You are not logged in.
The text seems to have some "color fringes".
Try to adjust native Gnome font settings using gnome-tweak-tool. You may try and duplicate in Gnome configuration found in /etc/fonts/conf.avail.infinality/10-base-rendering.conf.
I just couldn't understand why Mr Steffens would say that "it's making Cantarell tolerable".
Because people perceive things differently? Because there is not a single, 'correct' perception?
You keep expecting things to be what you want instantly, without even trying to examine how they really work. You have Cantarell available in three formats: TTF, OTF and T1. If you had spent a little time learning and experimenting, you would have gained enough knowledge to fine tune both Cantarell and the rest of your collection. A lumberjack's approach won't work here, unfortunately. However, that seems like your choice.
@brebs: It's not the first time when a T1 font looks best. I somehow missed the presence of Cantarell in T1 format: was it already available in TeX Live 2012? Anyway, thanks for the hint.
:: Registered Linux User No. 223384
:: github
:: infinality-bundle+fonts: good looking fonts made easy
Offline
I just edited my earlier post, adding screenshots, assuming that no-one would post again so quickly
Perhaps cairo needs the "respect" patch, to apply all the custom fontconfig rules, in Gnome 3? I wouldn't be surprised
Offline
I tried the infinality-bundle. It fixes the "number issue" and it's much sharper. Still I prefer what text looks in 2.4 generic.
I don't know how to describe the problem with the bundle. It's kind of like I am using rbga AA. The text seems to have some "color fringes".I just couldn't understand why Mr Steffens would say that "it's making Cantarell tolerable".
Even if Cantarell is not a nice font, making it blur doesn't help at all.The point is, the gnome dev has once done something to make the font sharp.
(And Arch adopted it! Arch use the otf instead of the ttf!)
And now the new engine simply and only cancel the effect of that "real tweak". NOTHING MORE.Guess this is enough for me. I don't know if this is a matter of taste but I certainly could not convice myself that it is.
I'll certainly file a bug report somewhere if I didn't see the google+ post.
I'll stick with 2.4 until it's breaking things down or there are some good change. R.I.P. my text
Infinality's rendering is extremely customizable, you can tweak fontconfig settings with the script in /etc/fonts/infinality, and rendering settings in the /etc/profile.d/infinality-settings.sh file. I'm 100% sure you could probably fix whatever issues you had with its rendering. Infinality does in fact use rgba and subpixel aa by default for example, but all that can be changed.
I've got my system configured to use rendering somewhat simiar to ubuntu's (but a bit different) just as an example, there's many other presets and settings that can be tweaked to achieve totally different rendering: http://i.imgur.com/CzIImKh.png
Last edited by bwat47 (2013-07-06 19:25:25)
Offline
@bohoomil
I know there are different tastes. I just don't think the issue I saw was a matter of that.
I don't not only because of what I saw (or perceive), but as what I said, it was once considered a problem and fixed, and it just come back again.
That's why I don't think trying to use a different package or tune setting is the right solution or something I should do in this case. (Okay, true only considering I am a lazy guy)
(Because it looks like a regression caused by mistake.)
I don't think it's wrong or bad to stick with default/generic packages/settings. It's more sustainable for me.
Yet I am not criticizing your way. That's a matter of choice/preference/need.
Last edited by tom.ty89 (2013-07-06 19:57:09)
Offline
https://bugs.archlinux.org/task/35948
https://git.gnome.org/browse/cantarell-fonts/tree/NEWS
It makes the whole thing looks even more suspicious. Why Cantarell 0.0.12 works fine with the new engine while 0.0.13 doesn't?
Yet the bug report is closed. What a DELIBERATE CHANGE.
It turns out that all I want is autohint.
I forgot to look in my own config file, for where Cantarell has been tweaked. Here they are, the first tweak is by Infinality, presumably to fix a strange issue:
<!-- Fixes issues where Cantarell is not being displayed --> <match target="font"> <test name="family"> <string>Cantarell</string> </test> <edit name="family" mode="assign" binding="strong"> <string>Cantarell</string> </edit> </match> <match target="font"> <test name="family"><string>Cantarell</string></test> <edit mode="assign" name="autohint"><bool>true</bool></edit> <edit mode="assign" name="hintstyle"><const>hintslight</const></edit> </match> <!-- To stop 'e' from looking vertically compressed --> <match target="font"> <test name="family"><string>Cantarell</string></test> <test name="weight" compare="more_eq"><const>bold</const></test> <edit mode="assign" name="autohint"><bool>false</bool></edit> <edit mode="assign" name="hintstyle"><const>hintnone</const></edit> </match>
Try those in a new file, /etc/fonts/conf.d/91-cantarell.conf, to make sure they override your existing rules - still fuzzy?
The reason this don't work because the first autohint is true but the second is false. (I don't know why)
P.S. Unlike the bug poster, the 21-cantarell-hinting.conf from 0.0.12 works for me. It should definitely stay in 0.0.13 in my opinion.
And the funny thing is, Mr Steffens is actually the commiter of it. https://mailman.archlinux.org/pipermail … 82471.html
Last edited by tom.ty89 (2013-07-06 22:49:41)
Offline
:: Registered Linux User No. 223384
:: github
:: infinality-bundle+fonts: good looking fonts made easy
Offline
Because when I ask you again, you tell me to see your signature (which leads me to the infinality-bundle).
Okay it's my fault. I indeed kind of skip it when I saw "custom settings". But I tried. Just that I tried breb's maybe-wrong example.
But my instinct is correct. If 21-cantarell-hinting.conf hasn't been removed, everything will be fine.
I never needed something like infinality-bundle. Being lazy works better for me.
That's why I said I don't like tweak, because the truly necessary ones would usually be added in the Arch package already.
That's why I love Arch. KISS, yet sane KISS.
I just don't know why Mr Steffens remove his good tweak now. Anyway it's good to know that I can fix it myself.
Sorry and thanks guys!
Last edited by tom.ty89 (2013-07-06 23:32:09)
Offline
because the truly necessary ones would usually be added in the Arch package already.
Haha. Then there's the occasional madness, like cairo respect fontconfig.
Your fontconfig changes didn't take effect. That's VERY BAD, and it's been a soure of annoyance for years. Don't you think THAT'S a bug? So try the "respect fontconfig" patch. Because when I add fontconfig rules (in XFCE, with my "cairo respect fontconfig" patch), they DON'T get ignored.
@bohoomil, are you adding cairo-respect-fontconfig.patch to your cairo? I can't download your cairo source (wget says "No data received"). Edit: Ah, got it now. I see you're using Zhang's version. So I dunno why tom.ty89's fontconfig changes didn't take effect - maybe Gnome 3 so stubborn that it overrides cairo, or Zhang's patch is inferior to mine, or tom.ty89 would rather argue than apply fontconfig rule changes correctly
Last edited by brebs (2013-07-07 03:28:29)
Offline
According to tom.ty89's manifesto, he is lazy and doesn't find tweaking his system anything worth his efforts, so unless there's a key phrase he recognizes in the post he's reading ("bug", "regression", "mainstream plot", etc.), he won't waste his time going through it and just 'kind of skips it'. I believe his mentality must automatically blacklist those once suspected of wrong-thinking… Anyway, I don't think he even read what some of us were trying to explain because most of it violated his EULA, so let sleeping tom.ty89 lie…
As for my server: yes, it sucks because it gets the hiccups, but it's still better than the previous one which could choke all days long. I'd love to change it once again for something right as rain this time, but unfortunately that's impossible at the moment. Usually after a couple of minutes the content of the repository is available again, though.
The entire bundle thing has been tested in all available environments, including Gnome 3, and seems to be working quite well, even with the usual pains KDE/Qt or Gnome are notorious for. The user is only supposed to install it and go through a checklist to eliminate possible problems.
:: Registered Linux User No. 223384
:: github
:: infinality-bundle+fonts: good looking fonts made easy
Offline
Haha. Then there's the occasional madness, like cairo respect fontconfig.
And now the "autohint won't fix" is another ocassional madness. Maybe they think I should propose it to upstream
Your fontconfig changes didn't take effect. That's VERY BAD, and it's been a soure of annoyance for years. Don't you think THAT'S a bug? So try the "respect fontconfig" patch. Because when I add fontconfig rules (in XFCE, with my "cairo respect fontconfig" patch), they DON'T get ignored.
In your conf you set a true autohint AND a false autohint. I do notice that the there are two "test", but it simply works when the false one is removed.
Maybe you should take a look at this. Found in 10-autohint.conf in conf.avail.
This configuration is available on the major desktop environments.
We shouldn't overwrite it with "assign" unconditionally.
Most clients may picks up the first value only. so using "append"
may simply works to avoid it.
But is it relevant at all? Not sure. I am the kind of jerk that don't study things.
@bohoomil
It's not really just about one word. You said freetype 2.5 is working fine for you.
Of course it is. You are using a whole different set of packages. But I didn't know that.
Besides, we are using the most recent freetype2 in infinality-bundle,
Worse still, I somehow misunderstanded every piece of information in this sentence.
So I just thought that what you're suggesting is not what I am looking for, because I THOUGHT YOU WERE HAVING FUN WITH THE GENERIC 2.5.
Instead of fixing my problem, it will actually just FURTHER BEAUTIFY IT A BIT ON TOP OF WHAT I SAW. (You know what, that's the case in generic 2.4)
But I considered and still consider what I saw is INSANE.
Last edited by tom.ty89 (2013-07-07 08:08:15)
Offline
In your conf you set a true autohint AND a false autohint. I do notice that the there are two "test", but it simply works when the false one is removed.
And your point is what? You seem to be missing my points completely.
My previous fontconfig ruleset, which included a tweak for *bold* Cantarell, *worked*. You can see from the 3 screenshots I posted earlier in this thread, that it was taking effect and making a visible difference.
I have since removed the bold tweak, because it's not appropriate anymore (freetype is continually getting bugfixes), and I've switched to the PFB variant of Cantarell (which is narrower), rather than TTF. I may check out the OTF variant later, to see if there are yet more slight differences.
Now then, on to my real point. YOU are the only one who has fontconfig rules that don't take effect. So good luck debugging that, and yes that means effort on your part. Run an app from the console, and look for any messages, especially moans from fontconfig.
Offline
@brebs
My point is, I wondered if I have misunderstood you points, or maybe it only works with infinality-bundle. Maybe when someone shares my case and reads this thread, he would have known the bold tweak should be left out, especiall as what you said, it's not appropriate anymore.
Anyway, I've got what I want now. Thank you.
@end
Here is my redundunt summary:
Cantarell OTF provides far better rendering when comparing with the TTF. Yet the new Adobe CCF engine made that improvement completely gone.
Considering one is using generic freetype2:
With freetype 2.4, autohint is merely an optional enhancement. Yet there was a Arch patch in Cantarell 0.0.12, even symlinked to conf.d by default.
With freetype 2.5, autohint become a must-have for Cantarell. But then the patch is removed completely since 0.0.13, not even found in conf.avail.
Since the patch probably won't be added back again forever, one can symlink or copy 31-cantarell.font in conf.avail to conf.d.
Then open 10-autohint.conf in conf.avail and extract the following line:
<edit name="autohint" mode="append"><bool>true</bool></edit>
put it on the 31-cantarell.conf in conf.d so that the file become:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE fontconfig SYSTEM "../fonts.dtd">
<!--
Define aliasing and other fontconfig settings for Cantarell.
-->
<fontconfig>
<!-- Generic name assignment -->
<match target="pattern">
<test qual="any" name="family">
<string>Cantarell</string>
</test>
<edit name="family" mode="assign" binding="weak">
<string>Cantarell</string>
</edit>
<edit name="autohint" mode="append"><bool>true</bool></edit>
</match>
</fontconfig>
In case anything unwanted comes up, one may try to remove the following part:
<edit name="family" mode="assign" binding="weak">
<string>Cantarell</string>
</edit>
Last edited by tom.ty89 (2013-07-07 11:27:01)
Offline