Thanks for the feedback.
Kleber
Em sexta-feira, 21 de agosto de 2020 01:22:36 BRT, Werner LEMBERG
escreveu:
>> Today the code is like this:
>>
>> error = FT_New_Size( t42face->ttf_face, );
>> t42size->ttsize = ttsize;
>> FT_Activate_Size( ttsize );
>>
>> My initial
>> Today the code is like this:
>>
>> error = FT_New_Size( t42face->ttf_face, );
>> t42size->ttsize = ttsize;
>> FT_Activate_Size( ttsize );
>>
>> My initial suggestion was:
>>
>> error = FT_New_Size( t42face->ttf_face, );
>> if(error)
>> return error;
>>
Hi Alexei,
I'm not expert in this system, but I think that it is not a good pratice. Sorry
if i caused any misunderstanding.
Kleber
Em quinta-feira, 20 de agosto de 2020 16:43:47 BRT, Alexei Podtelezhnikov
escreveu:
Kleber,
Why would you prefer uninitialized t42size->ttsize to NULL?
Kleber,
Why would you prefer uninitialized t42size->ttsize to NULL? Please
answer this question.
Alexei
On Thu, Aug 20, 2020 at 3:20 PM Kleber Tarcísio
wrote:
>
> Hi everyone,
>
> I appreciate your attention!
>
> Today the code is like this:
>
> error = FT_New_Size( t42face->ttf_face,
Hi everyone,
I appreciate your attention!
Today the code is like this: error = FT_New_Size( t42face->ttf_face,
); t42size->ttsize = ttsize; FT_Activate_Size( ttsize );
My initial suggestion was:
error = FT_New_Size( t42face->ttf_face, ); if(error)
return error;
> In the first case, I think we need return because ttsize can be NULL and
> t42size->ttsize = ttsize; could get a NULL value.Is it a problem, no?
> In the first case, I think we need return because ttslot can be NULL and
> slot-> ttslot = ttslot; could get a NULL value.Is it a problem, no?
Hi Kleber,
> https://github.com/klebertarcisio/freetype2-1/commit/aab3c841f08c3a67914fbd0167ef4ff174116ccf
What is wrong with a couple of housekeeping lines before returning the
error? Why do you think that it is necessary to drop everything and
run for cover? In fact, libraries are not supposed
Hi Free Type Team, how are you?
My name is Kleber and I'm a PhD student. I'm working on a technique to help
developers to find weaknesses in configurable systems. I would like to know
what you think about this change: