexmode: never heard of this problem, CirrusSearch should not do anything special with parser functions, it just uses ContentHandler to fetch its data (which might call the parser to generate the ParserOutput) [20:32:59] dcausse: ty... I'm just trying to figure out what may stop the SMW parser functions but not the other extensions right now. [20:45:29] hexmode: (might be unrelated to your problem but...) if your wiki follows the MW master branch we recently changed some function signature in ContentHandler that might break smw if it extends ContentHandler [20:45:50] no, this isn't master [20:46:15] it is 1.38