29 Jun 2020 |
_hc | for example, I tried a couple times to write tests for the wiki stuff, and it just wasn't feasible | 10:05:53 |
_hc | and every time I touched the wiki stuff, something broke | 10:06:05 |
_hc | we can add the full JSON API without touching the wiki code | 10:06:19 |
_hc | then once we no longer need the wiki, we can do a big purge | 10:06:29 |
Bubu | In reply to@eighthave:matrix.org donno if/how its in use. Seems like something to purge once the wiki goes away. removing it before that could trigger a lot of annoying debuging Huh? The wiki definitely does not fetch anything from f-droid.org. | 10:19:13 |
Bubu | Ah, you mean other wiki code depending on this. I don't think so. Its the last thing that gets done in index.py - main | 10:40:52 |
_hc | the code that writes out that file is the wiki code | 10:57:11 |
Bubu | ...No? | 11:04:05 |
Bubu | Download image.png | 11:04:08 |
Bubu | the wiki code runs after that though | 11:04:44 |
Bubu | I mean, I'm fine with doing away with it later. I just don't think it has anything to do with the wiki 🤷 | 11:05:36 |
_hc | ah, right, its stats, not wiki | 11:15:44 |
_hc | I would only touch code that we are actively working on | 11:15:58 |
_hc | its really the only feasible way to deal with legacy code | 11:16:13 |
_hc | treat the legacy code like an API as much as possible, regardless of how ugly it is inside | 11:17:14 |
_hc | refactor only the parts of the code that are directly being changed | 11:18:22 |
_hc | there are many books and articles on the topic | 11:18:38 |
_hc | so in this case, update_stats is still in use, and unless latestapps.dat is actively causing problems, it should just stay there until that whole block is removed | 11:22:55 |
Bubu | If anyone is interested, I am doing CI with fdroidserver here: https://gitlab.com/Bubu/fdroidclassic/-/jobs/615596931 via https://gitlab.com/Bubu/fdroidclassic/-/blob/master/.gitlab-ci.yml#L25 and https://gitlab.com/Bubu/fdroidclassic/-/blob/master/fdroidrepo/fdroid_gencibuild.py | 12:12:36 |
Bubu | this was a lot more tricky to get to work than it should be :( | 12:12:50 |
Bubu | In reply to @eighthave:matrix.org so in this case, update_stats is still in use, and unless latestapps.dat is actively causing problems, it should just stay there until that whole block is removed it was causing problems in my repo when I enabled stats | 14:23:33 |
Bubu | that's why I looked at it | 14:23:38 |
Bubu | it was writing an empty file and then next run complained about an empty stats fiel being present or something | 14:24:24 |
Bubu | * it was writing an empty file and then next run complained about an empty stats file being present or something | 14:24:28 |
Bubu | and well, instead of fixing it we could just remove this if nobody is using it anymore | 14:24:44 |
Bubu | hence my asking | 14:24:46 |
Bubu | (I just now remembered why I started looking at this) | 14:26:19 |
Bubu | * it was writing an empty file and then next run complained about an empty stats file being present on the next run or something | 14:26:52 |
Bubu | * it was writing an empty file and then next run complained about an empty stats file being present or something | 14:34:44 |
_hc | I don't know if its used | 14:40:24 |