Re: [Devel] Bug in FT New Memory Face?

[Michael Pfeiffer]

Hello David,

thanks for the clarification.

Kind Regards,
Michael

>> I develop under BeOS and AFAIK it does not support memory-mapped I/O.
>>
>Ah ! So there is a difference..
> 
>> I think it was my fault to free the memory for the font file after
>> the call of FT_New_Memory_Face(). I thought this was safe because
>> the documentation for FreeType 2 does not mention that
>> FT_New_Memory_Face possesses the memory for the font file.
>> Please correct me if I am wrong.
>>
>Indeed, the font data that is fed to FT_New_Memory_Face can be located
>in ROM, or coming from a memory-mapped file, so there is no way
>to make FT2 "own" this piece of memory.
>
>This means you need to make the memory block "live" until the
>corresponding call to FT_Done_Face !