Re: [ft-devel] [PATCH RFC] src/sfnt/ttsbit0.c: fix broken pointer overflow checks

2013-01-25 Thread Sean McBride
On Thu, 24 Jan 2013 15:58:14 -0500, Xi Wang said:

Many compilers such as gcc and clang optimize away pointer overflow
checks `p + n  p', because pointer overflow is undefined behavior.

And clang can detect signed overflow at runtime if you enable the debug flag 
-fsanitize=undefined described here:

http://clang.llvm.org/docs/UsersManual.html#controlling-code-generation

Could be useful to build freetype with that flag and see what it finds...

Cheers,

-- 

Sean McBride, B. Eng s...@rogue-research.com
Rogue Researchwww.rogue-research.com 
Mac Software Developer  Montréal, Québec, Canada



___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel


Re: [ft-devel] [PATCH RFC] src/sfnt/ttsbit0.c: fix broken pointer overflow checks

2013-01-25 Thread Sean McBride
On Fri, 25 Jan 2013 17:48:46 -0500, Xi Wang said:

 And clang can detect signed overflow at runtime if you enable the
debug flag -fsanitize=undefined described here:

Yeah, if you have input data to trigger the overflows.

Yeah.  Ideally freetype would have some unit tests, and some fuzzing.  But oh 
well.

Cheers,

-- 

Sean McBride, B. Eng s...@rogue-research.com
Rogue Researchwww.rogue-research.com 
Mac Software Developer  Montréal, Québec, Canada



___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel


[ft-devel] Subpixel Hinting is Broken

2013-01-25 Thread Infinality

Hi all,

I've spent the last few hours trying to figure out what was wrong with 
my code (because I naturally assume that something is messed up with 
*it*. :D  )  Well, it turns out that some recent change to Freetype has 
broken subpixel hinting.


Compile the last release with subpixel hinting and rendering enabled, 
and run ftview on Courier New (Windows 8 version, but others may work):

ftview -r 96 10.0 /usr/share/fonts/windows8/cour.ttf

Hit 'L' twice and the down arrow twice, and Esc.  Fails = 0.

Then, compile the latest git with subpixel hinting and rendering 
enabled, and repeat.  I get 713 fails.  I tried reverting certain more 
obvious things with the changes since the last release, but wasn't able 
to narrow down the cause.


Any ideas on what might have caused this?

Thanks,
Erik


___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel


Re: [ft-devel] Subpixel Hinting is Broken

2013-01-25 Thread Infinality

Alexei's latest commit seems to have fixed the issue.


On 01/25/2013 10:36 PM, Infinality wrote:

Hi all,

I've spent the last few hours trying to figure out what was wrong with 
my code (because I naturally assume that something is messed up with 
*it*. :D  )  Well, it turns out that some recent change to Freetype 
has broken subpixel hinting.


Compile the last release with subpixel hinting and rendering enabled, 
and run ftview on Courier New (Windows 8 version, but others may work):

ftview -r 96 10.0 /usr/share/fonts/windows8/cour.ttf

Hit 'L' twice and the down arrow twice, and Esc.  Fails = 0.

Then, compile the latest git with subpixel hinting and rendering 
enabled, and repeat.  I get 713 fails.  I tried reverting certain more 
obvious things with the changes since the last release, but wasn't 
able to narrow down the cause.


Any ideas on what might have caused this?

Thanks,
Erik




___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel


Re: [ft-devel] Subpixel Hinting is Broken

2013-01-25 Thread Werner LEMBERG

 Alexei's latest commit seems to have fixed the issue.

Are you aware of the bisecting method?  Have a look at `git help
bisect'.  For such situations it is the easiest way to find out what
commit exactly is the problem.


Werner

___
Freetype-devel mailing list
Freetype-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/freetype-devel