Comments on: fontAvailable: Test to see if a font is usable

By: tlrobinson

tlrobinson — Wed, 11 Mar 2009 06:49:29 +0000

To handle the case where you’re checking the font it’s checked against, you can just check against two different fonts.

Here’s the implementation in Cappuccino: http://gist.github.com/280north/cappuccino/blob/89516d4a3817f0c098afa220486e36211ceab24d/AppKit/CPFontManager.j

By: lalitpatel

lalitpatel — Wed, 11 Mar 2009 06:13:18 +0000

By: JimmyP22

JimmyP22 — Tue, 10 Mar 2009 17:10:52 +0000

You should probably do some cleaning up in the plugin. Multiple checks will mean a tonne of useless spans within the DOM.

By: ajaxery

ajaxery — Tue, 10 Mar 2009 15:52:47 +0000

Should probably set the font-size as well. Don’t ya think?

By: Jordan1

Jordan1 — Tue, 10 Mar 2009 15:50:45 +0000

I can’t believe this is hosted on Google Code. It seems more suited as a blog entry.

By: howardrauscher

howardrauscher — Tue, 10 Mar 2009 15:38:17 +0000

@ChrisCasciano
~85% of of home page views had tahoma. Even thought this number was a rough approximation and has to be taken with a grain of salt, I was hoping the number to be a lot lower so I would have a good reason to push for a wider used font like Arial.

I have been experimenting with a couple of different methods to make it more accurate. Testing against a matrix of generic families is a good idea.

By: ChrisCasciano

ChrisCasciano — Tue, 10 Mar 2009 15:11:16 +0000

Sooooooo, spill the beans Howard – How many of your users had Tahoma installed?

As for those worried about a few edge cases – this test could probably be beefed up by testing against a few different faces rather then just a fake font — for example test against “fake font” but also the monospace, serif and sans-serif generic families so you have a number of matrix to work off of. Even the most controlled environments [either by user preferences or by device type] will probably have different fonts set up for the different generics.

By: karf

karf — Tue, 10 Mar 2009 15:07:31 +0000

I think it would be better to set the font family to “monospace” and then to “__FAKEFONT__, monospace”. Then comapre the widths. This will of course not work for monospaced fonts, but use of monospaced font is a very rare case.

By: howardrauscher

howardrauscher — Tue, 10 Mar 2009 15:04:33 +0000

Yeah, testing for the default font is definitely a case where it breaks. I a couple of different revisions where I try to handle that case, and they all pretty much fail. The reason I submitted it to Ajaxian is so I could get ideas and some feedback.

@WebReflection
I never thought about checking for the height of the test element. Thanks

By: JimmyP22

JimmyP22 — Tue, 10 Mar 2009 14:43:44 +0000

This will fail when the default font is the same as the one you’re querying.

By: WebReflection

WebReflection — Tue, 10 Mar 2009 14:40:38 +0000

yep, that’s what I meant when I talked about times new roman but generally speaking, if charater width is the same of the default font, even if the height is different this function will return false.
Finally, as I already said, that element should be removed before the return ;-)

By: WebReflection

WebReflection — Tue, 10 Mar 2009 14:13:27 +0000

un-removed element a part, is this method that reliable? As example, times new roman returns false on windows … and I wonder how many fonts with the same character size could return false as well.