2009/2/23 Dominik Vogt <[email protected]>:
> On Mon, Feb 23, 2009 at 09:21:03PM +0000, Thomas Adam wrote:
>> 2009/2/23 Dominik Vogt <[email protected]>:
>> - I can't say I like the magic numbers in the malloc line:
>
> :-P
>
> I couldn't think of anything better, and I think the previous
> allocation was too small.

Perhaps -- I did toy with looking at the actual size of the
charmap_t[] array at each instance it was being called (once as per
key_bindings, the other as win_contexts) -- but no worries, expect a
potential comment on this nevertheless, since if this ever comes back
to haunt myself and you, I don't want to be left scratching my head as
to why it's * 8 +1.  :)

>> allmods = safemalloc(sizeof(table->value) * 8 + 1);
>>
>> However, I know why they're there -- I might add a comment there if I
>> really feel strongly about it.  :)
>>
>> This approach is better in that it reduces memory allocation a little.
>>
>> I don't have a problem with it, and it seems to work well.
>
>> Your changes to  fvwm/bindings.c:print_bindings() aren't as extensive.
>>  I like (as trivial as it sounds) the block around the code after the
>> case statement -- that does make sense.
>
> I like to use that to limit the lifetime of stack variables.  And
> it's often a good way to document blocks of code.

Oh absolutely -- it's a technique I use often and embarrasingly
neglected to use here.  Thanks though for catching it.

>> The only query I have in print_bindings() is I am well aware I have
>> done nothing with "b->Action2", and whilst it is set in places, I am a
>> bit unsure when and if it is actually ever used -- at least, it hasn't
>> had any noticeable effects when dumping out bindings.  Perhaps though
>> it's worth considering -- and I'd be happy to do so if I understood
>> its use.
>
> It's only used in FvwmIconMan.

Thanks again -- that was my understanding as well, but it is always
good to be sure.

-- Thomas Adam

Reply via email to