#metabrainz

bitmap: Is there something I am missing from Lock? Something in the code perhaps?

Do I have to do Read Lock?

what's the problem with it right now? I'm not sure what to look for

Just doesnt cluster all the files

it might help if I could see a diff of the branch against master

ok

maybe push the branch to github

ok

bitmap ^

yup, looking

LordSputnik pushed 2 commits to bookbrainz-data-js: https://github.com/bookbrainz/bookbrainz-data-j...

is there a reason you're not just doing self.count_dir.int == self.total_dir.int

?

the len(set(dir_progress)) == 1: thing is confusing

bitmap: That makes more sense haha. chaging

cahnged for file too

Now there was a reason. Maximum recursion depth hit

It is virtually the same thing however. Just checks if both of them are the same

how is max recursion depth hit?

let me run the code to see what's going on

ok

send it or you got the chnge?

just copying what's at https://www.diffnow.com/?report=hi0yr

ok

Discourse. :-)

basic setup done, remaining things: setup S3 account for backup

opatel99: oh, I can't enable the setting without the full branch, I guess

configuration (logos, titles, etc...)

see if we can import old forums

bitm let me push

many import scripts exist, but we'll only know if they work for us when we'll try

i'll set up a parallel install of discourse for import tests

bitmap: pushed to pull request

ok

do I need a really big folder for it to break?

ok, now it breaks

I see at least one issue in the code

there's no guarantee that self.count_file.int on the next line is what you incremented it to

you should have the increment method return that value if it's important

also, setting count_file on the instance seems wrong, if it's only applicable to the new_files in that function

a different thread could easily be reading from the self.count_file you assigned from another

zas: Edits for subscribed editors page should be working on beta now.

LordSputnik pushed 1 commits to bookbrainz-data-js: https://github.com/bookbrainz/bookbrainz-data-j...

LordSputnik pushed 1 commits to bookbrainz-data-js: https://github.com/bookbrainz/bookbrainz-data-j...

LordSputnik, might wanna touch the copyright header in the files you added. :)

bitmap: Sorry my internet died.

opatel99, wb

wb?

Welcome back.

Oh haha

So this lock thing. Hair loss level? Or am I making it hard on myself?

I don't think I would've reccomended publishing this as a GCI task :)

recommended?

why?

Recommended, yes.

too hairy, I guess

(Nice mnemonic: related to "commend".)

bitmap: What do I do haha

What if I delay lookup until the end too? The album art is really hogging threads

I'll try to think if there's a simpler approach. this counter stuff doesn't seem like a good idea

bitmap: This is my 6th day on this :/

7th now my bad.

bitmap: Would coming back to this task be petty?

When 3-5 hours turns into 20plus hours, I start to get worried. ^

yeah, idk, it might take me longer than 3-5 hours to fix too

bitmap: You want to test out a tagger.py? :)

I'm not a mentor for this task, though, so I'll refrain from discouraging you any further until Mineo reappears :P

Oh no. I want to finish this. I really do. I just need my Mineo to guide me :). No school tomorrow, so I can stay awake until 3am when he is awake.

Lets try this. http://hastebin.com/raw/cetacunano

it stopped at [359, 490] for me

with 4440 unmatched files left

bitmap: What is the Redis version that’s running on teto?

Version: 2:2.2.12-1build1

Thanks.

bitmap: Do you see anything blatantly wrong?

I'm sort of confused becuase I thought add_files and _file_loaded always ran in the main GUI thread anyway

I have tried ... 47 ways to approach this. None of them fully work.

My closest one was when Mineo only had 1 file load without metadata

and another was when Mineo said it worked, but I had too many variables.

well, do you still have the version that worked?

I can taste the success, just cant savour it yet

Yeah

bitmap: I believe this was the one with too much mess

and this was the one with some minor success: https://gist.github.com/anonymous/40060d5c71ffb...

well the first one still left 3333 unmatched files for me

which I had to cluster manually

bitmap: I am a mess

bitmap: How did the second fare?

one sec

not finished yet, but...it appears to be working much better, if you ignore the maximum recursion depth errors

yeah, looks like it got all the files this time

So...should I take off from there?

I'd probably start from there and figure out why calling self.cluster within _file_loaded is bad

I guess this might be something you already talked to Mineo about, though

bitmap: Are you referring to

that's not the exact error I get, but it's a symptom of the same cause

What would be your approach bitmap ?

hmm

this isn't a good solution, but I'd try seeing if it works after removing QtCore.QCoreApplication.processEvents()

in def cluster

I have a feeling that that's modifying stuff in the same thread we're working in

bitmap: I just tried that. Lots of unmatched files

and still got that loopdy loop

ah :/

Can you try

I wont bother you much after this.

sure

well, on first try it clustered everything without any errors :o

bitmap: Praise?

did it work for you, too?

bitmap: yep. 2 files unmatched, but that may just be the clustering technique