Lucas Kramer
Lucas Kramer
After taking a look, the problem is the interfering override of `functiondefs` by `cilk_syncStmt`. This isn't really possible to avoid, since computing `env` at this point requires the value of...
Yeah, the fix for this (a bit contrived) case would be to make `functiondefs` occur on `Expr`. But the problem with the cilk extension (and the halide extension, btw, but...
It appears GCC raises an error with a direct goto into a statement-expression, but they certainly do propagate, as I can do a computed goto! ``` int main(void) { ({l1:...
I am currently working on fixing this in the feature/env branch, amongst some other random env refactoring.
Good catch - if you want, you can try making a pull request to fix this.
I don't know why we wrap declRefExpr in parens- @tedinski? The testing directory in ableC is for interference testing, which has been broken for a while. To actually check if...
If removing the parentheses from declRefExpr doesn't seems to break anything, I would say go ahead and do that. On Jun 1, 2017 1:17 PM, "Patrick Stephen" wrote: > I...
Yeah, handling this specially might be the best option... Currently the halide extension is using some hacks to get around this issue when generating omp pragmas. On Jun 1, 2017...
I don't believe gcc automatically includes this, either. I don't see a reason why we should do otherwise?
Yes, `make` has an implicit rule for compiling .c files that uses that make variable (not actually an environment variable, but if it is provided as an environment variable, but...