Re: [debug-early] emitting early debug for external variables

[Aldy Hernandez <aldyh at redhat dot com>]

On 03/19/2015 02:08 AM, Richard Biener wrote:
> On Wed, Mar 18, 2015 at 10:28 PM, Jason Merrill <jason@redhat.com> wrote:
> > If you move the call to rest_of_decl_compilation we could go through and
> > prune the debug info for unused decls at dwarf2out_finish time, the way we
> > do with unused types.
>
> True.  Note that the varpool nodes eventually get created once the
> first use is seen.
> So I wonder how much garbage we create by unconditionally creating the node
> in rest_of_decl_compilation (yeah, header files, of course - probably
> a similar issue
> for unused function declarations?).
>
> I'd prefer to do early_global_decl from rest_of_decl_compilation (and shun the
> symtab/varpool walk - but that would require the FEs would hand off each global
> that possibly needs debug info through rest_of_decl_compilation).
>
> To prune globals during early(!) dwarf2out_finish you should be able to use the
> symbol table (not sure if we prune all unused symbols, but surely the list
> of references should be empty).
>
> Richard.

Thank you both.

I have moved the debug early generation for _symbols_ to 
rest_of_decl_compilation.  I'm not so so brave as to move 
FUNCTION_DECL's and such.  Besides, things are working fine.  Let me get 
through the rest of my gdb regressions. :).

I am now running gdb tests for every patch, in an effort to fix the 
plethora of regressions I have caused over the past few months.  The 
current patch exposes no regressions for guality.exp, or for the gdb 
testsuite.  For that matter, it fixes 4-5 gdb regressions.

I will prune the unused DIEs in a subsequent patch; actually much later, 
when I'm done regression hunting.

Let me know if you have any problems with this.  I am committing to the 
branch.

Aldy

Attachment: curr
Description: Text document