arial narrow is broken since Fedora 8

Julian Sikorski belegdol at gmail.com
Wed Dec 10 13:21:27 UTC 2008


Nicolas Mailhot pisze:
> 
> Le Mer 10 décembre 2008 01:54, Kevin Kofler a écrit :
> 
>> Please do not ignore real-world usability in your quest for
>> perfection.
> 
> This is not a quest of perfection this is getting font and text bugs
> fixed.
> 
> The freetype autohinter has progressed because we've enabled it in
> Fedora despite its problems and told people to report bugs upstream
> instead of helping them enable the bytecode interpreter and ignore the
> problem.
> 
> OO.o has started working on OpenType CFF support because we told them
> plainly we would not stop merging fonts in this format or prioritize
> OpenType TTF just so people didn't notice that unlike other apps, OO.o
> didn't work.
> 
> Ligature support was fixed in Firefox because we didn't try to hide
> them and so people complained upstream of upstream bugs.
> 
> Likewise ligature support was lately fixed in freetype, again because
> we didn't hide the problem and people complained in the right place
> (upstream issue trackers, not general-purpose downstream lists).
> 
> When we tried to hide a problem by removing triggering glyphs
> font-side there was 0% progress on fixing application-side and some
> upstreams still argue we should restore the hiding so they don't have
> to bother.
> 
> Red Hat tried to avoid font problems by not merging anything that
> looked like it would trigger application bugs, and had to shell some
> millions later for Liberation; complaining at the same time the FLOSS
> font scene really was not active enough for them to rely on it. Well,
> if you want activity you have to support this activity not ignore it
> and hope things will magically perfect themselves without distro-side
> exposition.
> 
> Your proposition is made of 100% pure un-adultered FAIL. And facts
> back me up on this.
> 
> You'll find scores of people to rewrite spontaneously media players,
> MUAs, or the distro boot chrome, but people won't work on font
> problems unless users complain to them, and users won't complain if
> you hide or diffuse the problems. They'll just note Fedora font
> support is crap, without pointing to any specific fact.
> 
> Just Google for 'linux fonts', and you'll find many such reports,
> culminating around 2006, which incidently is when Fedora decided to
> get its feet wet, and released Fedora 6 with DejaVu LGC as default,
> breaking the status quo and starting the virtuous circle of upstream
> fixes.
> 
> True, even with users complaining, some bugs take ages to get fixed,
> but the way to accelerate this is to get more users to complain, not
> remove the complains. If some of your favorite DEs or apps are not
> fixed yet organise fellow users and put some pression upstream (or,
> better, find someone to submit upstream a patch).
> 
> I personally think our current course is the best to « [lead] the
> advancement of free, open software and content. »
> 
> Anyway, I'm sick of repeating the same arguments in different forums.
> Here's the deal: you disagree with our current font strategy, so go
> convince FESCO. If FESCO agrees with you, I'll happily give you the
> Fonts SIG keys, and let you manage as you wish from now on. I
> personnaly do not intend to waste my personal time trying to improve
> the Fedora font situation if one of the precious few levers I have at
> my disposition, users complaining upstream of problems, is removed.
> 
> And that's the last thing I'll write on the subject.
> 

Looks like the current strategy works then, but quite slowly in some
cases. The main issue I see here is that freedesktop bugzilla seems to
be a black hole - I added a comment to bug #13416 half a year ago and
got zero response from upstream developers, so please forgive me my
slight frustration at this point - it's not that I didn't try, but if no
one upstream gives a damn about the problem, I tried to find allies
here, who could possibly have more convincing power. Attracting
attention - that was the purpose of my post.

Julian




More information about the fedora-devel-list mailing list