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 ping [dbxout_typedefs]


Per Bothner <per@bothner.com> writes:

> Zack Weinberg wrote:
>
>> You never responded to the concern I raised about the typedefs still
>> being in reverse order in languages other than C.
>
> I did some simple experiments, comparing before and after
> using cc1plus and jc1.
...

> Before patch:  C in reverse order, C++ and Java right order.
> After patch:  C correct order, C++ and Java in reverse order.
>
> Gdb doesn't seem to care (though older debuggers did).
>
> I think this is a weak argument in favor of the patch.
> We don't support stabs debugging for C++ or Java very well,
> so it is more important to get stabs right for C.
>
> A stronger argument in favor of the patch:  It removes
> a case of potentially unbounded deep recursion, which
> could crash the compiler, at least on small-stack hosts.

These are both good arguments, and gdb not caring is an important data
point (did you try the gdb test suite, or just do ad-hoc experiments?)
I am going to say that you may check in this patch in 24 hours, unless
someone registers a substantive objection in that time.  However, you
are then on the hook to produce patches to get the order right for all
other supported languages before 3.4 ships.

zw


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