Hi,
This is my weekly report:
---
This week I set out to:
- Implement the parsing of {init,set,shutdown} directives for the C
and C++ front-ends.
- Continue with the parsing in the Fortran front-end, based on the
question I posted previously.
Over the week:
- I reviewed the code and implemented pragma directive recognition
in the C++ front-end. However, I'm currently on the learning curve
of the front-end and experiencing some unexpected behavior during
parsing. As a result, most of my time was spent debugging the
parsing for these directives.
- I started looking into how other directives resolve clauses with
an enum type argument declared in Libgomp (Fortran front-end) and
began implementing the integration of OACC_CLAUSE_DEVICE_TYPE using
omp_mask to accept the OpenACC device_type clause.
- I posted a question regarding this
(https://gcc.gnu.org/pipermail/gcc/2026-May/248262.html), but I
think I'm figuring out the answer on my own. I'll likely wait
for the code review to agree on the naming of the new clause
inside omp_mask.
Current status:
- The C++ front-end recognizes the {init,set,shutdown} directives,
but there are still some bugs that I'm currently fixing.
- Fortran now accepts the device_type clause, and I'm currently
working on including the arguments in the parsing logic.
I'm taking one week longer than expected (compared to my formal GSoC
proposal), but I remain positive as this is mainly because of the
front-end learning curve. Regardless, I will put a bit more effort to
finish implementing these directives and move on to the next agreed
task.
---
Best,
Sebastian Galindo
