[WIP] Experiment with generating all API data statically #311
Conversation
|
Thanks for your pull request, @wilzbach! |
source/app.d
Outdated
| @@ -33,7 +33,7 @@ void startMonitoring() | |||
| sleep(1.seconds()); // give the cache a chance to warm up first | |||
| while(true){ | |||
| if (s_mirror.length) s_registry.mirrorRegistry(s_mirror); | |||
| else s_registry.updatePackages(); | |||
There was a problem hiding this comment.
uhh you do realize this code is the same as the production code and you removed package updates with this?
There was a problem hiding this comment.
Yes of course I do. That's why it's called experiment / WIP. It crashes sometimes when trying to update the registry while wget/curl is fetching. I just wanted to see how far I can get with a simple curl loop.
Anyhow, I can/will remove this once --no-monitoring is a thing (-> #310)
So I guess this would have to use a different approach, like
|
|
how about limiting packages to a select few? I doubt we will need all packages. And whenever packages are needed for a test they can get added (such as when people might want packages with different readme names, with different languages, with different metadata, different logos, different repo types, etc) |
|
This is not about previewing PRs, this is about having a static registry mirror on a CDN. See: #309 |
|
oh ok nvm then, though for netlify PRs I think that would be a good idea |
FYI: They are already limited because at the moment there's no way to reach a package if not via the main list view and dependencies of such a a package. So if you go to https://dub.pm at the moment (that's the static snapshot = to what you see on the previews), only ~ 20 packages are reachable. Maybe a bit more with their dependencies. |
|
Not sure that's the best move, #309 (comment). |
See #309
A start, probably
curlis too slow and this should would need to be done directly in the registry.TODO: