This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix PR52977


On Thu, 26 Apr 2012, Diego Novillo wrote:

> On 4/26/12 9:35 AM, Diego Novillo wrote:
> > On Thu, Apr 26, 2012 at 06:43, Dodji Seketeli<dodji@seketeli.org>  wrote:
> >
> > > I guess it's also worth noting one limitation of PPHs that is, if I
> > > believe the wiki:
> > >
> > >     In essence, the only headers that can be pre-parsed are those that
> > >     produce the same result when they are compiled in isolation or as
> > >     part of another translation unit. So, header files that are affected
> > >     by pre-processor symbols defined before inclusion are not going to
> > >     be considered (e.g., stddef.h).
> > >
> > > [1]: http://gcc.gnu.org/wiki/pph#Scope_and_limitations
> > >
> > > How hard would it be to drop that limitation?
> >
> > It's an explicit non-goal, actually.  If you relax this requirements,
> > you might as well re-parse the header file.  The work needed to make
> > flexible PPH images will rob you of most/all the performance you were
> > looking for.
> 
> Expanding a bit on this point.
> 
> One of the goals of PPH is to act as a bridge towards C++ modules, which is
> currently being discussed for inclusion in the next C++ standard.
> 
> While it is uncertain what modules will look like, we want to use this
> experience to inform the committee on the different tradeoffs and issues we
> found.  We will publishing something shortly, and discussing it at the Prague
> meeting.
> 
> I a future world with modules (if it comes to be), files acting as modules
> will need to have exactly one meaning, regardless of where they are imported
> from or what was imported before them (think Java/Go/Python modules).

And btw - the PCH implementation shares this limitation.  All tokens
consumed are part of the PCH - changing the state when trying to use
the PCH will invalidate it (or crash gcc).  In this way PCH is even
more limited than PPH because you can only have a single PCH file
and that single PCH file must represent the very first tokens of
the compilation unit it is used in.

Richard.

-- 
Richard Guenther <rguenther@suse.de>
SUSE / SUSE Labs
SUSE LINUX Products GmbH - Nuernberg - AG Nuernberg - HRB 16746
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]