[Bug symtab/30656] [gdb/symtab] optimized out static handled inconsistently

["vries at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=30656

--- Comment #7 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #2)
> (In reply to Tom de Vries from comment #0)
> > I wrote a dwarf assembly test-case simulating an optimized-out static
> > variable:
> 
> Also reproduced with a plain C test-case:
> ...
> $ cat whole-program.c
> static int aaa;
> 
> int
> main (void)
> {
>   return 0;
> }
> $ gcc-12 whole-program.c -g -flto -O2
> $ nm a.out | grep aaa
> $
> ...
> 
> Corresponding DIE:
> ...
>  <1><104>: Abbrev Number: 2 (DW_TAG_variable)
>     <105>   DW_AT_name        : aaa
>     <109>   DW_AT_decl_file   : 1
>     <10a>   DW_AT_decl_line   : 1
>     <10b>   DW_AT_decl_column : 12
>     <10c>   DW_AT_type        : <0x110>
> ...
> 
> Difference in behaviour:
> ....
> $ gdb -q -batch ./a.out -ex "print aaa"
> No symbol "aaa" in current context.
> $ gdb -q -batch -readnow ./a.out -ex "print aaa"
> $1 = <optimized out>
> ...

FTR, while writing the test-case I realized I couldn't reproduce this, I
probably got the system gdb and devel gdb mixed up.  Anyway, this does
reproduce with system gdb based on 12.1.  This points to the fact that we've
regressed in terms of not requiring to expand the symtab in order to find the
language of main, I suppose because we're not tracking the language of entries
in the cooked index.

For the test-case I split up the source in two CUs, to make sure that

-- 
You are receiving this mail because:
You are on the CC list for the bug.