12 Oct 2020 |
_hc | or https://f-droid.org/repo/status/build.json | 14:52:59 |
_hc | oops, seems its not in the JSON. should be IMHO | 14:53:11 |
mimi89999 | Oh. I forgot that it was there. | 15:27:26 |
jochensp | mimi89999: it's in there? Which key? | 15:28:43 |
mimi89999 | On the page `guest RAM | 15:29:08 |
jochensp | ah | 15:29:30 |
13 Oct 2020 |
| marzzzello joined the room. | 10:23:12 |
| marzzzello left the room. | 10:23:12 |
mimi89999 | uniq: So you also have a QEMU buildserver? | 19:41:57 |
uniq | mimi89999 I can bootstrap one if I want to, but I usually need my disk-space for my day job | 19:47:04 |
mimi89999 | I can wait for Bubu... | 19:47:57 |
wb9688 | Isn't hc more the server guy? | 20:06:27 |
| marzzzello joined the room. | 20:13:45 |
| marzzzello left the room. | 20:13:46 |
14 Oct 2020 |
| marzzzello joined the room. | 01:14:14 |
| marzzzello left the room. | 01:14:15 |
| marzzzello joined the room. | 09:20:34 |
| marzzzello left the room. | 09:20:34 |
smichel17 | Bubu: Is there a reason data processing happens in this order instead of handling triple-t and fastlane metadata at the same time? | 15:22:46 |
smichel17 | https://gitlab.com/fdroid/fdroidserver/-/blob/master/fdroidserver/update.py#L2247-2252 | 15:22:54 |
smichel17 | (eg, instead of nesting them both under the same if repodir == 'repo': ) | 15:23:32 |
smichel17 | Aside, I don't work with python much — what's the recommended way of installing dependencies not-systemwide these days? | 16:11:50 |
wb9688 | venv | 16:13:38 |
wb9688 | `python3 -m venv /path/to/venv` to create one, then `source /path/to/venv/bin/activate` to activate it, then you could just use `pip install whatever` to install whatever into the venv | 16:14:53 |
wb9688 | (I'd recommend running `pip install wheel` before installing any other package into the venv) | 16:15:24 |
_hc | smichel17: the order of operations in fdroid update is basically just organic, as things were implemented. I'll bet the order there was not a design decision as much as "it worked that way, so let's keep it that way unless there is a reason to change it". Then to test this, we can do before/after runs of fdroid update and compare the resulting index-v1.json | 16:27:57 |
smichel17 | okay. I'm just looking through the source now and figuring out the best way to approach things. Currently feeling some tension between "the current approach is quite fragile" and "I'd like to avoid large changes if I can help it". But I think it might might sense to do a larger refactor and perhaps unify how triple-t and fastlane metadata are handled. | 16:36:00 |
smichel17 | Also we are currently doing some metadata merging, so it turns out we will need to decide what to do there, after all. But, I will bring that up in fdroidserver#829 when I've thought it through | 16:37:16 |
[gibot] | [server] #829: Support additional fastlane metadata locations - https://gitlab.com/fdroid/fdroidserver/issues/829 | 16:37:26 |
_hc | I'm open to a refactor there, the current implementation leaves a lot to be desired. I didn't have many example apps to work with when I started it | 16:41:55 |