search

gbdev / rgbds

Rednex Game Boy Development System - An assembly toolchain for the Nintendo Game Boy and Game Boy Color
https://rgbds.gbdev.io
MIT License
1.35k stars 172 forks source link

Tests fail on github release archive #955

Closed mid-kid closed 2 years ago

mid-kid commented 2 years ago

Steps to reproduce:

wget https://github.com/gbdev/rgbds/releases/download/v0.5.2/rgbds-0.5.2.tar.gz
tar xf rgbds-0.5.2.tar.gz
cd rgbds
make
cd test/fix
./test.sh

Errors appear, like:

/home/mid-kid/rgbds/test/fix/version.gb out.gb differ: char 333, line 3
./test.sh: ./gbdiff.bash: /bin/bash^M: bad interpreter: No such file or directory

How to fix in repo:

for x in contrib/gbdiff.bash test/fix/*.flags; do
  cat "$x" | tr -d '\r' > tmp; cat tmp > "$x"; rm tmp
done

Though it probably should be fixed in gitattributes.

ISSOtm commented 2 years ago

I guess the fix would be to disable auto-CRLF in our CI script, as well as re-generating all affected tarballs.

mid-kid commented 2 years ago

Regenerating tarballs after this long is kinda evil since most packaging scripts rely on it staying the same for verification. And failing tests is super minor. But if you want a list, it's just 0.5.1 and 0.5.2.

mid-kid commented 2 years ago

Also why not fix the gitattributes?

ISSOtm commented 2 years ago

Because it appears to work fine on Windows installs despite using CRLF, if I am to believe our CI.

ISSOtm commented 2 years ago

As for regenerating the tarballs, you're right.

Rangi42 commented 2 years ago

I think e35585960c5b814a00e216fdc6f7fa1503134250 and 830df360ae7dd5f045d177c5bc57746a9a6d337e resolved this too. Edit: Nope, the reported problems still occur with rgbds-0.6.0-rc2.tar.gz.