This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: RFC: CLI files naming
- To: Fernando Nasser <fnasser at cygnus dot com>
- Subject: Re: RFC: CLI files naming
- From: Andrew Cagney <ac131313 at cygnus dot com>
- Date: Wed, 15 Nov 2000 01:16:11 +1100
- Cc: gdb at sources dot redhat dot com
- References: <3A101FD5.C38FD29D@cygnus.com>
Fernando Nasser wrote:
>
> As part of the separation of the command line interface code from the
> gdb core (libgdb), I will need to create files related to this interpreter.
>
> We have placed MI related files in the gdb/mi subdirectory,
> the TUI related files in the gdb/tui and the Insight files in the
> gdb/gdbtk one. Of course, CLI files should go into the gdb/cli subdir.
The TUI isn't exactly a source of good examples :-)
> However, for historical reasons, the MI and Insight files got prefixes
> in their names. The reason was that they previously resided in the gdb
> directory itself and this was a way to group them together.
>
> So, do I follow the tradition and name files like:
>
> gdb/cli/cli-cmds.c
The tradition also dates back to the days when the debug info didn't
contain the path to the file and, consequently, you needed to have
globally unique file names.
> or do I eliminate the redundancy and go for shorter names like:
>
> gdb/cli/cmds.c
>
> Note that the long names have an advantage when debugging. The MI,
> for instance, have a file called mi-main.c. If it was just main.c
> (of course we could choose another name) it would clash with the
> gdb/main.c when we tried to set a breakpoint at main.c:36 for instance.
I've this memory of regretting the choice mi-main.c thinking it should
have been mi-top.c
> Any preferences?
No. I went and tossed a coin and it came up tails - heads was short
form - so the longer form consistent with mi/mi*.
Andrew