View source for Thread:Support/ in messages/reply (2)
In Portal:ksw, the English name for the S'gaw Karen language is not shown, it only have the native name. Please add the English name like any other language portal.
Typo Wikimedia:Wikipedia-ios-diff-thanks-sent/en: "set" should be replaced with "sent".
I spotted an edit which cannot in any sense of the word be considered an translation. This edit is , which adds an Icelandic male name to an field named "Add language". This is grossly inappropriate and I am requesting that his "offline translator" rights are removed due to this. You really can't expect other translators to clean up an mess like this one.
Is this a regular occurence or just a one-off? In assuming good faith, I would believe it's just an honest mistake. I could see myself doing the same thing – thinking I'm copypasting in one thing, then not noticing I copied the wrong thing. It can happen.
It is the worst edit of several bad ones (of those I have found, anyway). It has also happened that I have fixed his translations and then he just re-adds the old translation, that either can´t be considered an translation or contains a word that does not exist in the language.
It would be better if the user had normal translator rights only. An translator can in theory use their offline translator rights so that they use an old translating file and update from that. The problem with that is that that user would end up ignoring other peoples edits while doing so. I suspect that he has uploaded the same offline translator file more than once, and ignored any improvements by others in the proccess. So, I belive he would be an better translator with normal Translator rights only, as it would show him the most recent improvements every time.
It does not help matters either that there is an lack of vandalism tools here. There is no AbuseFilter, TitleBlacklist and so on.
@Snævar: OK, fair enough. But have you tried talking to him to tell about the problem? I can't see anything in the user talk page, but maybe you have some other communication channels. In any case I'm hesitant to call it vandalism, it seems more a case of not being careful enough. If you want I could post something on his talk page, but it might be better if you do it since you know the specific issues much better than I do, obviously.
Not sure where to submit a username change request. So I request it here.
I want to change my username as "A-lú-mih". Thanks.
Hi, can you provide me list of message groups which have less than 10 messages?
just use the Translation tool (from the sidebar).
- Select your language.
- Uncheck the box restricting the list to hide fully translated groups.
- Press the button to get the list of message groups
- Then click on the "Expand all" link displayed
- Click on the column header to sort the expanded list by messages count.
For the translations created by FuzzyBot that are created already fuzzied, the translation interface should be submitting them for our review, because they're fuzzied, but does not.
For example, on 9 May 2019, FuzzyBot inserted the following two translations, with summary "Importing a new version from external source":
- Wikimedia:Wikipedia-library-0907f2-Which authorization method doe/pt
- Wikimedia:Wikipedia-library-fc7cbf-Which authorization method doe/pt
Let's look at the first link, for example:
If you start editing the message directly from this page, it is definitely fuzzied:
But the edit interface does not show it in the list of "Outdated" strings. It is none of the six presented here:
The message appears in the list of "All" strings, as translated and not flagged as requiring any action whatsoever:
Worse still, if you open the message itself here through the translation interface, the software does not even show you that it is fuzzied. The whole !!FUZZY!! thing is totally hidden:
So, just to confirm again, if you go to edit the message directly from the edit interface, as follows (greyed out menu item):
you can confirm that the message is effectively fuzzied:
So, it has been there since 9 May needing review, but not to my knowledge.
Thanks for the report. I'll file task and see if I can mitigate this.
I have run script that fixes this for existing imports. https://phabricator.wikimedia.org/T237412 tracks how to prevent future cases of this. https://translatewiki.net/w/i.php?title=Special:Translate&group=wikipedia-library&language=pt&filter=fuzzy&action=translate contains a lot of fuzzy messages now.
Message group stats was not updated by the script, but they will hopefully "repair" itself over time.
Is FreeCol stopped to update/translate? Because they updated on 2016. If they are really stopped for that, oof.
It's not stopped, they just haven't been able to put out a release yet. We'll consider this in https://phabricator.wikimedia.org/T228072
I’m in conflict with Verdy_p about these messages:
Verdy_p wants to turn
<. However, entities are escaped in these messages where they are included, so the messages are displayed as they are (with entity codes). You can see that on Special:PageMigration (you need to be a translation admin), trying to type anything in the field then submitting.
You give a proof with a specific page used only by admins. But there's no double encoding of ampersands in Mediawiki. Everywhere else the "<" have to be escaped to avoid incorrect matches as candidate tags. valid numeric entities and at leastr the few basic standard characters (which are not "named entities" in HTML, but part of its required syntax) are implicitly recognized (they do not require any DTD, their supprot is mandatory in all HTML and XMS parsers, and in MediaWiki itself): "lt", "gt" for tags, and quotation marks and required for attributes and ampersands for URLs; the support is extended to numeric entities since HTML4 and in all versions of XML and mandatory in HTML5. It is mandatory since ever in all versions of MediaWiki (initially however it had limited support for Unicode and allowed numeric entities mapped to 8-bit only codepoints before UTF-8 became mandatory, which occured when HTML was fixed to require Unicode-only mapping of numeric entities, forbidding mappings to legacy codepages, even if the pages themselves would still be encoded with legacy charsets, this no longer applied to numeric entities). You're living in the early non-standard age of HTML of the end of the 1980's when MediaWiki did not exist, Translatewiki.net did not exist and HTML was full of ioncompatible proprietary versions derived inconsistantly from the initial CERN invention.
I'm not really sure but I have found this ticket on Phabricator: https://phabricator.wikimedia.org/T227244
I don't see any other security-related ticket, tho. However, as far as I know, this extension will be used on WMF wikis do you can assume it's secure enough.
Also thanks for your reply,
I'd say that technically the extension is ready, but the user experience is in alpha stage, so the translation may be also alpha as long as it has not been tested; there will likely be more to translate and explain how this should work and what ot do in case of problems of configuration and why this new Auth method would be preferable or not to another (the Phabricator task explicitly questions about how to work with multiple libraries trying to do the same thing, how they can be unified using a replacable hook allowing to evaluate also the implementations in terms of effective security, performance, cost of maintenance, debugging in case of problems. Also the testbed is still not in place, and they wonder how they'll recruit testers with usable inputs (or if automated monotoring can generate better results). In terms of pure UX, the testbed will receive some comments, several proposals will be made, as well as A/B tests. How many testers will be needed will also condition how many languages will be needed for these tests (they probably don't want to support many languages for now, but have enough inputs for specific needs in some countries where other authentication methods (possibly using thrid party services) are not supported very well, or have no good support and fow which alternate less demanding methods may be needed, but also more training for their users.
Still in all countries, 2FA is not enough known and not generalized, many users complain about how the current methods are implemented, and they also fear about privacy risks this causes to them (notably in countries like China with very active user monitoring by authorities, or with extreme social/economical and criminal risks like Russia or Mexico, or legal risks like US, or countries where users have no other choice to conenct to Wikimedia than using foreign VPNs, or slow mobile connections, or shared Internet access, or whose emails are heavily monitored
As well there's a worldwide threat caused by malwares that attempt to steal people's identity, and 2FA will sonn be not enough, we'll need more factors in a very near future. But how we'll configure all these to work correctly together may become really complex for many people, that will finally opt to protect their privacy by delegating everything to a commercial service provider like Apple, Google, Facebook or Amazon, and tie all their existence to a permanent dependency to these big players, creating lifetime monopoles and splitting the world in distinct proprietary universes.
I'll consider this. It's not on top of the TODO list so it can take a while.
$1 on both, the old and the new username. Shouldn't it use $1 for the first and $2 for the new one?
$2 is the proposed new name, it still does not exist under this name so it cannot be queried for gender. Both names refer to the same user currently named $1, who has the gender settings (these user settings will be applied later to the new name once it is applied; then $1 will no longer have its own settings, it may be deleted/hidden, or kept as a public redirect (possibly protected against deletion or modification of its user pages and talk pages). So this is the same gender of $1 which would apply to both names.
Note that the gender may be needed only in a few language for grammatical forms. If a translation does not need it, the GENDER marker must still be present but may containt empty text (in the English text, the text is the visible user name, not the gender itself, and there's no alternate gender variant of that name and no need to add any other personal gender particle which may be required in addition to the specific user names (I think these gender particles are needed only in some native African languages; in some Slavic languages, there are sometimes variations of user names depending on case on specific constructs, which are also dependant on gender).
during the last few dates I noticed that on the page Special:ContributionScores for my User for Last 7 days (Top 50) the number of contributions is marked as following:
Rank Score Pages Changes Username
49 9 9 9 Robby
while on  you may see that I made at least 50 edits on 31 August till 2nd September.
So there seems to be something broken on this.
49 or 50 edits, this is not a huge difference, and the statistics are not counted immediately; your last few edits were probably made after the statistic batch was run (so it counted only 49). Look at the statistics after 24 hours, you should see evolve it (but the count may still be off by the edits you made in the last few hours after the batch ran). Verdy p (talk) 23:35, 1 September 2019 (UTC)
I've the same problem since the last days of August. I'm on 27th overall ranking but my changes are stopped on 38201 (and the scores is stopped too). May be there's a bug?
You are mistakken 49 is the rank not the number of changes or pages these were 9 and there is quite some difference between 9 and over 50.
OK, but your actual statistics is not 50 but over 70,000. Still my comment remains valid: the statistics are not computed all the time continuously but with batches running periodically, so it's normal that it can ignore your most recent edits over the last shortest period. As well, the results depend on the activity of the job queue (which is very active on this wiki even if there's few edits, as the longest jobs are those created for SemanticWiki, some of them running for hours may lock temporarily not just some pages but also the updates of related statistics (which will be deferred). These statistics are not intended to be actualized in real time, they are just estimates. Computing them has a huge cost on the server, so they cannot be refreshed instantly, even if this is based on a live SQL request (which is already costly and quite long to reply for this page, because it has to lookup all existing users to sort them and then extract about 50 of them in each "top list"; the statistics however are quite accurate for least frequent users, but must frequent users are only roughly sorted, especially for the shortest period of 7 days). Note also that the computed statistics do not show the actual period of dates on which they were computed: the period may be off by more time than what you expect, and the "last 7 days" may be 7 days before the last run of statistics offline jobs which may have occured up to 7 days ago (possibly more of the database had long job lists, with some jobs taking several days to complete and deferring other pending Statistics jobs to run that, most probably, have lower priority than Semantic jobs and regular jobs built in MediaWiki).
This statistics page is not certainly the one to consult if you want to get your own personal edit statistics, that you can get using the live Mediawiki API:
For example: https://translatewiki.net/w/api.php?action=query&list=users&usprop=editcount&ususers=Robby
The same for Joe Taras: https://translatewiki.net/w/api.php?action=query&list=users&usprop=editcount&ususers=Joetaras (note: you must use the actual username, like in your user page, not the displayed username in your signature, and URL-encode it if needed in the URL) This gives a total on all periods. Computing this over a period is much more complex as it requires scanning your history, so this cannot be done directly with the live API. Doing that on all existing users (just to get the Top 50 lists) is even more complicate and costly.
Ok for me. Usually I've checked my stats and I've seen updated. Maybe there's a lot of traffic. No problem. Thank you for your reply
Dear, I must return on this issue. I've understood the performance is important but, if we can't see the updated information about edit count and other, please remove some functionality from Translatewiki, as follow:
The page SupportedLanguages is unuseful because if I check information about an user the information are not update, as the following image:
The page Contribution scores is unuseful because is not update, as the following image:
And these widgets
I'll wait for your response.
Have a nice day
Hi, I have one question. When I created my account here on translatewiki.net, I was asked what languages I know. So I chose English, Slovak and Czech. I translate English messages to Slovak through the Translate window and sometimes I just look for Czech messages that appear on Slovak wikiprojects since Czech is a fallback language for Slovak and because we don't have many contributors in Slovak at the moment, there are many Czech messages appearing on Slovak wikiprojects. I want to first translate all of them to Slovak and then look at the others. However, I can only see English messages and translate them to Slovak (with the help of Czech). Can I somehow see only all to-Czech translated messages which don't have Slovak translation?
Special:SearchTranslations should help, but why not focus on untranslated messages regardless of the status in other languages?
Special:SearchTranslations works only when you know the text of the translation but I need the very same interface as when I translate random messages from English to (language I choose). What would happen if I didn't choose English in the beginning? Then it wouldn't be able to show me English translations as the source. I want to achieve something like that, the source would be Czech. I thought I can change/add languages I know also after the registration but I haven't found anything in preferences...
Why translate Czech translations at first? Because I firstly care about what translations are (mostly) seen, then about the others which might not be seen that often. (As I said, Czech is a fallback language for Slovak and I first want to make sure that Slovak texts (not the Czech ones) are appearing on Slovak projects.)
Right, I had forgotten that we didn't finish the part of the work which involved adding a different source language. The interface is designed to allow it, though, so if you have some programming skills it might be easier to send a patch for that than to use manual methods.
What makes you think that the strings translated into Czech are the most seen? After the group of "top"/most used messages is translated, we have the main Wikimedia messages group which offers plenty of suggestions. Some of those messages are very important, for instance the new search features and notifications which are very visible.
I don't have programming skills to do that so could you please recommend me somebody who would be able to help me?
Seeing those Czech messages on Slovak projects makes me think that they are most seen.
Maybe, if I can suggest something, if it is not too hard and time-consuming, adding an option to switch the source language would be a very useful thing (at least for me).
Who is this user, and why he opens account which is IP address? See Special:Log/newusers.
That IPv6 is "mine", very strange because I'm logged in here.
If you look for it, you'll see that it belongs to the IPv6 space allocated by Free.fr for its fiber accesses (the IPv6 translates algorithmically into a regional location and default IPv4 address, normally, but this does not apply to me because my IPv4 (which is not shared by 4 users like it is by default in Free.fr, which uses NAT by assigning them 4 distinct ranges of 13684 port numbers for TCP and UDP, instead of the classic 65536 port numbers) was allocated to another space where I have a dedicated IPv4 address not passing through NAT.
This is visibly a bug of the "TranslateUserManager" bot, loosing some data about user sessions when we are connected via IPv6, or when some subrequests are coming with different address. Or may be there's a bug in this wiki in its antispam systems, which incorrectly assume that an IPV4 cannot be used simultaneously by two unrelated users. Anyway this is also a privacy bug that should be signaled (it does not conform to Wikimedia policies), and fails to map the related IPv4 and IPv6. I can't explain why this happens, but sessions are incorrectly managed by the frontends of this wiki.
This may be related to recent bugs in this wiki, with failure in its front end caches, also loosing sessions, mixing queries, corrupting the frontend caches with unrelated translations, and not returning the expected data and making incorrect logs for user accounts that it could not get.
I've never opened any such account and in fact the log is wrong as there was no account created at all. All this comes from an internal bot on this wiki.
This bug is visible since last August 15. I never had my IP "logged" like this even if I used this wiki before and did not change my ISP settings or subscription and was active even before.
Everyone sees their own ip. The users in those log entries no longer exist and those log entries should not be there.
What do you mean ? We don't see the same things in logs ? How can you be certain we have the right to see "deleted" log entries? Just because at one time we are using the same IP ? Anyway I do see all these log entries (about 4-5 listed each day since August 15, not any one before, but I see them in the log just since a few day; I many have not noticed that before). So you say that these log entries "no longer exist", but then why do I see them ? Was there a bug that caused IP accounts being created for logged in users, then immediately deleted after logon, but recreated later and repeatedly ? This is clearly a bug: there should not be any account creation before we submit an actual edit (just reading the wiki or logging in should log nothing, or you have a serious privacy issue for RGPD).
@Verdy p: The IP address I see is the one I'm using. It's not an IPv6 address. I agree it's a bug, and should be fixed, but there is no privacy breach when everyone sees their own IP address.
I thought that everyone saw my IPv6 address in this page, which then does not doisplay the same thing for everyone. Anyway these entries are spurious, It reveals that there are effectively lot of logs being made for non-sense account creations by a bot, which then gets rapidly reverted. I suspect this occurs each time we connect on the site. But in fact I tried to disconnect and waited for some minutes before revisiting once, then wait for other minutes, and come back later to login: 3 actions are logged. Any visitor from any IP may have an IP account instantly created, and logged, even if that accoutn is not used at all for any write action, jsut simply visiting. This means that this site is easily attackable: you can have lot of logs filling your disks and the database, and the site won't resist at all. This may explain why the site is sometimes SOOO long to refresh itself and has many antique statuses that can take sometimes days to be reflected.
The source text to translate from English in MediaWiki:Sanctions-row-button-execute/en actually contains a Korean word (not sure if this really means "Execute" as an imperative verb, and no indication for the capitalization of the button). Please fix it, thanks. Verdy p (talk) 01:11, 22 October 2019 (UTC)
Hi everyone! I'm from Georgian wiki. I'd like put some effort in wiki translation but as I figured I don't have rights yet. Can anyone help me out in getting this fixed? Thanks.
This message cannot be correct in French: the adjective must be "actives" after "campagnes" (feminine), but "actifs" after "cours" (masculine).
The English source incorrectly assumes that adjectives are invariant. And the grammatical gender of nouns (like the replaced curse type) is variable depending on each language, do not assume they are simple to predict as this also depends on the choice of translated termes (ex. if instead of "cours", which is masculine with invariant number, we chose the synonym "leçon" or "leçons" in plural, which would then be feminine.
Note that even the plural, marked by the final "s" in adjectives "actifs" or "actives", may become singular depending on the number of campaigns or curses displayed. But as well don't assume that plurals are always fomed using a final "s". If needed pass a number parameter to allow us to use. As there's no number parameter, the unknown number is assumed to be plural (for most frequent cases except initially for a very new program or campaign which is still incomplete if it's limited to only one short course of about half an hour to complete or a single wiki page to read or review).
Please split this message in two versions, one per type, don't use such patchwork. And its not simple to find another way to translate a <adjective+noun> phrase (note also that the order is swapped to <noun+adjective> in French and other languages).
[critical bug] MediaWiki:Withoutinterwiki-legend/fr : translation mixups: updates are going to the wrong page
Translations are sent to the wrong page. Trying to set them correctly, the submission is accepted, but when we refresh the page, we still see an unrelated message, not for this resource, and no change is applied (and no error ever displayed).
And if we look deeper, we see that what was submitted has in fact changed another page, creating a mixup, which then affects multiple other pages.
Very probably problem of index corruption (page or version ID's) in the database (I think there are duplicate ID's in some tables that should be unique), or a previous bad use by some prvilege admin of native SQL update.
As well the histories are mixed up across pages.
A maintenance should be performed to check the database integrity (notably check all table indexes).
It is also likely a problem of the background job queue, which seems to be stalled (recent changes are not fed correctly, statistics are out of sync, page counts are wrong...)n or its backlog of pending jobs is too large, or there's a lack of storage on the server.
I just opened the page MediaWiki:Withoutinterwiki-legend/fr and it really shows the unrelated text, but clicking on "Edit" shows the text correctly.
This was a real bug (https://phabricator.wikimedia.org/T235188) and is what is referenced by the site notice at top. It was affecting transwlatewiki.net stince the beginning of month. The bug was a caching bug on the server. Still it affects some resources displaying incorrect "fuzzy" state (yellow icon) which seem to reoccur again. The issue was real and very strange. Its resolution is partial, as some other glitches are still occuring; the resolution seems to be non definitive because the impact is still not well understood. That's why the site notice at top remains.
Quite certain that these issues started appearing since our deployment last Wednesday (8th Oct, 2019). We should not be seeing this issue on twn any more, but if you do, please ping on the Phab ticket or here.
Actually this started around September 24-26: many edits made after that date (or even reviews) were affected. And this continues even if caches were tentatively purged.
See for example with a new incorrect edit saved based on the wrong unrelated message: https://translatewiki.net/w/i.php?title=Wikimedia:Wikipedia-library-c0f799-Renewal_confirmation/fr&action=history
As long as the new patch for disabling the use of Memcached (by Fuzzybot and other background jobs, including bots exporting data or importing data from external sources) is not applied here (it is still being tested by Mediawiki developers, hoping to find the effective origin of this bug), we still see those bad data. It seems that there's a design bug in how multiple APIs are working together across multiple services, multiple hosts, and in very different security contexts.
The dates of 24-26 September does not match the 8 October deployment. Something occured earlier that had unexpected impact. For now there's no clear isolation of the bug, and there are propably multiple levels of unchecked assumptions between the various components, maintained separately but with a sequencing of events that may have changed and were not supposed to be out of sync or occuring in different orders, because they were not covered by any existing unit tests.
However the bug seems related to a probable bug in internal memory management in PHP for "function closures" (a very tricky piece of code in PHP). The PHP engine effectively used may also have an effect (for example Wikimedia is currently actively working on coming back to PHP7 instead of HHVM, and the past switch from PHP/Zend to HHVM had already caused similar problems some years ago, because PHP is still not defined with very strict semantics). As well there are strange bugs in Memcached itself (still using unsafe assumptions about memory allocation made by C compilers, plus the effects of some deployed mitigations in processor firmwares, used to solve 90% of problems while also adding their own few % of problems).
Thanks Wikimedia works in improving the unitary tests and give more powers to linters to detect false assumptions. But there are tons of compiler/linter warnings to check now: some early solutions to detect them can cause some parts of the software to be simplified, changing the behavior that was previously just "unspecified" but worked as expected. Ideally MediaWiki should be tested by running it on very different architectures and OSes. This is not really the case, but it would help detect more unspecified behaviors and isolate more bugs, while enforcing better software quality by improved portability. However Mediawiki is highly develkoped now in terms of pure performance on the existing and very costly Wikimedia architecture.
But the bug T235188 is now proven to affect also Wikimedia wikis (including Commons, Meta-Wiki and Wikidata, but also some Wikipedias that use the translate extension for some local project pages or for running some tools or because they need to support pages with several orthographies when automatic transliterators are not reliable enough).