This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFA] Remove make_cleanup_free_section_addr_info


On 2018-03-15 10:44 PM, Tom Tromey wrote:
>>>>>> "Simon" == Simon Marchi <simon.marchi@polymtl.ca> writes:
> 
> Simon> Did you consider changing the section_addr_info struct to use entirely
> Simon> by an std::vector<other_sections>?  Functions that return a
> Simon> section_addr_info could return a vector directly (by value? is that
> Simon> how we say it?), so we shouldn't need any custom unique pointer.
> Simon> Then, other_sections::name can be made into a string, removing the
> Simon> need to free it by hand.
> 
> I didn't really consider it, but thanks for the idea.  It turned out to
> be more invasive, but it did remove one weirdness from
> syms_from_objfile_1.  Previously that function did:
> 
>       local_addr = alloc_section_addr_info (1);
> 
> ... however, it doesn't actually need 1 entry -- the way this is set up,
> the minimum that can be requested is 1, but the returned
> section_addr_info actually has num_sections set to 0.  Obscure!
> 
> In the new code this is replaced by an empty vector.
> 
> Let me know what you think.  I did re-run this through the buildbot.
> 
> One idea I had is that maybe other_sections should have a constructor.
> That would simplify some of the code.

Hi Tom,

I took a quick look, it LGTM, but indeed having a constructor in other_sections
would make it nicer IMO.

Simon


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]