Re: AIX bootstrap failure (was Re: Hot/cold partitioning fixes)

[Caroline Tice <ctice at apple dot com>]

Before I go and attach these labels to the function structure, I
just wanted to double check on something.  The last time
I tried adding fields to a structure, there turned out to
be a problem because people did not like the size of
the structure to be increased (admittedly there are *far*
more basic_block structs and edge structs that function
structs).  So I just want to make sure this is not going to
be a problem, before I add these labels to the function
structure.

Any objections to my adding fields to the function structure?
Anyone?

-- Caroline Tice
ctice@apple.com

On Apr 1, 2005, at 2:30 PM, Caroline Tice wrote:

> The variables are:
>
> unlikely_section_label
> hot_section_label
> cold_section_end_label
> hot_section_end_label
>
> They are used in assemble_start_function and assemble_end_function,
> where they are written out.  They are also used in dbxout.c and
> dwarf2out.c, where they are used to calculate the sizes of the text
> sections for the debugging information.
>
> -- Caroline
> ctice@apple.com
>
> On Apr 1, 2005, at 2:21 PM, Geoffrey Keating wrote:
>
> > Caroline Tice <ctice@apple.com> writes:
> >
> > > I was thinking that since I already have global variables that I was
> > > putting the old strings into, I would just use those global
> > > variables putting the result of calling ASM_GENERATE_INTERNAL_LABLE
> > > into them.  Is this a bad idea?
> >
> > Where are these global variables used?
> >
> > Yes, in general having random routines depend on the contents of
> > random global variables is a bad idea.