astral
astral copied to clipboard
astralapp
Infinitely duplicating stars under the untagged category
Astral is loading forever by repeatedly adding all my stars to the untagged stars category resulting in duplicates and making the app useless. As an example:
Occurs in Chrome Version 73.0.3683.103 (Official Build) (64-bit) and Firefox 66.0.2 (64-bit) on Ubuntu 18.04.2 LTS.
Reproducible by loading the page.
Hm, I can't seem to reproduce on my own account. I don't know if it's because you have a lot of stars, but I'm not seeing this issue on my end. Is there any JS errors in the console? I'll check the server logs in case there's something there.
@caccavale Can you try hitting the "refresh" button? That will bust your cache entry and refetch your stars
Hey thanks for the quick response. An aside: I really like the project so I'm hopeful I can use it :upside_down_face:, its funny that the same reason I want to use it might be the problem.
Did a refresh of my stars and watched some console errors:
It seemed to stop refreshing the stars so I refreshed the page and it reproduced the original error. I'm not a huge JS dev but I'm vaguely competent. I'll see if I trace that error.
Just attempted another refresh after I let the page load for a bit. A similar two console errors came up when my Untagged Stars reached about 3000 or 1/2 my total stars. No idea if that is relevant but the page had been refreshing for a decent while.
And:
After the errors the page ends up stuck with ~1/2 of my stars being Untagged when in reality all by 5 are untagged. I'm assuming this puts it into a state when refreshing the page triggers the original bug.
:upside_down_face: Another refresh of the page and then clicking to reload the stars moves progress to
Though sometimes doing the same thing results in the error occurring at a much lower star count. I'll try and reproduce this tonight locally so I get better debug.
So I have a couple guesses what the issue could be:
- All stars get cached with memcached as they get pulled in, so that subsequent reloads are fast. Each memcached entry has a 1mb limit, so there is a chance you're stars actually don't all fit in in the cache object
- Something with GitHub's API. I checked the server logs and noticed some errors coming back right from the GitHub API, but there wasn't any proper detail. I added an extra log check for those failures, so if you could hit try another fetch again, I can hopefully(?) get more detail as to what the problem is.
Sorry for the inconvenience 😞
Don't worry about it! Astral is really cool and your work is greatly appreciated. I'm not surprised my dumb amount of stars is presenting itself as an interesting test case. I'll try and get it working locally and maybe ask some more questions in the process.
