Bug 85256 – freetype-2.1.9: bytecode interpreter not enabled (bindist in USE)

• Product: Gentoo Linux
• Status: RESOLVED INVALID

[Description Konstantinos Maras 2005-03-14 13:43:01 UTC]

I upgraded to freetype-2.1.9-r1 and i got a ewarn message:
 * Bytecode Interpreter is disabled.
I checked in X: it was indeed disabled (fonts a bit uggly...)
It seems the cflag trick doesnt work anymore

I tried the upstream recommended method:
a #define in freetype-2.1.9/include/freetype/config/ftoption.h)
and it works like a charm!
I will attach the patch and a new ebuild

thanks for all the fine work - sorry for my bad english

Reproducible: Always
Steps to Reproduce:
1.put bindist in USE somehow...
2.emerge sync
3.emerge freetype

Actual Results:  
bytecode interpreter in freetype library was disabled

Expected Results:  
bytecode interpreter in freetype library must have been enabled

[Comment 1 Konstantinos Maras 2005-03-14 13:46:16 UTC]

Created attachment 53455 [details, diff]
upstream recommended method for enabling the bytecode interpreter (the patch)

[Comment 2 Konstantinos Maras 2005-03-14 13:47:47 UTC]

Created attachment 53456 [details]
upstream recommended method for enabling the bytecode interpreter (the ebuild)

it works here

[Comment 3 foser (RETIRED) 2005-03-14 18:08:43 UTC]

It should be exactly the same thing (patch vs our current method), the define overrides the flag. It always worked, i don't see why that should be different now.

The test doesn't necessarily have to be right, don't take that as clue. It's old and it's checking for API that might have changed.

PS. your header patch is not very clean, it leaves half of the 'comment' syntax in.

[Comment 4 Konstantinos Maras 2005-03-15 14:00:32 UTC]

I checked it again (with a 'cleaned' patch, your remark was right):
 the cflag-thing definately doesnt work here,
 and the little grep test is still working ok

i dont have any peculiar configuration (updated libc, gcc, x11),
if you like i can post my "emerge --info"
or even screenshots showing the difference in font rendering... :)

thanks again

[Comment 5 foser (RETIRED) 2005-03-15 14:16:39 UTC]

it's not that i don't believe you, but I'm pretty sure it works for me and everyone I know out there. With USE=bindist it is supposed to use the autohinter, so it's nothing more than natural that you see a difference, it would be a problem if you didn't. Now which one you prefer is a matter of personal preference and also a bit of 'getting used to'.

[Comment 6 Konstantinos Maras 2005-03-16 09:15:20 UTC]

>> With USE=bindist it is supposed to use the autohinter, so it's ...

...the enlightment!
after a little bash reading and internet searching
i have to say that u r right and i am wrong
(i thought that it was the opposite thing:
 use=bindist --> use the pattented code)
lets close this bug
and open another one saying
"bindist USE flag: totally meaningless and confusing name and description"

sorry for consuming your time foser

[Comment 7 Konstantinos Maras 2005-03-16 09:24:09 UTC]

my mistake

i thought that it was the opposite thing:
 use=bindist --> use the pattented code

[Comment 8 foser (RETIRED) 2005-03-16 09:52:40 UTC]

regarding your other comment. It is bindist because that should be on for binary redistribution (livecd's etc.), so we don't use possibly patent encumbered code on them. No need to open another bug for that, we won't change it.