I've got a wiki that is built at two places: * a static copy, aimed at being viewed without any web server, using a web browser's `file:///` urls => usedirs is disabled to get nice and working links * an online copy, with usedirs enabled in order to benefit from the language negotiation using the po plugin I need to use mirrorlist on the static copy, so that one can easily reach the online, possibly updated, pages. But as documented, "pages are assumed to exist in the same location under the specified url on each mirror", so the generated urls are wrong. My `mirrorlist` branch contains a patch that allows one to configure usedirs per-mirror. Note: the old configuration format is still supported, so this should not break existing wikis. OT: as a bonus, this branch contains a patch to support {hashes,arrays} of {hashes,arrays} in `$config`, which I missed a bit when writing the po plugin, and decided this time it was really needed to implement this feature. --[[intrigeri]] > Ping. --[[intrigeri]] [[!tag patch]] >> (I'm not an ikiwiki committer, opinions may vary.) >> >>> In my opinion, you're an ikiwiki committer! --[[Joey]] >> >> This would be easier to review if there weren't a million merges from >> master; perhaps either leave a branch as-is, or rebase it, or merge >> only at "significant" times like after a release? >> >> I believe Joey's main objection to complex $config entries is that >> it's not at all clear what [[plugins/websetup]] would do with them. >> Would something like this make a reasonable alternative? >> >> $config{mirrorlist} = ["nousedirs|file:///home/intrigeri/wiki", >> "usedirs|http://example.com/wiki", "http://example.net"]; >> >> From how I understand tainting, this: >> >> $untainted{$_} = possibly_foolish_untaint($tainted->{$_}) >> >> probably needs to untaint the key too: >> >> my $key = possibly_foolish_untaint($_); >> $untainted{$key} = possibly_foolish_untaint($tainted->{key}); >> >> --[[smcv]] >>> You are fully right about the complex `$config` entries. I'll >>> convert this to use what you are suggesting, i.e. what we ended up >>> choosing for the `po_slave_languages` setting. >>> >>> About the merges in this branch: Joey told me once he did not care >>> about this; moreover the `--no-merges` git log option makes it >>> easy to filter these out. I'll try merging tagged releases only in >>> the future, though. >>> >>> --[[intrigeri]] >>>> FWIW, I don't care about merge commits etc because I review >>>> `git diff ...intrigeri/mirrorlist` -- and if I want to dig deeper >>>> into the why of some code, I'll probably checkout the branch and >>>> use git blame. >>>> >>>> I agree with what smcv said, my other concern though is that >>>> this is such an edge case, that supporting it just adds clutter. >>>> Have to wonder if it wouldn't perhaps be better to do something >>>> using the goto plugin and cgiurl, so that the mirror doesn't have >>>> to know about the configuration of the other mirror. --[[Joey]] >>>>> I have implemented something using the cgi + goto in my (history >>>>> rewrite warning) mirrorlist branch. Please review, please pull. >>>>> --[[intrigeri]] >>>>>> Ping? I've merged 3.20110321 in my `mirrorlist` branch and >>>>>> checked it still works properly. --[[intrigeri]] >>>>>>> Joey: ping? I've rebased my `mirrorlist` branch on top of >>>>>>> 3.20120419, and checked it still works properly. I really >>>>>>> would like to see this functionality merged in time >>>>>>> for Wheezy. --[[intrigeri]] >>>>> concerning goto/cgiurl, what about having that as the default in >>>>> mirrorlist, but keeping ``nousedirs|file:///home/intrigeri/wiki`` and >>>>> ``usedirs|http://example.com/wiki`` valid for cgi-less cases? >>>>> that would keep typical installation with a clutter-less configuration, >>>>> and support more individual setups too. >>>>> --[[chrysn]] >>>>>> I would not mind. On the other hand Joey was concerned about >>>>>> cluttering the code to support edge cases, which I fully >>>>>> understand. The case you (chrysn) are describing being even >>>>>> more specific than the one I was initially talking of, I think >>>>>> this should not block the merge of the branch I have been >>>>>> proposing. Support for the usecase you are suggesting can >>>>>> always be added later if needed. --[[intrigeri]] >>>>>>> Well, that came out nice and clean. [[done]] --[[Joey]]