Re: [PATCH] disallow duplicate group names

Ali Gholami Rudi <[email protected]> writes:

> Hi Shawn,
>
> On Oct 17 2008 12:12 +0330, Shawn wrote:
>> Ali Gholami Rudi <[email protected]> writes:
>>
>>> +  if (groups_get_group_by_name (ARG_STRING (0)))
>>> +  if (groups_get_group_by_name (ARG_STRING (0)))
>>> +  if (groups_get_group_by_name (ARG_STRING (0)))
>>
>>>  rp_group *groups_find_group_by_name (char *s);
>>> +rp_group *groups_get_group_by_name (char *s);
>>
>> Doesn't groups_find_group_by_name work for these cases?
>
> No.  It finds the first group that starts with the given prefix (it
> uses str_comp()).  I've tried adding an "exact" parameter to
> groups_find_group_by_name(), but IMO it wasn't any cleaner.

Ahh, I see.

> Maybe group_find_group_by_name() can be changed to find the exact
> match? AFAICT, that usually means only an extra tab key: "gselect def"
> no longer works for switching to the "default" group (also gselect
> won't be case insensitive anymore).  Doing so will also solve the
> situation when there are two groups in which the name of one groups is
> the prefix of the other.  For instance when "a" and "ab" groups exist,
> it is impossible to switch to "a" group when "ab" has a lower group
> number (when not using group number in gselect, of course).

Give groups_find_group_by_name an argument: exact_match. If its
non-zero then then do an exact match. if its 0 then do what it
currently does. update the existing code, patch up your code, and
bipass the secondary manifold using a 0.04 micron phase discriminator.

-Shawn