Jump to content

Incubator talk:Importing from Incubator

Add topic
From Wikimedia Incubator

Also replace {{#titleparts: {{PAGENAME}} | 2 }}}} with "". --Mkratz 01:25, 3 September 2010 (UTC)Reply

Is it used somewhere? If so, feel free to add it to the page. SPQRobin 19:49, 4 September 2010 (UTC)Reply
At least I'm using it myself to build templates that are easily portable across projects. --Mkratz 15:27, 6 September 2010 (UTC)Reply

translations

[edit source]

I somehow doubt the need to translate the How-to part. It is only needed for those who import (and, apart from that they should know how to do it with their experience, they can also speak English). --MF-W {a, b} 14:17, 29 July 2012 (UTC)Reply

Always be bold. I agree with you. Ebe123 (Talkabout it|contribs) 21:18, 31 July 2012 (UTC)Reply
Implemented. The how-to is now on a subpage. --MF-W {a, b} 01:26, 5 August 2012 (UTC)Reply
[edit source]

As discussed at User talk:MF-Warburg#Redundant pipes left behind after import and User talk:SPQRobin#Importing into tcywiki, the way step 5 was being implemented resulted in links like [[Page|page]] being left behind in some newly-imported wikis. While technically this causes no problems, "culturally" it may confuse some users into thinking such piping is necessary, so it should be avoided if possible. I haven't noticed such problems in recently imported wikis, so does that mean this has been "fixed"? Should the fix be documented here? - dcljr (talk) 22:37, 10 July 2017 (UTC)Reply

I am not aware of anything. --MF-W {a, b} 00:41, 14 July 2017 (UTC)Reply
[edit source]

I'd like to add a recommendation to submit a Phabricator task to run rebuildall.php and updateArticleCount.php after the importing of content into new wikis is finished. As explained in the rebuildall.php documentation, "You should run this script after importing some data into your wiki. It will populate the categorylinks, pagelinks, and imagelinks tables and refresh the search index." I've noticed in the recently created (and not yet counted) Shawiya Wiktionary that the wiki "thinks" that most main-namespace entries are dead-ends (do not link to other entries), even though almost all of them actually do contain wikilinks that are provided by a call to a template. Presumably this could be fixed by running the first script. The reason for running the second script is merely to avoid having to wait up to two weeks for the article count of the wiki to be determined or corrected (since it is run on the 1st and 15th of each month). Objections? Other opinions? - dcljr (talk) 01:25, 24 November 2019 (UTC)Reply

If rebuildall.php is considered overkill, then replace with refreshLinks.php, which just rebuilds the pagelinks, categorylinks, imagelinks tables. - dcljr (talk) 07:53, 2 December 2019 (UTC)Reply
BTW, the Shawiya Wiktionary has now been counted (statistics initialized on 1 December at, I think, c. 5am UTC), but the problem I'm pointing out above is still affecting the wiki. Merely recounting a wiki is not sufficient to fix this kind of problem. Making edits (possibly null edits) to all the pages on the wiki would fix it, but the script approach would be much faster. [grin] - dcljr (talk) 08:46, 2 December 2019 (UTC)Reply
FYI, the refreshlinks.php script was run on the Shawiya Wiktionary on December 8, 2019, and fixed the problems I was talking about above. Another recent example (October 2019) of a problem on a newly created wiki being fixed by running refreshLinks.php is phab:T235843. - dcljr (talk) 10:44, 8 January 2020 (UTC)Reply

Titles that differ only in the case of the first letter

[edit source]

Since titles here in the Incubator are case-sensitive, there are sometimes page titles in test wikis that differ only in case (i.e., capitalized vs. not), even though the target wiki for importing (say, a Wikipedia) is case-insensitive (for the first letter). As an example "in the wild", there are Wp/knc/kǝri and Wp/knc/Kǝri. The first of these got imported to the newly created Central Kanuri Wikipedia as knc:Kǝri, while the latter did not get imported, apparently. I don't know if that was intentional. In any case, my question is, should we include a warning about case (in)sensitivity in these instructions? Presumably, if the equivalence of these page titles wasn't noticed before importing, one version of the page would get imported "over" the other. Right? And that wouldn't be good… - dcljr (talk) 03:55, 1 February 2025 (UTC)Reply

Yes, I've noticed the problem a few months ago when I started doing imports myself. I reported this ticket to prevent it: https://phabricator.wikimedia.org/T380064 , but there hasn't been much action on it yet, as far as I know. @Jhsoby can correct me if I'm wrong.
One of the problems specifically for the knc import is that when I noticed the articles with small letters, I couldn't change it because editing in the Incubator is locked after the domain is created.
Perhaps someone should just find all the titles that have small letters, fix them, and prevent creation of new pages with small letters using MediaWiki:Titleblacklist or the AbuseFilter or something. Better ideas are welcome. Amir E. Aharoni (talk) 14:36, 1 February 2025 (UTC)Reply
Just for the record, do you know what would have happened if you had imported both pages (kǝri and Kǝri)? I assume both sets of edits would end up in the history of the resulting page (just like importing the same page multiple times apparently causes duplicate edits in the page history). Right? Would you get a warning when that happens, so you'd know? Just wondering… - dcljr (talk) 03:40, 2 February 2025 (UTC)Reply
I don't really know :( Amir E. Aharoni (talk) 06:14, 2 February 2025 (UTC)Reply