Re: [PATCH] add-symbol-file-from-memory command

[Roland McGrath <roland at redhat dot com>]

>   . The first line of the doc string, up until the first \n, should
>     be a short description of the command, since that is what GDB
>     displays when several commands are listed (e.g., by `apropos' or
>     `help files' commands).  [Yes, I know: `add-symbol-file' that is
>     already there has the same problem.]

Ok.  I used add-symbol-file as the model.  How about this text?

"Load the symbols out of memory from a dynamically loaded object file.\n\
Specify an expression for the address of the file's shared object file header."

>   . It is not entirely clear what could ADDR be.  Is that a numerical
>     address, or something more flexible?  The reason that this is
>     important is that the command's completion function should be set
>     according to the possible arguments it could accept; as written,
>     the command's completion will try to complete on symbol names,
>     which I'm not sure to be appropriate.

It uses `parse_and_eval_address', again modelled on the add-symbol-file
command's implementation.  I think that makes generic completion, including
symbol names or whatever, the appropriate thing.  

> Also, if this code is approved, please add to the manual a
> description of the new command.

Ok.