You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some intentionally-error-triggering codegen tests are failing since error_status is not being updated (it stays at 0 ('clean')). Same goes for global_error_status. Stepping through some of these tests in GDB reveal that they are in fact going to the handlers, but error_status fails to recognize that.
Possibile culprit(s):
Not being updated in the error handler(s)
Not being 'passed along' all the way down and gets overwritten
Possible solution(s):
Have error_status be specific to each module/structure/compiler pass:
Have an expr_error_status, decl_error_status, etc. for each of the passes
Or typecheck_error_status, codegen_error_status, etc.
Pass it in as a pass-by-pointer parameter instead of returning it
This is a minor issue. This does not block overall code generation (if anything it's just a losing a sanity check or 2).
The text was updated successfully, but these errors were encountered:
Ttile.
Some intentionally-error-triggering codegen tests are failing since
error_status
is not being updated (it stays at 0 ('clean')). Same goes forglobal_error_status
. Stepping through some of these tests in GDB reveal that they are in fact going to the handlers, buterror_status
fails to recognize that.Possibile culprit(s):
Possible solution(s):
error_status
be specific to each module/structure/compiler pass:expr_error_status
,decl_error_status
, etc. for each of the passestypecheck_error_status
,codegen_error_status
, etc.This is a minor issue. This does not block overall code generation (if anything it's just a losing a sanity check or 2).
The text was updated successfully, but these errors were encountered: