This is the mail archive of the
mailing list for the GCC project.
Re: [libiberty] add obstack macros (was Re: PR #53525 - track-macro-expansion performance regression)
- From: Dimitrios Apostolou <jimis at gmx dot net>
- To: Ian Lance Taylor <iant at google dot com>
- Cc: Dodji Seketeli <dodji at redhat dot com>, gcc-patches at gcc dot gnu dot org, Andrey Belevantsev <abel at ispras dot ru>, DJ Delorie <dj at redhat dot com>, Tom Tromey <tromey at redhat dot com>, Jason Merrill <jason dot merrill at gmail dot com>
- Date: Sat, 4 Aug 2012 19:40:53 +0300 (EEST)
- Subject: Re: [libiberty] add obstack macros (was Re: PR #53525 - track-macro-expansion performance regression)
- References: <alpine.LNX.firstname.lastname@example.org> <email@example.com> <alpine.LNX.firstname.lastname@example.org> <CAKOQZ8x05E9RsoCPGCj4AZ+yypA7=ZAwSf=hs=XsSry=XPVDvg@mail.gmail.com>
On Fri, 3 Aug 2012, Ian Lance Taylor wrote:
2012-08-04 Dimitrios Apostolou <email@example.com>
type-safe macros for obstack allocation.
(XOBFINISH): Renamed argument to PT since it is a pointer to T.
+/* Type-safe obstack allocator. You must first initialize the obstack with
+ obstack_init() or _obstack_begin().
This should recommend obstack_init, or obstack_begin, but not
_obstack_begin. Also obstack_specify_allocation and
obstack_specify_allocation_with_arg are OK, so really it might be
better not to list the functions, but simply say "You must first
initialization the obstack."
Grep reveals that 9 out of 10 times we use _obstack_begin(), and we set
alignment to 0 (isn't it suboptimal?).
+ T: Type, O: Obstack, N: Number of elements, S: raw Size,
+#define XOBSHRINK(O, T) obstack_blank ((O), -1 * sizeof (T))
+#define XOBSHRINKVEC(O, T, N) obstack_blank ((O), -1 * sizeof (T) * (N))
These are hard to use safely. I'm not sure we should define them at all.
I've already used XOBSHRINK and it looks clear to me, but I could use
obstack_blank() directly if necessary.
+#define XOBFINISH(O, PT) ((PT) obstack_finish ((O)))
For XOBNEW, etc., we use (T *) rather than (PT). Using (PT) seems
error-probe--it's the only use of the obstack with a different type
parameter. Why not use T rather than PT here, and return (T *)?
I'd have to change many (about 60) occurences of XOBFINISH if I change
that. I'd go for it if I was sure it's what we want, it can be a separate
patch later on.