Re: [PATCH] Fix line table regression

[Tom Tromey <tom@tromey.com>]

>>>>> "Simon" == Simon Marchi <simark@simark.ca> writes:

Simon> It does fix the issue in do_mixed_source_and_assembly I have seen.

I'm going to check it in.

Simon> I like the idea of having a separate type to catch things at compile
Simon> time.  I would like for it to be used in other objects using unrelocated
Simon> addresses.

I'll take a quick look.

Simon> I'm just afraid that some code paths will always need to be
Simon> handle both relocated and unrelocated addresses, so the distinction
Simon> may not always be completely clear cut.  For instance, your patch shows
Simon> that gdbarch_addr_bits_remove works with both relocated and relocated
Simon> addresses.

Yeah.  The contract on some of these things is a bit vague for my taste,
as well.

Simon> If this becomes more widespread, I think it would be nice to have some
Simon> helper methods like this:

Simon>   unrelocated_addr objfile::unrelocated (CORE_ADDR);
Simon>   CORE_ADDR objfile::relocate (unrelocated_addr);

Simon> It would make the code a bit clearer about what the intention is, versus
Simon> just doing additions and subtractions.

I think they'd have to take the section index as well, but yeah.

Tom