Re: [PATCH 1/5] Introduce build_debug_file_name

[Pedro Alves <palves@redhat.com>]

On 07/02/2015 12:18 PM, Gary Benson wrote:
> Pedro Alves wrote:
>> On 06/16/2015 10:42 AM, Gary Benson wrote:
>>> This commit introduces a new function build_debug_file_name which
>>> concatenates a series of filename components into a filename.
>>> find_separate_debug_file is updated to use build_debug_file_name.
>>> A later commit in this series will extend build_debug_file_name to
>>> correctly handle "target:" prefixes, so it is convenient to have
>>> filename building pulled out into one function.  For now the only
>>> functional change here is that the original code sometimes
>>> generated filenames with repeated directory separators while the
>>> new code does not.
>>
>> I'd drop the "debug" from the function's name.  Sounds like a
>> candidate for reuse elsewhere to me.
> 
> Should I put it somewhere else, maybe common-utils.c?

I dislike common-utils.c for "kitchen-sink" reasons, though.

There's common/filestuff.c, but that looks more for low/OS level
file things.

There's substitute_path_component (and gdb_realpath / gdb_abspath)
in utils.c.  Maybe put it next to substitute_path_component, which
seems to be in the same "family" of function, and then (at some point)
we would move all file name/path manipulation routines to its own file.

WDYT?

Thanks,
Pedro Alves