About the caching problems, i forgot to give an update. I contacted pigeons and we spoke about it, he thinks the problem could be on the reverse proxy in front of the web server (i didn't even know we had a reverse proxy). He'll take a look. I'll keep you updated.

I think we should consider having another person with access to the backend. pigeons is very busy lately and we are behind some updates quite a lot. Like the Onion mirror, matomo (analytics) and now the caching problems.

Caching problems?

You mean the too aggressive caching?

yep

.merge+ 1344

Added

.merge+ 1346

Added

Ops wrong message :/

Looks like some merged CCS proposals are from people asking to do work that was already done by other contributor. I don't know if they put that in consideration, since they were proposing to "translate" not to "review".

i'm leaving some comments here and there, but i don't know how you want to proceed about that luigi1111.

Looks like a bunch of proposals about translations were approved quickly. I hope they were discussed thouroughly during community meetings and not simply "cheap translations: ok", but looking at how some overlaps, that don't seem to be the case.

i'm not in -community anymore and nobody contacted me for clarifications, so i guess we have to deal with the situation as it is

ErCiccione[m]: when updating a user guide, do I have to update all languages?

yes, you can do that immediately, or wait for feedback and then add the changes to all the other languages. I suggest the latter

btw i'm refraining from editing user guides and moneropedia entries for now, unless until we changed the stupid system we use.

ping rehrar regarding translation ccs ^^

I pinged Erciccione in community and during meetings. I thought it was that problem where he shows up as guest again.

I didn't know he wasn't in community anymore.

They were discussed, yes.

Also, ErCiccione[m], the work that was already done seems to have been done by the proposer. In regards to the greek, right?

e.g. german website and gui are fully done

so not sure why they are getting funded again

any others ErCiccione[m]?

rehrar: i haven't been in a couple of weeks, i am in basically all the other channels tho. I have to take a look, i read them only today.

ErCiccione[m]: how do we go about undoing their stuff on weblate? I'll help.

This is kinda good, because we have something to point to when we add the 'prior Monero Project volunteer work required' rule, which is being seriously discussed now

I took a quick look. It's gonna be easy to revert the changes to monero-site, harder with the GUI. Commits for the GUI get squashed on that repository. I think i can manage to revert stuff using the reflog, but it will take a bit.

rehrar: would be useful to have a list of what's overlapping with past CCS and what are the languages that need to be checked

and for which project (gui, cli, site, etc)

the issue seems to be that even before they were merged, they were translating

probably wanting fast payout

languages of merged proposals: German, Dutch, French, Italian

Polish, Serbian, Spanish, Ukranian in Ideas. And since they start before merge, it may be helpful to keep an eye on those.

German, Dutch, French, Italian <- This are all scamming proposal. Selsta noticed they are using an automated translator. i'm checking right now.

yes, I thought that was what we are discussing

it's why I offered to help revert on Weblate if needed. The proposals are bad so they need to be gotten rid of.

Man, that last sentence. You'd think I didn't write.

This is gonna take some time. I locked all translations in the meantime.

Someone added shite to the translations and stuff got added on top ?

ErCiccione[m]: let me know if and how I can help

moneromooo: they didn't get added, but they are commit in the repository on weblate

rehrar: thanks, best would be to take a look at the translations and make sure they are google translated. There are some weird behaviours that i don't fully understand.

Greek for example looks scammy because behave like the scammy users of the other languages, but the translations seem to be legit

Greek was the first one, and was unique for a couple of days. The others all popped in after that copying it. It has a higher chance of being legit.

I'm asking around seeing if there is any native greek speakers that can confirm whether the translation is quality or not.

yeah, point out big blocks of text like this: translate.getmonero.org/translate/g…78439187783762568&offset=622#nearby

seems google translate but then a word or two change.

gotcha

Ok. To sum up: !186 !183 !184 and !181 are scam proposal. They used the deepl translator to make translations in those languages. The github and gitlab profiles they posted are all created around the time they opened the proposal and they are all empty. !180 (greek) seem to be legit.

If i have to guess, i would say the two folks who opened the proposal for greek realized they could game the system and decided to open the other proposals. Of course, it's a guess,

but the behaviour of the users is similar, including the way they posted/reviewed translations.

The GUI, CLI and monero-site repository on weblate need to be fixed. Commits got squashed in some occasions and the reflog is useful.

rolling back the git repos will take some time. After that i will have to remove all their pending suggestions (luckily not many).

They put some effort in this, they really tried to hide their behaviour on weblate and they were quite efficient in that. Luckily selsta spotted a gross mistake that made him suspicious and they got spotted.

To avoid further issues like this i suggest to approve only translators who contributed in past or that can prove their experience, which is the metric i personally usually use for CCS-related translations.

Also, more time before merging would have helped. I managed to take a look at most of the (merged) proposals only today

rehrar FYI ^

and luigi1111 too ^^

Weblate is locked until i fix the mess

ErCiccione[m]: what are your thoughts of banning translation related proposals entirely?

ErCiccione[m]: also, whoever it was went above and beyond. They showed up to the community meeting (like, all of them) and started talking and mingling and discussing their previous experience translating for other projects.

i don't think would be a good idea. Translators and especially reviewers are very needed. Some things will never be translated unless somebody is paid to do it. Some things are very long and other are complex. Volunteer translators tend to skip those.

they had explanations for why their proposals were similar, and answered other questions

I was definitely initially paranoid when I saw all of these proposals be so similar, but they basically said they had worked together on another project, and one told the others about this, and they copied the open proposal (Greek) for formatting.

which, unfortunately, is plausible

yeah, they really put effort into it. It was a properly studied "attack", i don't blame you, wasn't easy to spot.

ErCiccione[m]: ok. Thanks. I'll relay to core, who will be deliberating hopefully soon on this.

ingenious move to have two people per proposal so one can review

Ok. Keep me updated pls

and, if I can be so bold, this attack comes suspiciously close on the heels of all of the other attacks on all fronts of our ecosystem by a particular group

yeah,they also spreaded the work evenly, they didn't just use an user to make suggestions and one for reviewers. They even left some suggestions still pending.

Don't know. My money are still on the two initial greek folks who staged the thing

but yeah, your guess is as good as mine

Correction from my statements before: s/ the reflog is useful./the reflog is useless. it's all cherry-picked commits, so it will take a bit more effort

I think we can easily demand existing accepted translation work ON Monero pre CCS opening

demand/require

yes

it's too hard to track down outside work in a general sense