Closed alexnask closed 8 years ago
Oops.
Those failing tests passed at some point earlier in the commits.
That will teach me to not test the final version on my local machine.
Anyway, I'll be on it
Actually, they pass on my machine >_>
Well, that is weird af.
Seriously, those tests run fine on my machine...
Well, I'm giving up for today before I fry my brain.
Cool stuff :D good job
Thank you :)
Hopefully I will find out what is causing Travis to choke on those two tests swiftly tomorrow and have it merged.
What version of rock is Travis using for bootstrap? That might be an issue...
Travis is just using the latest make rescue build.
Anyways, it manages to build itself with c_rock, then build rock with rock, so if it gets up to the tests, it means the resulting rock is good.
pretty sure travis is using quick-rescue? careful
Travis does:
- make -s rescue
- make clean self
- make clean self
So it makes c_rock, c_rock makes rock (make -s rescue), rock makes rock (make clean self), rock makes rock (makes clean self).
I don't think it could be a bootstrapping issue anyway since it is failing on tests that used to pass.
Anyway, I'm on it right now
I can now reproduce the two errors on a fresh install on my linux distro, so I'm ready to go to town :)
I'm merging this but I'm sure there will be tonnes of bugs related to templates to fix, so please report as many as you can if you play with them :)
Things to note at the moment:
Foo new(42) => T == Int
or stuff like that) My focus will now be on #895 which bugged me quite a bit while making this (and there are a couple of workarounds here).
They can be extended from, either partially (e.g.
Bar: class template <T> extends Foo<Int, T>
) or fully (e.g.IntWrapper: class extends Foo <Int>
).In addition to adding type templates, this PR closes #964.
See fasterthanlime/nagaqueen#29 for the (one line) grammar changes.