-
-
Notifications
You must be signed in to change notification settings - Fork 373
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix recent JVM breakage (Rakudo::Unicodey) #4290
Conversation
This looks like an oversight from 817d113. The old code augmented class Str which had and attribute $!value -- but the new class Rakudo::Unicodey does not.
We can't even refer to NFC and friends as a return type, since they are completely unknown.
For unknown reasons using native arrays in the core doesn't work. It looks like the array is created as VMArray and not VMArray_i. The resulting error is (in a couple of spectest files): VMArray representation does not implement push_native I'd say we are still in the 'make it correct' phase for the JVM backend, so it doesn't matter if the code runs a bit slower (by not using NQP ops).
So not even an |
I've just pushed 496f986e6c in the hope it will stick on the JVM backend. If it doesn't, let me know, and I'll make it use an untyped array. Thing is that I will need do make all of the |
I'm afraid that is the case. As far as I can see native arrays are mis-compiled into non-native arrays in the setting. So, after 496f986e6c the JVM backend builds, but method
Do you refer to upcoming changes? Because for now things seem to work if an untyped array is used (as in 6014cfa).
|
Hopefully fixed with b88e1ca3bf Please close this PR if so :-) Thanks for the nudging and the PR! |
It looks like we finally found a solution: #4293 -- and it doesn't even require many quirks. The main discovery for me is that
is miscompiled in the setting (on the JVM backend, that is), while
seems to work. That's good to know. I'm closing this PR. |
No description provided.