Re: [PATCH 2/7] gdb: Add an is_declaration field to each symbol

[Andrew Burgess <andrew dot burgess at embecosm dot com>]

* Tom Tromey <tom@tromey.com> [2019-07-29 14:21:45 -0600]:

> >>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:
> 
> Andrew> Track if a symbol is a declaration or not.  At one point we could
> Andrew> possibly figure this out based on the LOC_UNRESOLVED address class of
> Andrew> the symbol, but, for Fortran we mark some symbols as LOC_UNRESOLVED
> Andrew> even when the DWARF supplies an address, this is because the address
> Andrew> supplied by the DWARF is actually wrong.  For details look in
> Andrew> dwarf2read.c and look for references to gFortran bug #40040.  I have
> Andrew> confirmed that current versions of gFortran still have the issue
> Andrew> mentioned in that bug report.
> 
> That bug looks like another instance of the copy relocation problem.
> So, I wonder if these workarounds could be removed if/when this series
> lands:
> 
>     https://sourceware.org/ml/gdb-patches/2019-06/msg00612.html
> 
> (I still haven't looked at Pedro's review in detail, but I will get to
> it sooner or later.)

I checked your copy relocation series does indeed make parts #2 and #3
of my series redundant.

I'll address the other feedback with this series and rebase once your
series lands.

Thanks,
Andrew