nimyaml doesn't work on devel anymore

[ringabout]

Issue 1

git clone https://github.com/flyx/NimYAML
nim r test/tserialization

Gives

stack trace: (most recent call last)
C:\Users\blue\.choosenim\toolchains\nim-#devel\lib\core\macros.nim(1685, 10) getCustomPragmaVal
C:\Users\blue\Documents\GitHub\NimYAML\test\tserialization.nim(134, 7) template/generic instantiation of `suite` from here
C:\Users\blue\Documents\GitHub\NimYAML\test\tserialization.nim(559, 8) template/generic instantiation of `test` from here
C:\Users\blue\Documents\GitHub\NimYAML\test\tserialization.nim(562, 9) template/generic instantiation of `load` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(1341, 14) template/generic instantiation of `construct` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(1319, 19) template/generic instantiation of `constructChild` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(1166, 18) template/generic instantiation of `constructObject` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(416, 19) template/generic instantiation of `constructChild` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(1145, 20) template/generic instantiation of `constructObject` from here
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(904, 25) template/generic instantiation of `constructObjectDefault` from here   
C:\Users\blue\Documents\GitHub\NimYAML\yaml\serialization.nim(840, 29) template/generic instantiation of `getCustomPragmaVal` from here       
C:\Users\blue\.choosenim\toolchains\nim-#devel\lib\core\macros.nim(1685, 10) Error: O doesn't have any custom pragmas

It works on Nim 1.4.8

Related PR probably:
#11526
#17725
etc

Similarly
arraymancer, weave don't work on devel too.
mratsim/Arraymancer#505
mratsim/weave#175

[ringabout]

Issue 2

nim lexerTests

/home/runner/.nimble/bin:/home/linuxbrew/.linuxbrew/bin:/home/linuxbrew/.linuxbrew/sbin:/home/runner/.local/bin:/opt/pipx_bin:/usr/share/rust/.cargo/bin:/home/runner/.config/composer/vendor/bin:/usr/local/.ghcup/bin:/home/runner/.dotnet/tools:/snap/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin
/home/runner/work/NimYAML/NimYAML/yaml/private/lex.nim(136, 6) Hint: 'lineWithMarker' is declared but not used [XDeclaredButNotUsed]
io.nim(866)              writeFile
Error: unhandled exception: cannot open: /home/runner/.cache/nim/tlex_d/test/tlex.json [IOError]

nim r tests/tlex works and nim lexerTests works on Nim 1.4.8, quite strange.

[narimiran]

Related PR probably: #11526

git bisect shows exactly this PR that broke nim r test/tserialization.

I'll see if I can find the same for Arraymancer and Weave.

[Araq]

Why is "nimyaml" not in the list of important packages?

[ringabout]

I'll see if I can find the same for Arraymancer and Weave.

Nice!

IMO, the package which calls compiler APIs or Macros deeply is likely to be broken. If the author want to maintain them, they should submit it to important packages(a compile test at least). For instance I ever wanted to try https://github.com/disruptek/dust but it had been broken at that time.

In other word, now I think https://github.com/beef331/nimscripter, https://github.com/SciNim/Unchained and the stuffs of https://github.com/haxscramper should be added to important packages. We can also collect the intentions to important packages periodically on Nim forum or consider the packages on "Nim monthly".

[disruptek]

Sure, because Nim needs more excuses to not fix bugs?

No thanks. Leave my stuff out of it and fix the damned language, please.

While I'm at it, DUST broke because Nim was changed for IC. Where is IC?

[Araq]

While I'm at it, DUST broke because Nim was changed for IC. Where is IC?

It's stagnating, some compiler devs have to fix showstopper bugs, others have moved on to help with "continuations". ;-)

[narimiran]

Issue 2; nim lexerTests

According to git bisect, this was broken by #18100 (cc @timotheecour).

[timotheecour]

fixing issue 2 in #18555