On 2/11/14, Nicolas Mailhot nicolas.mailhot@laposte.net wrote:

Le Lun 10 février 2014 12:34, Alec Leamas a écrit :

...

• The upstream spec [2] does a lot of stuff in %post: (mkfontscale,

mkfontdir...). How should I cope with this?

Legalities aside the upstream spec just uses a packaging style that was current in Fedora at the start of the millenium. It tries very hard to make stuff like X core fonts work, when Fedora and all knowledgeable people have long since determined it was hopeless and should be left to die quietly (go wayland go).

I guess the only reason it was not updated is that everyone with a clue has realised it was not a good idea to propagate an obsolete set of fonts that can not be updated or fixed due to legal reasons.

So to answer you questions:

1. you won't lose anything significant by using our current font packaging

templates instead of the upstream one 2. it's a very very bad idea to package those fonts and expose them to anyone. It encourages developer laziness "I can hardcode windows fonts metrics they're available under Linux", it's a legal trap "I can bundle windows fonts with my app, after all the Linuxes do it", it's a lie (the actual MS fonts people get with windows are several generations later, with lots of technical fixes and coverage enhancements, so you'd be pushing spoiled goods to innocents).

Really there is a ton of nicer and more modern fonts on Google fonts library waiting for someone to package. Just ignore the mscorefonts "common knowledge" which has been cargo-culted in Linux user FAQs for years and you'll render a service to everyone involved. We're not in the 90's any more there are lots of other opentype fonts to choose from (don't you realise "core fonts" means "we are crushing you now Netscape". That's ancient history, a few browser wars away!)

-- Nicolas Mailhot

Hmm...

Thanks for input! It clarified some things I somehow suspected without really knowing.

My overall idea so far has been to try to cope with these "Fedora post installation" guides by incorporating the things often recommended in rpmfusion. This is basically a "give users what they want" strategy which of course cannot be used in each and every case.

You have good reasons not to give users these fonts. So, this then boils down to "Give users what they should have" vs "Give users what they want". For me, this is not a clear-cut decision.

Part of this is also a question how these fonts works in different locales. Karel Volny gave some interesting remarks on this in [1].

Anyway, I have submitted three review requests 3176, 3177 and 3178 for mscore-fonts, mscore-tahoma-fonts and cleartype-fonts. Perhaps there are different trade-offs in each case, dunno.

I'm definitely open to just throwing some or all of these requests in the trash, should this be the right thing to do. After all, I write this after installing all the stuff, and it does *not* look better on the screen, on the contrary :)

Cheers!

--alec

[1] https://bugzilla.rpmfusion.org/show_bug.cgi?id=2483#c53