This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [OB PATCH] Fix doc for -add-inferior
- From: Pedro Alves <palves at redhat dot com>
- To: Yao Qi <yao at codesourcery dot com>
- Cc: Simon Marchi <simon dot marchi at ericsson dot com>, GDB Patches <gdb-patches at sourceware dot org>
- Date: Tue, 04 Jun 2013 15:36:54 +0100
- Subject: Re: [OB PATCH] Fix doc for -add-inferior
- References: <I-love-christmas at ericsson dot com> <51AD401C dot 7090806 at codesourcery dot com> <51ADEC29 dot 7030907 at ericsson dot com> <51ADF29D dot 5010609 at codesourcery dot com>
On 06/04/2013 02:58 PM, Yao Qi wrote:
> On 06/04/2013 09:31 PM, Simon Marchi wrote:
>> I like "thread-group" better also. Since changing an existing MI
>> interface could break applications that use it, what is the policy in
>> GDB about changing things like this?
>
> In general, we can't change the existing interface.
Agreed.
> However, this change fixes the inconsistency between code and doc, and it should be fine to change the code to comply with the doc, IMO.
That's unfortunately not really a valid justification, so I can't agree with that.
Valid justifications would be "we don't need to care that frontends break
because ..." (e.g., "this never worked as is"), or
"frontends won't break because ...".
Any existing frontend that is using -add-inferior will be expecting to
see "inferior" there in order to extract the inferior id. It's quite natural
that other frontends authors hadn't even noticed the 'reality vs documentation'
divergence, as -add-inferior resulting in "inferior" borders on
so-obvious-who-needs-to-check-docs-anyway output. IMNSHO, it's better to
just document the over-3-year-old current behavior.
--
Pedro Alves