Re: Invalid postscript generated from simple Ink

[Fred Kiefer]

Am 19.07.2016 um 14:26 schrieb Riccardo Mottola <address@hidden>:
> Fred Kiefer wrote:
>> thank you for your analysis. It looks like line 271 in the file 
>> NSFontDescriptor.m needs an added test for face being null. But you are also 
>> correct that this case should not happen,  looks like we commented out 
>> setting the font face in the attributes dictionary in GSFontInfo.m line 796 
>> for some reason. I cannot remember when and why and don't have the time to 
>> inspect this until after my holidays. Adding the first fix should help you 
>> for now.
> 
> thanks. I will test and add the small fix. It is a safety feature, but for we 
> will loose the face information.

We wont "loose" it just not try to use it when it isn't present.

> Checking the file history, I think you never "removed" support for the face 
> attribute, but checked it in that way many years ago. Look at this commit:
> 
> https://github.com/gnustep/gui/commit/e1bb31f0d2851291efbdde2c2e596baca94120ab

Thank you for looking this up. I remember that change now, it was something I 
did to get Emacs to compile with GNUstep again. We needed basic 
NSFontDescriptor handling for that. I implemented that by taking some code from 
mystep and adding the rest myself.
But I still have no recollection on why I commented the face attribute out. 
Maybe I need to enable that bit and try the results.

> My only guess is that Face and Trait somehow overlap in concept and you 
> implemented Traits ? e.g. if I have a "Face" which is "Oblique"... is 
> automatically the Trait Italic ? or can I have "both" ?

No, they are in some regard similar but not identical, we have traits with no 
face name and the other way around.

Fred