On 03/05/2010 01:53 PM, Tom Tromey wrote:
> Keith>  One other issue that I uncovered: DW_AT_MIPS_linkage_name appears to
> Keith>  be necessary for Ada. I have a patch that I used to address this (for
> Keith>  one of our internal releases), but it is probably not complete. [In
> Keith>  other words: it's now just as broken as it was before.] I can submit
> Keith>  this as a follow-up, if so desired.
>
> By this, I assume you mean that after your patch, Ada continues to work,
> but still uses DW_AT_MIPS_linkage_name; and that you have a follow-on
> patch that is incomplete but which attempts to remove the reliance on
> DW_AT_MIPS_linkage_name?

The patch simply keeps/uses MIPS_linkage_name for Ada. Almost nothing 
else is changed. I've attached it. With this patch, Ada tests now fail 
as much as they did before. I'm guessing there is a lot more to do, though.

It is incomplete in the fact that I did not spend much time with it, and 
this patch is quite hacky. If we want to incorporate something like 
this, I'll need to play with it a bit first. Probably want to try to 
move this from dwarf2_name to dwarf2_physname, if that's even possible. 
[DW_AT_name and DW_AT_MIPS_linkage_name don't appear to be related in 
any way in the test cases I've seen.]

Keith