|
Hmm? This does work for me (tried just today!). I've experienced this bug only when using the "Discogs import userscript". You are right I was using the Discogs Import Userscript, so the problem must be to do when seeding the data but that doesnt mean there isnt a problem. Aurélien - are you saying this used to work with the Discogs Import userscript, or just in general? I was just saying this used to work in general (at the time of my last comment we were not aware that is was related to seeding of release editor with parameters). Reproduced this on the current server, using with the following steps: 1. install murdos' import script from http://userscripts.org/scripts/show/36376 No recordings are selected, it seems to use the seeded tracklist over the tracklist selected on the duplicates tab. If you try the test case again but select the 3 track release I think the problem is clearer. But this use case, also shows a more serious problem: Im not sure whether in this case it should show nine tracks, or three tracks but clearly should be consistent. My preference would be it shows nine tracks, with the first threee tracks replaced by the release we selected to base it on, and the other six tracks untouched. I added a fix for the issue as I described it in master as commit 121294d529fa138c64c803e01c5414d3a2475604. But I believe that fix doesn't solve the issue as it was described by Paul in the previous comment, so I'm not closing this ticket. In my tests it now selects the actual tracklist from the Duplicates tab, throwing away any seeded tracks. IMO this is correct behaviour, and the bug as described in the ticket is fixed now. I can understand that you would like additional seeded tracks to show up in the tracklist, it sounds useful in some cases (adding a release with a bonus track), but it also sounds potentially confusing. In any case, I suggest a separate ticket should be opened for that. |
||||||||||||||||||||||||||||||||||||||||||||||||||
Anyone monitoring Jira and allowed to set priority and milestone, please make this important: it used to work, so this a regression, and Paul is at least the second user encountering this behavior.