I use the name of my bot here on Phabricator as a username, because accented letters are prohibited for some reason (hey, it's 2019!). Sorry if this confuses somebody. I updated my profile to make the use of singular clear. :-)

Why do you think so?
Limit is more familiar to people because of SQL use, and this is a similar role.

After all, the renaming was correct. :-)

Now I see. Someone implemented this behaviour on purpose here:
https://github.com/wikimedia/pywikibot/blob/master/pywikibot/page.py#L1272
But why?

I think this should alwqys remain in the scope of the script. Replace.py could show me the proposed change, then I, as a human editor may decide what to do. (Edit the page manually, tell the other editor who placed the teamplate etc.) Suppressing spelling errors on the page is usually not the primary aim of using such templates. Also, I often use replace.py just for collecting pages with a specified text or pattern, which does not affect the content of the page. Why should page.py hide these pages from me?

In T205205#4608866, @Framawiki wrote:

This page according to the backlog provided.

Pywkibot does check the templates, otherwise you wouldn't get this message.
As far as I see, replace.py at line 819 does not handle exceptions at all, but I am not familiar with asynchronous save.
Please provide the link of the page where the exception occured.

Please consider that

• {{done}} or any similar template may have parameters (at least |~~~~)
• some delay may be neccessary and should be implemented (we don't want people to mark sections as done at 2:00 am and the bot to archive them at 3:00 am before anybody could have seen the changes). I think it So "algo = old(30d) OR marked(done|2d) would be better (where | is not neccessarily the best choice for separator character)
• {{done}} or any similar template also may appear in an inactive position (HTML comment, nowiki, quoting template etc.).

I wrote a related script. Unfortunately it is hardcoded for Hungarian Wikipedia and therefore it has a lot of Hungarian text (as I didn't think at that time, that it would be useful for others).
https://hu.wikipedia.org/wiki/Szerkeszt%C5%91:BinBot/frissdump.py
This script does not download the dump, but refreshes a template with the date. Users interested in dump watch this template. So the first part of the task is solved: to watch the new version. It needs cron timing.
Somebody may find it useful.

Actually, T100360 is about deleted revision, but with deleted edit summary the situation is really the same.
Example: https://hu.wikipedia.org/w/index.php?title=S%C3%BClys%C3%A1p&curid=45101&diff=19117621&oldid=19117618

Thanks!

I didn't contribute to bot.py, but someone will surely explain the reason.

In T147178#3567967, @matej_suchanek wrote:

The funniest fact at the end: this script has sync/async saving swapped, compared to other standard bots. In other words, with -always ([a]) it saves synchronously, in manual mode ([y]) asynchronously.

In T147178#3568092, @Framawiki wrote:

As this bug is still alive a year later, and it only concerns an output that is not essential, I'll submit a patch to temporary disable it.

"The form will replace the content of the account's user and user talk pages as requested and expected"
Does not work for me.

Some use cases that are not covered by multiple fixes:

Could you provide that custom fix? NOT the screenshot, rather the text.

The video is not clear enough, the text is not readable and the music gives no plus. Could you please share the exact command or the fix you used? Let's forget this video thing.

In T99365#2809197, @Multichill wrote:

In T99365#2809160, @XXN wrote:

Are there some statistics on the compat usage in last 3 months?

Given the fact that it's broken on WIkimedia wiki's, I assume no edits.

Are these all autogenerated docs?

OK. Replace.py is very important for me as I heavily use it for multiple purposes, and I am highly interested in its performance. I contributed a lot to the compat version, but now I have troubles both with using the core and with coming back to development, but I will look inside the problem when I am able.

In T57078#3337396, @Dvorapa wrote:

In T165911#3296955, @valhallasw wrote:

Then a better piece of advice would be "Remove u only from scripts with no pending changes", wouldn't it? I think a patch only removing us can be revised on express lane.

@Xqt wrote at the mentioned page: "Anyway we shouldn't change old code for it". Why? Us will always cause problem in Python 3000.

Will this transition affect bots, or will the API hide the whole change?
Is there any chance to enhance the 255 characters limit?

Per @Pmlineditor's comment

Limit is more intuitive in my opinion.

Abbreviation of IPv6 decreases readability more than the idea of Tgr. Do we know about automated tools (AWB, PYwikibot, Lab tools etc.), how they recognize IPv6 and have a method for expanding? IPv6 usualy appears in total length in logs so they may have not had the reason to develop it. Anyway, modifying the message may be a breaking change for some tools. Perhaps it would be a good idea to prexfix abbreviated messages for the sake of bots and tools with a special character, so that they spare the analysis of every line.

I think, the link from user:$1 may also be omitted if necessary. Watchers of recent changes usually want to know about the guy whose edit was reverted, sometimes about the reverter. The previous editor is uninteresting and so is worth two clicks (the first being the difflink) in the rare cases.

Sure. :-)

If T6714 will be resolved, this becomes obsolete. Otherwise this may need another solution (e.g. reducing links).

T145604 may be relevant. Is it worth to deal with it in two steps?

Thank you!

Hungarian Wikipedia uses this template to prevent the section being archived even if expired:
https://hu.wikipedia.org/wiki/Sablon:F%C3%BCgg%C5%91ben
It may have a parameter.
Currently we use an individual workaround for that, but we woluld like to do this with the framework.

Thank you. You are basically right, but I would prefer not to use -user if the majority of pages can be treated as bot (for keeping recent changes readable).

I also don't know how much experience you have with frustrating console encoding problems, but I have suffered from them for years. While saved fixes always worked well, it was not the same with command line arguments. I had really, really many problems with that. Now, -save opens a relaiable UTF-8 file using codecs. That is proved. I don't feel the energy for trying new methods and experimenting again once something had already worked fine.

I don't believe that listpages.py can handle this funcionality: https://hu.wikipedia.org/wiki/Szerkeszt%C5%91:Bin%C3%A1ris/Fixes_and_functions_HOWTO

I will set up core but that takes more. For the moment I visited https://www.mediawiki.org/wiki/Manual:Pywikibot/Scripts which states: "Compat equivalent: pagegenerators.py and get.py." I also looked into the code and it does not have the complexity of fixes. I don't need a simple pagegenerator for this task, I need replace.py with -save that uses the same fixes and exceptions (included T144693) that the real replacements use. My fixes are rather sophisticated and I don't see any sign in listpages.py that it could provide this level.

I don't know listpages as I always used compat.

@jayvdb, done, you see the mentionings above. This is my response to your "add a comment here with the top 5 missing features, so we can fix them first" sentence.

See also T144692.

I know a workaround for this in compat, but that feature has not been ported to core yet. Hopefully will be soon...

@yuvipanda: You don't have to know Python, you only have to know regexes, and copy the function. Thanks for quoting! However, there is a possibility to enhance the bot with this feature.

I also implemented the 'include' key to 'exceptions' dictionary in replace.py. That is very much missing from core, too. Lack of these features (with the above mentioned) makes core totally useless for me as I would face all the difficulties with that I began to work back in 2008. All these should be just ported from compat.

solve_disambiguation.py is also missing -first parameter that was implemented in compat.

First of all: I implemented many new parameters to replace.py. -save, -savenew, -savexc, -savexcnew and edit counters which I remember now. These are missing from core, and I would really appreciate if someone ported them.

Dear Merlijn

Merlijn: the real problem is that on one side developers say "no, nobody declared that core has to implement full funcionality of compat" and on the other side compat issues are automatically set to lowest priority.

The problem definitely exists. I stull cannot log in with my admin account on huwikipedia.
I do nned the bot to delete plenty of pages, but I can't.
It is a shame that compat users are handled as secondary despite the fact that everyone knows, that core does not have a full functionality.

"Is this strange behaviour something we want to preserve, or is it just a historical quirk?"
Or just survived tests, as a rare case.

OK. Thanks for the detailed explanation.

Sorry, you are right, I used the wrong option.
However, after undeleting page it listed tha other pages which refer to the deleted one.

I just met https://phabricator.wikimedia.org/T57689 and was surprised to see that some existing funcionalities of compat are not inplemented to core, but may be requested as extra feature.

There are four similar parameters in compat, and all of them are required for usability of core. I added them to trunk someday and I use all of them on a regular base. They are -save, -savenew, -saveexc and -saveexcnew. Other option is not to deprecate compat. :-)
These options are really useful and neccessary for everyday spelling correction tasks.

Either on Checkuser-l, or I am wrong.

@MZMcBride: As far as I remember, this was discussed on mailing list. The tool should remain in the scope of checkusers. AbuseFilter is open to too many people, even if the filter is hidden. I think data protection wins here over pure technical consideration.

Some UAs are really common, and would match plenty of users. How could we define a minimal expected complexity?

I will keep in mind and suggest the idea if I have one. :-)
For example the replacement could depend on other parts of the same text (not just as 'text-contains').
As I use compat, I didn't know this feature of core, and I have no experience with the speed of using multiple fixes vs. one fix at the moment.
Title may still be useful, because separating one fix to multiple means you may duplicate/multiplicate other parts of fix (edit comment, other exceptions), what is usually thought to be harmful redundancy in programming.

I will check it. However, being able to use the page object in a function gives much more flexibility.

Xsize has the point. :-) I run several regexes within a fix and I don't want to separate them (details: http://wikimedia.hu/w/images/9/9b/Efficient_and_flexible_text_manipulation%2C_spelling_correction_and_page_collections_with_Pywikibot.ppt).
Only one of them should not work at only one series of articles. E.g. the title
ur' \(döntő\)$',
should be excepted for this pair of (old, new) while not for the others.

That was really quick, thank you! I wrote general as project because I would like to see it on compat, too. Could you port it to compat please?

I want to search in namespace 1 and get results from namespace 0.
Thus I should use -namespace twice which is not possible.
Can your proposal cover this case?
I imagine a clean ToggleGenerator that wraps around any other one.