[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-trivial] [Qemu-devel] [PATCH] libcacard: remove useless initia
From: |
Michael Tokarev |
Subject: |
Re: [Qemu-trivial] [Qemu-devel] [PATCH] libcacard: remove useless initializers |
Date: |
Sat, 24 May 2014 00:59:51 +0400 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Icedove/24.5.0 |
So, should we apply this or not? It's been waiting for quite some time,
and during this time we've found a very good example of why it should
be applied (I think anyway).
Thanks,
/mjt
12.05.2014 13:20, Markus Armbruster wrote:
> Michael Tokarev <address@hidden> writes:
>
>> 11.05.2014 11:58, Alon Levy wrote:
>>> On 05/08/2014 08:19 PM, Michael Tokarev wrote:
>>>> libcacard has many functions which initializes local variables
>>>> at declaration time, which are always assigned some values later
>>>> (often right after declaration). Clean up these initializers.
>>>
>>> How is this an improvement? Doesn't the compiler ignore this anyhow?
>>
>> Just less code.
>>
>> To me, when I see something like
>>
>> Type *var = NULL;
>>
>> in a function, it somehow "translates" to a construct like
>>
>> Type *found = NULL;
>>
>> That is -- so this variable will be used either as an accumulator
>> or a search result, so that initial value is really important.
>>
>> So when I see the same variable receives its initial value in
>> the next line, I start wondering what's missed in the code which
>> should be there. Or why I don't read the code correctly. Or
>> something like this.
>>
>> So, basically, this is a cleanup patch just to avoid confusion,
>> it most likely not needed for current compiler who can figure
>> it out by its own. And for consistency - why not initialize
>> other variables too?
>
> I hate redundant initializers for yet another reason: when I change the
> code, and accidentally add a path bypassing the *real* initialization, I
> don't get a "may be used uninitialized" warning, I get the stupid
> redundant initialization and quite possibly a crash to debug some time
> later.
>
>> Maybe that's just my old-scool mind works this way.
>>
>> At any rate you can just ignore this patch.
>
> Please consider it.
>