This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] gc-improv merge (1/n): new libiberty interfaces
2010/5/5 Basile Starynkevitch <basile@starynkevitch.net>:
> I really think that you should leave everywhere the formal arguments
> names for documentation purposes. (and BTW, there might be a missing
> space before the left parenthesis above).
Thanks, will fix.
> So please add the formal
> arguments name (like size, hash_f, eq_f... above) or at least explain
> why you don't want to add them.
Unfortunately that's not an option, see DJ's reply.
> At last, I would like some descriptive comments if possible. I am not
> entirely sure to understand exactly what you want to do -especially
> without formal arguments names, and I believe a newscomer also could
> be lost.
I am all for additional comments/documentation, however I am not sure
what additional comments would help here. The data structures
previously had a single callback for all their allocation needs, now
they have one callback per type.
> Perhaps also the htab_create_alloc function should be
> documented as deprecated (in hashtab.h) if you believe it is becoming
> deprecated.
It isn't. The patch only adds a new way to allocate memory for data
structures. The old ways are perfectly fine.
> I also don't have a clear picture of when and if your gc-improv work
> would be entirely [or sufficiently] merged into the trunk. I am really
> scared of merging the trunk into the MELT branch before your gc-improv
> is done
Why? In general, the more frequent the merge, the easier it is.
Thanks,
--
Laurynas