[00:40:40] Skizzerz: I am getting a message saying that "The authenticity of host can't be established". [00:41:16] I created and added the key locally and uploaded via cPanel, too. [00:41:21] So not sure what this is about. [01:20:48] Delpy: that's because you never connected there before [01:39:31] Platonides: Oh, so its completely normal and secure? [01:47:18] yes [01:47:43] you get that for every first-time connect to a host [01:55:34] MaxSem: ok, great! [05:07:03] Hello everyone [05:07:09] I have to discuss a very serious problem [05:07:13] can anyone please help me [05:07:50] I have recently checked the following page that contains very private & confidential information about me: [05:07:52] https://deletedwiki.com/index.php?title=Farhan_Wilayat_Butt [05:08:16] Please be apprised that I am working in a very sensitive field (i.e. interfaith activities) which is a very controversial topic in Pakistan. [05:08:20] Sharing my personal and confidential details online is of course not safe for me. [05:08:39] I therefore want to request to delete the above page on urgent basis. [05:08:43] can anyone please help me out [05:08:45] anyone... [17:49:58] Hi all! I set up a MediaWiki instance several years ago (v1.17.0) and wrote a couple articles. I forgot about it, and recently want to return to it, but I want to update to the latest version. Since there are only ~20 articles, would it be easier to export/import those articles or should I actually perform an upgrade on top of my existing instance? [17:58:00] Totengeist: That's old, upgrading could be painful. I'd just xmldump and import. [17:59:51] G_SabinoMullane: great. Thanks! [19:27:38] Not possible to put a js file from a package pulled in via composer into a resource module? [19:51:30] JeroenDeDauw: if you use the 'vendor/whatever/whatever.js' path for the resource module, i don't see why it wouldn't work. but it probably may not pass CI checks. [21:14:15] hey, have a stupid question about the mediawiki api. i'm trying to grab an api token using a bot account i created in Special:BotPassowrds and pass it back to authenticate to my mediawiki instance. currently i'm grabbing the api token by doing the first step of `curl -s --cookie-jar cjar -X POST "https://$WIKI/api.php?format=json&action=query&meta=tokens&type=login&lgname=$BOTUSER&lgpassword=$BOTPASS"`, [21:14:18] and grabbing the api token with jq to assign to a variable in a script. my issue is that when i'm passing the token with `curl -s -b cjar -X POST "https://$WIKI/api.php?format=json&action=query&meta=tokens&type=login&lgname=$BOTUSER&lgpassword=$BOTPASS&lgtoken=$APITOKEN"` i'm getting "WrongToken" as an error trying to pass this info back [21:15:23] i'm not sure if my curl syntax is wrong or if i'm accessing the mediawiki api incorrectly? from what i was able to gather on the API:Login page, i need to do two requests. first to grab the token and second to pass the token back to authenticate with the same session [21:25:19] IDrinkMilk: your API queries look wrong. You want action=login, not action=query [21:25:27] for both the initial as well as the second [21:25:45] the password is not needed on the first request [21:25:58] req 1: action=login&lgname=USERNAME&format=json [21:26:13] req 2: action=login&lgname=USERNAME&lgpassword=PASSWORD&lgtoken=TOKEN&format=json [21:27:26] isn't action=login depricated? i get this warning when i use it "Fetching a token via action=login is deprecated. Use action=query&meta=tokens&type=login instead." [21:27:53] going to try the strings you pasted now, tabing to my shell :) [21:28:46] ah, I didn't see that in the documentation page [21:29:01] Fetching a token via it doesn't mean the whole thing is deprecated [21:29:11] but if mw reports that, then change req 1 to action=query&meta=tokens&type=login [21:29:13] &format=json [21:29:27] you do not need to include lgname, lgpassword in the token query [21:29:38] req 2 should remain as-is [21:31:40] hm, still getting {"login":{"result":"WrongToken"}} as the response with reg2 [21:31:45] req2** [21:33:29] token i grabbed is displaying as this [API-TOKEN]+\\ . i've url encoded the +\\ since that was used in the mediawiki example [21:34:14] try without encoding it [21:34:38] and it's only one \ [21:35:03] oh right, that's probably the actual reason [21:35:06] json escapes \ [21:35:31] thanks Reedy [21:35:38] oh right, derp. forgot about that... [21:36:07] okay, tried again... still getting WrongToken. this is trying with both the +\ as ascii and as url encoding [21:53:51] \\\\\\\\\\\\\ [21:56:33] IDrinkMilk: I'm able to repro, so I think it's something about how curl or your shell is handling the args that's mangling stuff [21:56:40] (tried in python and it worked) [21:57:21] heh, interesting... guess either curl or bash is messign with this. guess i'll have to brush off my python skills to wrap up this project for $employer :) [21:57:29] thank for your help Skizzerz and Reedy ! [21:58:08] well, depending on the project, there might already be something pre-written you can use [21:58:14] pywikibot has a lot of features [21:58:26] IDrinkMilk: are you using single quotes? [21:58:49] was using double quotes, but i'll try single quotes. h/o [21:58:55] I tried with single quotes and it was still failing for me =\ [21:59:35] how did you pass the token to curl? [22:00:39] yep, no dice. single quotes didn't work [22:01:16] I mean through which curl parameter? [22:01:25] i'm running req1 (from above) through jq to grab the token and assign it to a variable. when i run req2 i have lgtoken=$var [22:02:00] oh, sorry. i'm running curl -s -b cjar -X POST . cjar is in the current dir, holding the sessionid from running req1 [22:02:16] * Skizzerz was using -d [22:02:51] ah, -d doesn't urlencode by default [22:03:56] ok that worked [22:04:01] IDrinkMilk: lemme paste what I got [22:06:10] IDrinkMilk: http://dpaste.com/3R9STDK [22:34:50] AH! just got it working! thanks so much Skizzerz :D [23:25:21] So, I have a wiki with several thousand articles and it is currently being translated into 7 other languages. When I was dealing with 3 languages linking between the different sites was simple enough. But certain languages are approaching and exceeding 100 artlicles each, meaning I have thousands of links that need to be managed. My current imperfect system is to edit the sidebar of each site to have links [23:25:27] like this [[es:{{PAGENAME}}]]. for obvious reasons this is problematic. Unfortunately, the Wikidata model that Wikipedia uses seems a little more elaborate and time consuming than I would prefer. Can anyone suggest a simpler, yet elegant solution? [23:45:17] dp: Write a bot/script? [23:45:47] That can iterate through the few thousand pages and cross-reference them in a section on the page. [23:46:05] Or you could write a custom extension thing, but it'd prolly be annoying.