User talk:Citation bot/Archive 39
Latest comment: 3 days ago by GreenC in topic Don't convert URLs to all lowercase
 | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
| Archive 35 | ← | Archive 37 | Archive 38 | Archive 39 |
archive-link should be replaced with archive-url, not archive-date
Treat deadlink= the same as deadurl= (to convert to url-status=)
- What should happen
- [3]
- We can't proceed until
- Feedback from maintainers
audible.com
- Status
- {{fixed}}
- Reported by
- Questions? four Olifanofmrtennant (she/her) 17:30, 13 May 2024 (UTC)
- What happens
- So citation bot is changing audible listings to cite book.
- What should happen
- Cite web should be used.
- Relevant diffs/links
- [4]
- Replication instructions
- Add a source to audible.
- We can't proceed until
- Feedback from maintainers
Replace Cite Web with Cite Journal
- Status
- {{fixed}} several different ways to catch this and others
- Reported by
- —Carter (Tcr25) (talk) 13:30, 14 May 2024 (UTC)
- What happens
- changed Cite Web to Cite Journal
- What should happen
- should have left as web or possibly changed to Cite Archive
- Relevant diffs/links
- https://en.wikipedia.org/w/index.php?title=Stained_glass_windows_by_Harry_Clarke&diff=1223741434&oldid=1223665575
- We can't proceed until
- Feedback from maintainers
Not sure if Citation bot assumed that a the presence of a doi meant it was a journal, but the ref was pointing to a piece in the Trinity College Library archives collection. Attributed authors were picked up and added in ways that created CS1 errors. —Carter (Tcr25) (talk) 13:30, 14 May 2024 (UTC)
Don't add title=MSN
- What happens
- [5]
- We can't proceed until
- Feedback from maintainers
It appears that Citation Bot has a blacklist for generic titles. "MSN" is a common generic title that I see Citation Bot add, so it should be added to that blacklist. (I know the diff I linked is old, but I've seen this happen recently.):Jay8g [V•T•E] 00:31, 17 May 2024 (UTC)
AS Racine FC
Hi bot.can you help correct a typo on the AS racine fc page you just edited?
The page title "AS Racine FC" is wrong. The correct title is "AS Racines F.C".
Please correct it. RoaringEdits (talk) 14:18, 18 May 2024 (UTC)
Unlinking journal parameter
Hello,
According to the documentation at Template:Cite journal:
- work (required by {{cite journal}} and {{cite magazine}}): Name of the work containing the source; may be wikilinked if relevant.
However, it seems citation bot unlinks the journal= parameter (by default?). Is this intentional? Should the documentation be updated if wikilinking journals is discouraged? Or should citation bot be updated to honor an existing wikilink, if any? (Example diff. There's some other questionable changes in there but that's probably from bad input data and harder to deal with.). SnowFire (talk) 21:16, 22 April 2024 (UTC)
- Caused by the (JTS) being present for no reason. The bot expects
|journal=[[Foo|Everything]], not|journal=[[Foo|Something]], something else. 21:29, 22 April 2024 (UTC) Headbomb {t · c · p · b} 21:29, 22 April 2024 (UTC)- If this is intentional, then I feel like the documentation should be updated that this field needs to have only a single link. I don't really see the problem with the other approach, but I respect that maybe it's needed for bots / microformats rather than humans. (I saw quite a number of short references to "JTS" hence putting the acronym gloss in there, so it's not like a totally irrelevant piece of information to have somewhere.) SnowFire (talk) 21:46, 22 April 2024 (UTC)
date/year are the same
- What should happen
- [6]
- We can't proceed until
- Feedback from maintainers
Won't run on categories with only one item
- What happens
- Webpage reports
!Category appears to be empty
- Replication instructions
- Run the bot on any category with only one item. This happens both with the direct link on CS1 error/maint categories and with the category input on the Toolforge page.
- We can't proceed until
- Feedback from maintainers
Category-link formatting in bot's edit summaries is broken
- Status
- {{fixed}}
- Reported by
- Whoop whoop pull up Bitching Betty ⚧️ Averted crashes 15:21, 19 May 2024 (UTC)
- What happens
- When the bot makes an edit from a category run, the category link in the edit summary is missing one of the two closing brackets necessary to actually activate the link: [[Category:NAME].
- What should happen
- The category link should be properly formatted as a link, with two opening and two closing brackets: [[Category:NAME]].
- Relevant diffs/links
- https://en.wikipedia.org/w/index.php?title=Planking_(fad)&diff=prev&oldid=1224637690
- Replication instructions
- Run the bot on a category; any edits made during that run will show this bug.
- We can't proceed until
- Feedback from maintainers
TNT Title if "Request Rejected"
- What should happen
- [7]
- We can't proceed until
- Feedback from maintainers
Introduces ref error when citing Penguin publisher website
- What happens
- Introduces ref error by turning
{{cite web}}to{{cite book}}ostensibly when url=https://www.penguin.co.uk/books/295/295446/bullshit-jobs/9780141983479.html. - What should happen
- Citation bot should take no action. Instead of interpreting that link as referencing a book, citation bot should not assume. In this case, the publisher's webpage itself is being cited as a primary source, so a book it not being cited.
- Relevant diffs/links
- [8]
- We can't proceed until
- Feedback from maintainers
- This is the solution. Headbomb {t · c · p · b} 08:12, 22 May 2024 (UTC)
- Pretty grateful we get to go through every citation to a publisher's blurb or library catalogue record and add a hidden html comment so Citation bot can continue assuming we're all ding-dongs who don't know the difference between a book and a website.Having
saidsassed that, I'm not sure what percentage of us are ding-dongs who don't know the difference between a book and a website, and I'm of the opinion that the true solution is probably neither manually adding tens of thousands of hidden html comments nor endlessly tweaking exceptions into Citation bot's code, but mapping|website=to|via=instead of|periodical=in the context of {{Cite book}}, or rolling back the deprecation of the|periodical=aliases from {{Cite book}}, or having Citation bot output {{Cite conference}} or {{Citation}} to support more parameters. Folly Mox (talk) 09:32, 22 May 2024 (UTC)- The vast, vast, vast majority are ding dongs, where a link to the publisher website is meant to be a link to the book (much like links to Amazon are meant to simply reference the book, not the Amazon listing). There are exceptions of course, but those are the minority, and citing a publisher's website directly is often a failure of WP:IRS (for example here, you really should be citing Financial Times, New Statesman, and City AM directly, rather than the publisher of the book). Adding a comment is a very simple way to prevents bots from fucking up on the minority of cases where such a citation is intentional. Headbomb {t · c · p · b} 09:40, 22 May 2024 (UTC)
- I believe you, and I think I argued a similar point to your position when someone brought the same concern to this talkpage last year. I think I might just be grumpy this morning. Folly Mox (talk) 09:59, 22 May 2024 (UTC)
- @Headbomb, does the bot match this exact text or can I add to it? I would prefer the comment to mention that this comment is specifically for Citation bot, as otherwise I see hidden comments often removed from wikitext in mainspace. czar 14:19, 22 May 2024 (UTC)
- It only cares that there's a comment. See User:Citation_bot/use#..._the_bot_made_a_mistake?, specific parameter vs entire citation. Headbomb (alt) (talk) 14:59, 22 May 2024 (UTC)
- @Headbomb, does the bot match this exact text or can I add to it? I would prefer the comment to mention that this comment is specifically for Citation bot, as otherwise I see hidden comments often removed from wikitext in mainspace. czar 14:19, 22 May 2024 (UTC)
- I believe you, and I think I argued a similar point to your position when someone brought the same concern to this talkpage last year. I think I might just be grumpy this morning. Folly Mox (talk) 09:59, 22 May 2024 (UTC)
- The vast, vast, vast majority are ding dongs, where a link to the publisher website is meant to be a link to the book (much like links to Amazon are meant to simply reference the book, not the Amazon listing). There are exceptions of course, but those are the minority, and citing a publisher's website directly is often a failure of WP:IRS (for example here, you really should be citing Financial Times, New Statesman, and City AM directly, rather than the publisher of the book). Adding a comment is a very simple way to prevents bots from fucking up on the minority of cases where such a citation is intentional. Headbomb {t · c · p · b} 09:40, 22 May 2024 (UTC)
- Pretty grateful we get to go through every citation to a publisher's blurb or library catalogue record and add a hidden html comment so Citation bot can continue assuming we're all ding-dongs who don't know the difference between a book and a website.Having
Past and Present volume and issue numbers
- What happens
- For journal Past and Present, in {{cite journal}} given
|volume=26and|issue=26, bot omits|volume=and keeps|issue= - What should happen
- It should keep
|volume=preferentially to|issue=when the two are identical - Relevant diffs/links
- * https://en.wikipedia.org/w/index.php?title=Pliny_the_Younger&diff=prev&oldid=1223755369
- We can't proceed until
- Feedback from maintainers
Past and Present only has issue numbers. There is no volume number. That's erroneous metadata. See [9]. Headbomb {t · c · p · b} 23:10, 14 May 2024 (UTC)
Don't convert URLs to all lowercase
- What happens
- [10]
- Replication instructions
- This doesn't happen often and I'm not sure what causes it, but in this case it broke the link.
- We can't proceed until
- Feedback from maintainers
URLs are case insensitive (URL : "Although schemes are case-insensitive, the canonical form is lowercase and documents that specify schemes must do so with lowercase letters."), if something broke, it's not because of the casing. Headbomb (alt) (talk) 17:08, 24 May 2024 (UTC)
- Edit, I tried manually setting the casing and... well that website is some kind of special nutjob case because it IS case sensitive. I've never seen that before, ever. Headbomb (alt) (talk) 17:20, 24 May 2024 (UTC)
- @Headbomb (alt): You're misreading your source for that, I'm afraid. URLs are case sensitive.
- 'scheme' refers to the protocol element before the :// , e.g. http, https, ftps. That is case insensitive; hostnames generally are too (because DNS is); other parts of the URL are not.
- Check these out: https://yorril.uk/test https://yorril.uk/Test https://yorril.uk/TEST. This is perfectly normal web server behaviour (out-of-the-box Apache 2). Conventionally, these days, most people make their URLs fully lower case, but they don't have to. TSP (talk) 17:44, 24 May 2024 (UTC)
- The problem is that the parameter https://www.worldjewishcongress.org/en/legacy-of-jews-in-MENA/Iraq’s-First-Minister-of-Finance#:~:text is set to the value Iraq's%20First%20Minister%20of%20Finance%20%2D%20World%20Jewish%20Congress&text=Sassoon%20Heskel%20(1860–1932)%20was%20born%20into%20an%20established,the%20Middle%20East%2C%20and%20Asia. Which leads to some odd GIGO problems. If the the data was set to be a URL instead of done wrong, this would not occur. AManWithNoPlan (talk) 21:00, 24 May 2024 (UTC)
- Yes it has a URL but missing the
|url=key name. These errors are not very common, but I've seen enough of them, they exist. Not sure how to check. Probably slicing the citation string along "|" and making sure each segment contains at least one "=". It's imperfect since URLs (like this one) often contain a "=". If it detects a URL "https?://" then temporarily mask it out before doing the "=" test. -- GreenC 21:57, 24 May 2024 (UTC)
- Yes it has a URL but missing the
- The problem is that the parameter https://www.worldjewishcongress.org/en/legacy-of-jews-in-MENA/Iraq’s-First-Minister-of-Finance#:~:text is set to the value Iraq's%20First%20Minister%20of%20Finance%20%2D%20World%20Jewish%20Congress&text=Sassoon%20Heskel%20(1860–1932)%20was%20born%20into%20an%20established,the%20Middle%20East%2C%20and%20Asia. Which leads to some odd GIGO problems. If the the data was set to be a URL instead of done wrong, this would not occur. AManWithNoPlan (talk) 21:00, 24 May 2024 (UTC)