feat: generalize ResponseM to any MonadAff

[cakekindel]

No description provided.

[f-f]

@cakekindel what's the usecase for this?

[cakekindel]

I noticed that Response (and maybe some other structures) are generic over any MonadAff and am running HTTPurple in a custom pipe-based MTL, and figured this would improve the UX a bit.

See CustomStack for the concrete value-add.

I removed ResponseM because it's not super useful anymore but this change does not have to be breaking since it can easily maintain the previous API while just adding the general server' function I wrote.

Note that there has also been an open issue in httpure for several years if you want a little more context: #134

[sigma-andex]

thx @cakekindel , currently on vacation. Will have a look at it next week.

[EmpiricEmpire]

Will have a look at it next week.

Which one, is it the week on September 16 or the one on October 7-th?

[f-f]

I think this is a good change - CI is currently failing and the logs are not available anymore, @cakekindel could you have a look at this?

[cakekindel]

For sure!

…

On Wed, Sep 11, 2024, 10:39 Fabrizio Ferrai ***@***.***> wrote: ***@***.**** approved this pull request. I think this is a good change - CI is currently failing and the logs are not available anymore, @cakekindel <https://github.com/cakekindel> could you have a look at this? — Reply to this email directly, view it on GitHub <#16 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADBCZGB44MQVR5R3QPTO2QDZWBP25AVCNFSM6AAAAABOBFYAUKVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDEOJXGY2DEMBXG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[cakekindel]

@f-f can you approve the CI run?

[cakekindel]

I don't know how the tests ever passed - 2 main issues:

1. mockResponse didn't implement Stream#write correctly, so the tests that tried to write to those objects always timed out. Replaced the record with a MockResponse class extending http.OutgoingMessage
2. all tests that ran serve had race conditions where requests were issued before the server was actually running. For some, I used a serveAwaitReady function that resolves an Aff with {onStarted :: Maybe (Effect Unit)}. For others that already specified that, I just did the naiive thing of spinning until a request can connect

[cakekindel]

just noticed my git & gpg emails mismatched, force push was just changing the committer email

[sigma-andex]

@cakekindel awesome! thanks!

@f-f I don't know what the current process is to publish a new version with the new registry.

[f-f]

@sigma-andex if you cut a new tag it should be picked up by the registry, but switching out the spago.dhall with a spago.yaml should be good for future proofing

[f-f]

The registry currently has a cronjob to pick up all legacy packages once a day, but using spago publish will publish it immediately