Re: [PATCH] Add dw2-aranges.exp

[Kevin Buettner <kevinb@redhat.com>]

Hi Sasha,

The actual patches in this commit LGTM (once Keith's concern is
addressed), but I'd like to see some changes to the commit log...

> This test tests GDB is able to load a file with the
> .debug_aranges section Address size set to 0.

Maybe say:

    This test checks that GDB is able to load DWARF information when
    .debug_aranges has a section address size that is set to 0.

> The test was originally written by Jan Kratochvil to test
> commit 927aa2e778dce440f4de5de8fc37ead1683a804e
> Author: Jan Kratochvil <jan.kratochvil@redhat.com>
> Date:   Fri Dec 8 22:44:12 2017 +0000
> 
>       DWARF-5: .debug_names index consumer

I think that this can be written more compactly, perhaps like this:

    This test was originally written by Jan Kratochvil to test commit
    927aa2e778d from 2017, titled "DWARF-5:  .debug_names index
    consumer".

> The test was passing from the beginning.
> This test has been living in the Fedora tree
> for a long time, it seems to be a reasonable
> some bad debug information.

Maybe change this to:

    This test was originally written using a static .S file and has
    been present in the Fedora tree for a long time.

> If dwarf2/aranges.c is modified to turn off the
> address_size check, GDB will crash with SIGPE

s/SIGPE/SIGFPE/

(I think.)

> when loading the executable with the address size

s/the address/address/

> set to zero.
> 
> I modified the DWARF assembler to make it possible
> to set the Addres size to zero in a .debug_aranges section

s/Addres/address/

> and used a DWARF assembler to produce the assembly file.

s/a DWARF/the DWARF/

With those changes (assuming you're okay with them) and with
Keith's change to use "require dwarf2_support",...

Approved-by: Kevin Buettner <kevinb@redhat.com>