Closed mswsn closed 1 year ago
Following because I get this too. Can't find a pattern as to when. Might look into it codewise later
Looks like it crashes when it encounters a recaptcha URL. Which explains why this happens more often the more I keep importing. This was a crashing results
object from around line 214
results ['https://greenovamusic.bandcamp.com/album/back-ellin-crack', 'https://www.recaptcha.net/recaptcha/enterprise.js']
@mwassen I was able to temporarily make the crashes stop by setting search_max
to 1. Not great as it limits to one possible result, but recaptchas get filtered out.
Cheers, at least good to have a band-aid for now.
Sorry this took ages to respond to, currently looking into it!
@tomauty you are completely right regarding the recaptcha url! See below for a confirmation of your issue (see results number 7 and 18)
Have fixed it, see the equivalent below
Will get it released in a moment.
Done, try updating beetcamp to 0.16.1
!
I am still constantly getting a url error with beetcamp, have updated to 0.16.1 as far as I know with a direct git clone to the beets environment (is there a cli way of getting the version?)
I am not sure if it is the same issue, due to recaptchas. However, the traceback looks basically identical:
The error happens consistently on this link: https://derleemusic.com/album/off-the-beatn-path-2, while other albums by the same artist worked.
I tried to look around a tiny bit myself but did not get far with any coherent reasoning unfortunately.
Same error here. It's actually due to an artist's use of another domain name : For instance if you search for "Colors in the dark", the first bandcamp result will point to https://3six.net/album/colours-in-the-dark, which is not on bandcamp.com, same as your album link (but use the same backend). Then the regex defined in search.py is not matched, and the url empty, leading to this error...
@snejus imho, I would just change this regex to a simpler one with just 'http' and some query like '?from=search', given that all search results are likely to be bandcamp albums...
Okay, seeing this is re-appearing I am now testing the search against all album names I have in my beets database (1500+).
I can see that the first 300 searches gave me 80 results with a wrong URL (missing or some irrelevant HTML), each of which would cause the issue you above. This gives more than enough data to work with.
@shagr4th thank you for your input - ?from=search
is a great idea - see my comments under the PR. Aiming to get the fix out tonight.
@snejus Sorry to be off track, but what's the table
command you're piping to to show the response data?
@tomauty, it's a JSON prettifier based on rich. Since it's now in a somewhat of a more shareable state, see rich-tables. Early version is on pypi, so pip install rich-tables
should give you the executable. There aren't any docs or usage so far, but see the tests/json
folder in the repo and README
to see what they give. Then pipe in some data and see what you get 😉
This should now be fully fixed in 0.16.2
!
I'm occasionally getting crashes while using beets on an rpi4 and the culprit seems to be the beetcamp plugin.
Sometimes the hangup seems to be on particular albums and I'm forced to disable the plugin until that one is cleared. I've also noticed that the error always tends to occur if I have left beets alone for a few minutes.
Here's what I'm getting when using the
-v
flag: