Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 06:43

    blast-hardcheese on master

    Altering pr-label cacher set -e Update sbt-scalafix to 0.9.33 and 1 more (compare)

  • 06:43
    blast-hardcheese closed #1326
  • Nov 28 23:21
    codecov[bot] commented #1326
  • Nov 28 23:21
    codecov[bot] commented #1326
  • Nov 28 23:21
    codecov[bot] commented #1326
  • Nov 28 23:21
    codecov[bot] commented #1326
  • Nov 28 23:20
    codecov[bot] commented #1326
  • Nov 28 06:09
    blast-hardcheese closed #1312
  • Nov 28 06:09
    blast-hardcheese commented #1312
  • Nov 28 06:09

    github-actions[bot] on cli-v0.67.1

    (compare)

  • Nov 26 18:32
    blast-hardcheese synchronize #1326
  • Nov 26 17:55
    github-actions[bot] labeled #1326
  • Nov 26 17:48
    scala-steward opened #1326
  • Nov 24 18:17
    blast-hardcheese labeled #1325
  • Nov 23 23:45

    github-actions[bot] on guardrail-v0.69.0

    (compare)

  • Nov 23 22:49

    github-actions[bot] on scala-http4s-v0.69.0

    (compare)

  • Nov 23 22:48
    blast-hardcheese labeled #1324
  • Nov 23 22:47
    blast-hardcheese labeled #1323
  • Nov 23 22:45

    blast-hardcheese on master

    Removing build.sbt from release… (compare)

  • Nov 23 22:20

    blast-hardcheese on master

    Fix regression tests filtering … Add http4s 0.23.6 support Some final observations around … and 1 more (compare)

kelnos
@kelnos:matrix.org
[m]
hrm, i don't think i see the need for a core package, but i'm not really against it either
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
yeah, that's where I am. We already have some dev.guardrail.core package members
so it's a weird position to be in.
I'll sort it though, no worries.
kelnos
@kelnos:matrix.org
[m]
blast_hardcheese: regarding my PR comment about the slf4j-simple good. what do you think about separating out the CLI into its own module (which can then depend on sfl4j-simple)?
s/good/goo/
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
Wrestling with that now, actually, as part of the repackaging effort
kelnos
@kelnos:matrix.org
[m]
kk
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
I'll keep that as an option. I like it because it means the build tools don't need to contend with the CLI at all, as far as name collisions or anything
the downside is that the gradle plugin "uses" guardrail by just constructing CLI args and running the CLI directly 🥴
kelnos
@kelnos:matrix.org
[m]
doh ☹️ i guess that should be fixed too
(or in the meantime the gradle plugin can just depend on the cli artifact)
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
yeah
I really like the fact that coursier is able to just dynamically link and emit a standalone CLI artifact, btw. No need to publish the assembly like I used to do
kelnos
@kelnos:matrix.org
[m]
yeah, that's really cool
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
OK, here's #2. I split out and cleaned up the backreferences between the CLI runners, so now extending CLICommon is much more clean. Previously it was a mess because parseArgs was incorrectly part of CoreTerms, so we had a circular reference thing that mucked everything up. guardrail-dev/guardrail#1213
Also waffling about whether to merge ProtocolGenerators into one big interface instead of this somewhat split out mess, especially since we have useful copy methods.
Probably do that in a later PR if that does happen.
kelnos
@kelnos:matrix.org
[m]
blast_hardcheese: gave it a quick skim, looks good
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
kelnos appreciated
Nick Hudkins
@nickhudkins
@blast_hardcheese:matrix.org you still at Twilio and if so are y'all still actively using guardrail?
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
@nickhudkins: I'm not at Twilio anymore, but yes, they're still actively using it. A number of others in this channel do work at Twilio, though
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
@sideeffffect 👋 Thanks for your work on this upgrade! Once this is in I'll cut a release with your change as well as the http4s body validation failure stuff, likely later today Pacific US time
Considering we're doing this change all inside guardrail, as well as with Ross Baker's response, I think it should be safe to close the http4s PR as well
Ondra Pelech
@sideeffffect
:thumbsup:
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
0.65.0 released, with the new modules! sbt-guardrail is published, I'll do guardrail-maven-plugin tomorrow
Need to upgrade the syntax on the http4s samples so that the new plugin version compiles with those 🤔
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
Gonna need to get one of these going as well. Plan is to have one release-drafter configuration per module, accumulating PRs that impact that module, and when that release is cut it'll just run the release for that one module. release-drafter/release-drafter#935
unsure what the deal is going to be with the old guardrail module though. I think that'll just end up getting rolled into core, with a provided-scope dependency on the different modules. That way the interfaces stay the same, but when they change in core that'll also trigger a republish of the main guardrail dep.
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
That was much more frustrating than it probably should have been, but I think this should work. guardrail-dev/guardrail#1230
kelnos
@kelnos:matrix.org
[m]
that looks... complicated
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
kelnos Each module tracks commits that roll up to it, based on a file prefix filter. If you finalize the draft release for scala-http4s, it'll just release that one module. The next time a depending module is released, it'll pick up the most recent version of that module from the git tags and use that published artifact
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
I think the thing that would knit everything together would be a github action that just goes "release everything", where it knows about the reverse dependencies of the modules and bubbles everything back up to the main guardrail release
but it was great to see this actually work. I have a test repo here: https://github.com/blast-hardcheese/multi-release-drafter-test/releases
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
(also, aggravatingly, github doesn't permit on: release triggers to be filtered in any way, so my initial assumption about that being clean are also invalid. Gonna have to refactor that to avoid spam in the Actions tab)
but other than that, it does seem to be working. Releasing core now.
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
do'h. Everything worked, except of course the dependsOn versions ended up pulling their git desc version numbers instead of real versions.
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
OK, that time it worked.
kelnos
@kelnos:matrix.org
[m]
nice!
kelnos
@kelnos:matrix.org
[m]
blast_hardcheese: sbt-version-policy? better than mima?
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
I believe mima is used under the hood
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
Yeah, it does kelnos
kelnos
@kelnos:matrix.org
[m]
ah, cool
Ondra Pelech
@sideeffffect

Hello, is guardrail supposed to support specifying multiple response types?
As in https://swagger.io/docs/specification/describing-responses/#media-types

          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ArrayOfUsers'
            application/xml:
              schema:
                $ref: '#/components/schemas/ArrayOfUsers'

https://spec.openapis.org/oas/v3.1.0#parameterContent
When I did that with guardrail, the generated code contained only the last defined

kelnos
@kelnos:matrix.org
[m]
@sideeffffect: java or scala? which framework?
Ondra Pelech
@sideeffffect
Scala, http4s in particular
Ondra Pelech
@sideeffffect
To clarify my use-case -- I would need to have different schema based on the Accepted header, like
          content:
            application/json:
              schema:
                type: object
            application/x-protobuf:
              schema:
                type: string
                format: binary
                x-scala-type: fs2.Chunk[Byte]
blast_hardcheese
@blast_hardcheese:matrix.org
[m]
@sideeffffect: We only generate handlers for content-types that are known by guardrail internals. Without any changes, you can combine manually handled routes with the guardrail generated ones, reusing any guardrail-generated support classes where applicable
blast_hardcheese
@blast_hardcheese:matrix.org
[m]

So an interesting side effect of the module system that I hadn't considered, when a release-drafter release is cut, the tag is minted immediately. This has the effect of all subsequent runs compiling against the ABI exposed by the compiled modules in the source trees, but publishing POMs that refer to the tagged module versions.

The implications of this are that so long as you click release on the various drafts in the reverse-depends order, you can just release everything at once, and as the modules trickle up to bintray, they'll all be pointing at the right versions of stuff