search

bohoomil / fontconfig-ultimate

freetype2-infinality run-time settings => infinality compatible fontconfig => infinality-bundle
454 stars 38 forks source link

Slow font rendering in chrome #122

Open bjourne opened 8 years ago

bjourne commented 8 years ago

Thanks for maintaining Infinality! I was afraid it would fall by the wayside when the original author gave up maintainership.

That said, there seem to be some issue with Kubuntu 15.10 + Chromium + Infinality. I'm installing the packages from this ppa: https://launchpad.net/~no1wantdthisname/+archive/ubuntu/ppa With those packages font rendering is very slow in Chromium. There is noticable lag when switching between tabs or clicking links. Without the packages, Chromium is very fast.

bohoomil commented 8 years ago

Unfortunately, I'm not maintaining packages for Ubuntu so I can't help much here. Here we only report code related issues, not problems with distro-specific releases.

bjourne commented 8 years ago

Then who is doing the Ubuntu packages? Are you from now on supposed to patch and compile freetype and fontconfig yourself?

bjourne commented 8 years ago

And regardless, if infinality slows down font-rendering woudln't that be a code problem and not a packaging issue?

bohoomil commented 8 years ago
  1. The info reads: "For questions and bugs with software in this PPA please contact slow." Ergo: please contact slow.
  2. Maybe it's not infinality that slows down font rendering but some sort of misconfiguration of your environment? Did anyone else report a similar issue? Apart from it, Ubuntu releases are based on freetype2 v.2.5.2 and a pretty dated version of the infinality patch which is not supported any longer. The currently available patch in this repo is for freetype2 v.2.6.2. The difference between 2.5.2 and 2.6.2 code base is huge. Next, Ubuntu is using a different fontconfig-infinality configuration than the one available here. Are you sure your fontconfig setup is entirely correct? What do your logs say? Did you try tracking down the problem? There are dozens of reasons for an application to misbehave and unless you prove the problem is about code, it's just a statement that doesn't mean much...
bjourne commented 8 years ago

Wow, no. Your attitude is way to disparaging for me to bother with Infinality bugs.

behdad commented 8 years ago

Wow, no. Your attitude is way to disparaging for me to bother with Infinality bugs.

As a bystander, I want to chip in. Björne: I don't know what you didn't like about bohoomil's "attitude" but his comment and "attitude" was nothing short of polite and professional.

Maybe you never developed computer code, or maybe you are new to Free Software, but as bohoomil explained, just saying "slows down" is not going to enable upstream developers to be able to do anything for you. There's a thousand possible ways things might be going wrong on your system, and neither bohoomil nor anyone else can know without your cooperation. And reporting to packager is first step to start that process, unless you have debugged the problem already and know it's an upstream issue.

As a former and current maintainer of much of the text rendering pieces on the Free Software desktop, Ican tell you that it's comments like yours that we developers / maintainers find highly demotivating. But we learn to ignore those and keep it up, despite people demanding us to solve their problems even though we have no moral or contractual obligation to, and then sometimes they are plain rude and even abusive... Anyway, I digress.

bjourne commented 8 years ago

@behdad, it shouldn't matter who I am. Treat everyone with respect. If your position is anything other than that people who file bug reports do it to help the project then you are wrong.

But I'm not new, I've used Infiniality for 3-4 years, blogged about it, helped others install it and offered to help push it upstream to freetype & fontconfig where it belongs. I have my own projects which I handle incoming bug reports for.

Let me explain why @bohoomils comment is disparaging:

"please contact slow": Since the problem is not about packaging contacting him first is a waste of time. If he is currently involved in Infinality development he likely subscribes to this issue tracker already and is therefore implicitly contacted. His ppa page doesn't even have a bug tracker.

"Maybe it's not infinality that slows down font rendering but some sort of misconfiguration of your environment?" Yes, maybe but how should I know? I did my best to eliminate all other variables other than the infinality-related packages.

"Did anyone else report a similar issue?" Why even ask me this? How should I know?

"Are you sure your fontconfig setup is entirely correct?" How can anyone be 100% sure of that except for you, @behdad? I'm 95% sure it is correct.

"What do your logs say?" Which logs should I look for strange messages in?

"Did you try tracking down the problem?" Of course! Why would I otherwise go through the trouble reporting it here? It took me several hours to nail down that, with the fontconfig-infinality package, Chromiums font rendering is borked, without it, it is not.

Those questions are all rhetorical (asked to show why my bug report is invalid) and doesn't help us find the root cause. But I'm still here! If @bohoomil tells me how I could help him track down a likely bug in Infinality. But if he isn't interested, of course I won't bother.

behdad commented 8 years ago

I have no further comments. Looks like you know how to get people to want to help you.

goddesse commented 8 years ago

@bjourne This is a semi-packaging issue. Bohoomil doesn't support old code that is definitely slow and has other defects and the most recent update has a patch that should greatly improve rendering speed. You could ask slow to backport the most recent changes. slow would need to revert 5d0ca06 and 8b4d4e8 for compatibility with 2.5.2 and also check the default LCD filter parameters and make sure it's what infinality set as I believe those recently changed with 2.6.2.