DCE cleanup 2024

[Simn]

1. Add CfUsed and CfMaybeUsed to get rid of @:used and @:maybeUsed for fields.
2. Add CUsed to get rid of @:used for classes. The meta still exists for enums and abstracts because those don't have flags.
3. Remove m_if_feature, instead have DCE check @:ifFeature metadata itself.
4. Remove the removal of (now) CfMaybeUsed at the end of DCE. This is set by the typer, so removing it in post-processing is wrong.

The unit.js diff against development shows the removal of unit_issues__$Issue2785_Child, which makes sense to me because it's unused and the only @:keep exists on its parent class. The dump diff of the JVM tests shows 3000 removed @:used metadata and 45 removed lines. 13 of those are from the Issue2785 thing, and the other 32 are... I don't know where.

Edit: I think my diff tool was lying to me, if I avoid printing @:used there are 0 differences.

[Simn]

The JVM tests on windows 64 fail consistently, so there's a chance that it's related to the changes here somehow. I can't make sense of the CI log though:

test jvm
Changing directory to D:/a/haxe/haxe/tests/unit
Command: haxelib [git,utest,https://github.com/haxe-utest/utest,424a7[182](https://github.com/HaxeFoundation/haxe/actions/runs/7494491280/job/20414033497?pr=11485#step:13:183)a93057730fada54b9d27d90b3cb7065c,--always]
You already have utest version git installed.
Installing utest from https://github.com/haxe-utest/utest branch: 424a7182a93057730fada54b9d27d90b3cb7065c
  Current version is now git
Done
Command exited with 0 in 1s: haxelib [git,utest,https://github.com/haxe-utest/utest,424a7182a93057730fada54b9d27d90b3cb7065c,--always]
Command: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
You already have hxjava version git installed.
Updating hxjava version git...
Command exited with 1 in 1s: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
Command will be re-run...
Command: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
You already have hxjava version git installed.
Updating hxjava version git...
Command exited with 1 in 1s: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
Command will be re-run...
Command: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
You already have hxjava version git installed.
Updating hxjava version git...
Command exited with 1 in 1s: haxelib [git,hxjava,https://github.com/HaxeFoundation/hxjava]
The system cannot find the path specified.
javac 1.8.0_392
Cloning into 'java-lib-tests'...
The system cannot find the file specified.
Error: std@process_exit
Error: std@process_exit
An exception occurred in a neko Thread :
An exception occurred in a neko Thread :
std@lock_release
std@lock_release
Error: std@process_exit
test jvm failed
Error: Process completed with exit code 1.

[Simn]

Uhm okay, random commit fixed this... no idea what that was.

[andyli]

api doc gen was broken since this.

https://github.com/HaxeFoundation/api.haxe.org/actions/runs/7499155735/job/20415432376

              +gen.n *failed* | --> RUN haxe gen.hxml
              +gen.n *failed* | Error: Compiler failure: haxe.ValueException is missing field "value"
              +gen.n *failed* | Please submit an issue at https://github.com/HaxeFoundation/haxe/issues/new
              +gen.n *failed* | Attach the following information:
              +gen.n *failed* | Haxe: 5.0.0-alpha.1+ba272a6; OS type: unix;
              +gen.n *failed* | File "src/filters/exceptions.ml", line 387, characters 75-82
              +gen.n *failed* | Called from file "src/filters/exceptions.ml", line 387, characters 22-82
              +gen.n *failed* | Called from file "src/filters/exceptions.ml", line 455, characters 5-70
              +gen.n *failed* | Called from file "src/filters/exceptions.ml", line 488, characters 14-78
              +gen.n *failed* | Called from file "src/filters/exceptions.ml", line 559, characters 5-46
              +gen.n *failed* | Called from file "list.ml", line 92, characters 20-23
              +gen.n *failed* | Called from file "src/core/texpr.ml", line 131, characters 26-41
              +gen.n *failed* | Called from file "src/core/texpr.ml", line 140, characters 49-61
              +gen.n *failed* | Called from file "src/std.ml", line 26, characters 6-9
              +gen.n *failed* | Called from file "list.ml", line 121, characters 24-34
              +gen.n *failed* | Called from file "src/core/tUnification.ml", line 462, characters 10-15
              +gen.n *failed* | Called from file "src/filters/filtersCommon.ml", line 74, characters 23-91
              +gen.n *failed* | Called from file "list.ml", line 110, characters 12-15
              +gen.n *failed* | Called from file "src/filters/filtersCommon.ml", line 79, characters 2-45
              +gen.n *failed* | Called from file "src/typing/macroContext.ml", line 663, characters 2-54
              +gen.n *failed* | Called from file "src/macro/eval/evalPrototype.ml", line 319, characters 3-11
              +gen.n *failed* | Called from file "list.ml", line 234, characters 17-20
              +gen.n *failed* | Called from file "src/macro/eval/evalPrototype.ml", line 308, characters 17-801
              +gen.n *failed* | Called from file "src/macro/eval/evalMain.ml", line 393, characters 72-99
              +gen.n *failed* | Called from file "src/macro/eval/evalExceptions.ml", line 101, characters 5-8
              +gen.n *failed* | Called from file "src/macro/eval/evalExceptions.ml", line 116, characters 10-38

[Simn]

Thanks for catching that! I just tried to clone api.haxe.org but I'm having some trouble:

Receiving objects:  57% (14123/24776), 146.02 MiB | 1.80 MiB/s
Receiving objects:  58% (14371/24776), 146.02 MiB | 1.80 MiB/s
Receiving objects:  58% (14499/24776), 149.04 MiB | 2.31 MiB/s
error: RPC failed; curl 18 transfer closed with outstanding read data remaining
error: 719 bytes of body are still expected
fetch-pack: unexpected disconnect while reading sideband packet
fatal: early EOF
fatal: fetch-pack: invalid index-pack output

I've attempted a blind fix here: f6805db

[Simn]

Right, it wants --deph=1. Doing so lets me run haxe dox.hxml, haxe gen.hxml and then neko gen.n without issue. Is that the command CI is failing on? I can't quite tell in that log...

[kLabz]

That was #11477 (5.0.0-alpha.1+ba272a6 from above logs -- edit: and last successful build was with dec37b31a0af73de9ec59db040ce73e6cc37b345 so there's a few more commits in between) which happened to break "old" versions of hxparse which was fixed in Simn/hxparse@876070e

I updated the submodule, let's see what CI thinks ~ (still broken, though that does seem to be similar to what I have locally which does not break 🤔)

[Simn]

@andyli Both Rudy and I cannot reproduce this locally, and we can't tell what state the server has exactly when encountering this problem. Could you give us a hint how to replicate the behavior?

[andyli]

I merely discovered it since I got GitHub Actions emails about the broken build over api.haxe.org. I don't have much idea on how to reproduce it.

I tried it locally and here are what I discovered:

• Turn out that error already happens since ba272a6, but dec37b3 still works. I'm not sure about the commit in-between them since there are no builds on build.haxe.org.
• Somehow the error only occurs when HAXE_STD_PATH is NOT set.

[kLabz]

@andyli Should be fixed now; api.haxe.org CI is green again :)