>| Create a fontconfig alias that would create a virtual>| font-family "Nimbus Sans L Condensed". The following works for me: <match target="scan"> <test name="family" compare="eq"> <string>Nimbus Sans L</string> </test> <test name="width" compare="eq"> <int>75</int> </test> <edit name="family" mode="assign"> <string>Nimbus Sans L Condensed</string> </edit> </match> Nimbus Sans L urw fonts (unmodified) with width=75 (condensed) are re-assigned the family name Nimbus Sans L Condensed and are then selected by the css font-family: Nimbus Sans L Condensed. On Fedora: % fc-list "Nimbus Sans L Condensed": file style /usr/share/fonts/default/Type1/n019043l.pfb: :style=Regular Condensed /usr/share/fonts/default/Type1/n019044l.pfb: :style=Bold Condensed /usr/share/fonts/default/Type1/n019064l.pfb: :style=Bold Condensed Italic /usr/share/fonts/default/Type1/n019063l.pfb: :style=Regular Condensed Italic Bob T.
Le jeudi 06 mars 2008 ? 19:45 -0500, Bob Tennent a ?crit :> >| Create a fontconfig alias that would create a virtual > >| font-family "Nimbus Sans L Condensed". > > The following works for me:It works for you, when "working" is defined as making a specific font face available to web designers today ignoring all other considerations. It does not "work" for me where working is defined as a solution : ? which is not tied to a particular font (esp. one which is not actively maintained anymore, and pre-dates advances in screen legibility) ? which is not tied to a particular encoding set (direct result of tying to a semi-maintained font with limited coverage) ? which is not a fugly workaround we''ll have to keep ad vitam eternam (ie what is the migration plan mid-term when all the short-termists will have hardcoded the name of this font face everywhere) ? which is not a fugly workaround others are now going to ask for other fonts (because if we say yes there where will we stop?) ? which will not actively obstruct the resolution of the root bug (and we''ve seen how quick the mozilla developpers are to dismiss any need to work on it as long as there is a single condensed face available through family name hacks) In other words this snippet will offload problems from mozilla and you to the distributions foolish enough to accept it. And since Fedora wants problems fixed upstream instead of accepting the burden of piles of short-term workarounds that make their authors think the problems are solved and they can let the distro deal with straightening things out, I don''t think we want to accept it. Regards, -- Nicolas Mailhot
On 2008/03/07 11:10 (GMT+0100) Nicolas Mailhot apparently typed: ...> resolution of the root bug...> In other words this snippet will offload problems from mozilla... Mozilla''s Gecko (Firefox, SeaMonkey, Epiphany & others) is open source. https://bugzilla.mozilla.org/show_bug.cgi?id=3512 has no current assignee. Feel free to attach a patch, plagiarizing Webkit if necessary. :-) -- "Let us not love with words or in talk only. Let us love by what we do." 1 John 3:18 NLV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/
Le vendredi 07 mars 2008 ? 06:03 -0500, Felix Miata a ?crit :> On 2008/03/07 11:10 (GMT+0100) Nicolas Mailhot apparently typed: > > ... > > resolution of the root bug > ... > > In other words this snippet will offload problems from mozilla > ... > > Mozilla''s Gecko (Firefox, SeaMonkey, Epiphany & others) is open source. > https://bugzilla.mozilla.org/show_bug.cgi?id=3512 has no current assignee. > Feel free to attach a patch, plagiarizing Webkit if necessary. :-)Right now I''ll play the other well known open source tune "I disagree, feel fee to distribute your hack without me" -- Nicolas Mailhot -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: Ceci est une partie de message =?ISO-8859-1?Q?num=E9riquement?= =?ISO-8859-1?Q?_sign=E9e?Url : http://lists.freedesktop.org/archives/fontconfig/attachments/20080307/241e4fa7/attachment.pgp
>|Le jeudi 06 mars 2008 ? 19:45 -0500, Bob Tennent a ?crit :>|> >| Create a fontconfig alias that would create a virtual >|> >| font-family "Nimbus Sans L Condensed". >|> >|> The following works for me: >| >|It works for you, when "working" is defined as making a specific font >|face available to web designers today ignoring all other considerations. >| >|It does not "work" for me where working is defined as a solution : >|+ which is not tied to a particular font >| (esp. one which is not actively maintained anymore, and pre-dates >|advances in screen legibility) >|+ which is not tied to a particular encoding set >| (direct result of tying to a semi-maintained font with limited >|coverage) >|+ which is not a fugly workaround we''ll have to keep ad vitam eternam >| (ie what is the migration plan mid-term when all the short-termists >|will have hardcoded the name of this font face everywhere) >|+ which is not a fugly workaround others are now going to ask for other >|fonts >| (because if we say yes there where will we stop?) >|+ which will not actively obstruct the resolution of the root bug >| (and we''ve seen how quick the mozilla developpers are to dismiss any >|need to work on it as long as there is a single condensed face available >|through family name hacks) >| >|In other words this snippet will offload problems from mozilla and you >|to the distributions foolish enough to accept it. And since Fedora wants >|problems fixed upstream instead of accepting the burden of piles of >|short-term workarounds that make their authors think the problems are >|solved and they can let the distro deal with straightening things out, I >|don''t think we want to accept it. Blaming the lack of interest in font-stretch since 1999 on the snippet I circulated yesterday seems a bit of a stretch :+) It seems you inhabit a universe where font selection by applications is already as you want it and only Firefox is holding out. But the reality is that family and {regular,bold,italic,bold-italic} are the only choices available, not just in browsers, but in desktop applications of all sorts. AFAIK, LaTeX2e (and derived variants) are the only applications currently available that implement extended families. Until this reality changes, there is really no point in distributing fonts that are inaccessible to the vast majority of applications; if such fonts are distributed, they should be distributed in a way that allows their use, either changing the family name in the font or via the fontconfig hack I suggested. I would argue the same way for "modern" font families such as Vera and Liberation if they have condensed/expanded variants and don''t use family names to distinguish them from the conventional variants. The real problem with the Nimbus fonts is that distribution and development are fragmented; they''ve been distributed by ghostscript, abiword, TeX, and no doubt others, and some of the development has been unsatisfactory; so there is no evident reliable "upstream" to whom we can complain. But the reality is that they are the only fonts that can be found on every Linux box. Bob T.
Le vendredi 07 mars 2008 ? 11:16 -0500, Bob Tennent a ?crit :> Blaming the lack of interest in font-stretch since 1999 on the snippet I > circulated yesterday seems a bit of a stretch :+)I don''t blame the lack of interest on your snippet I blame your snippet in actively trying to perpetuate this lack of interest.> It seems you inhabit a universe where font selection by applications is > already as you want it and only Firefox is holding out. But the reality > is that family and {regular,bold,italic,bold-italic} are the only > choices available, not just in browsers, but in desktop applications > of all sorts. AFAIK, LaTeX2e (and derived variants) are the only > applications currently available that implement extended families. > Until this reality changes, there is really no point in distributing > fonts that are inaccessible to the vast majority of applications;It seems you inhabit a fantasy land where Nimbus Sans L Condensed is the only problem case, and where application developers are going to fix their bits without prompting while you remove the need for them to do so (by going on a workaround fest on all the fonts you can lay your hands on). Unfortunately? Nimbus Sans L Condensed is far from the only problem case and developers being busy people are perfectly happy to sit on problems a few more years if they find some fools to paper over symptoms for them in the meantime. A case in point is the infamous Firefox ligature bug, where removing ligatures from FLOSS fonts only led to 6 months of active inertia, and where the workaround had to be reverted before developers started looking at the problem (note that in the meantime there were lots of unhappy campers, both in users of FLOSS fonts that had no access to ligatures to play nice for Firefox, in users of non-FLOSS fonts that hit the bug but hadn''t the critical mass by themselves to get devs to look at the problem, and in users generally confused by the way the same font behaved differently depending on if someone had made a smartass surgery on it or not). To further contradict your thesis that ?{regular,bold,italic,bold-italic} are the only ?choices ?available, not just in browsers, but in desktop applications and that the only solution is to actively butcher fonts to appease broken apps, 1. The GTK2 font selector already supports extended faces perfectly 2. OpenOffice.org gained some access to extended faces in the past six months through a Fedora patch (the UI still stinks but the base capability is there, and UI aspects are being discussed now upstream). This because someone (me in that case) took the time to report the problem and convince OO.o devs instead of doing a OMG devs are not responsive let''s do a workaround font-side. Applications are on a collision course with complex fonts right now because Microsoft, Apple and Adobe have invested heavily in smart font formats such as OpenType (new fonts are for example a Vista selling point). At the same time, every major FLOSS font project released in the last months has used the OTF format since Fontforge and proprietary font tools besides now use it by default. The ?{regular,bold,italic,bold-italic} dumb font era is over and the sooner FLOSS apps identify the ways they need to cope with it the better for everyone involved. Delaying this work through workarounds like yours will only make the adaptation late and more painful. Therefore I''ll use whatever influence I have within the Fedora Fonts Special Interest Group to block any attempt to hide smart font problems under the carpet and reassure developers they can continue to assume last millenium rules apply. ?Save your energy for gathering support for the long-term fixes, don''t use it to promote short-termist bandaids. In case you wonder I wrote the current official Fedora font packaging policy. -- Nicolas Mailhot -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: Ceci est une partie de message =?ISO-8859-1?Q?num=E9riquement?= =?ISO-8859-1?Q?_sign=E9e?Url : http://lists.freedesktop.org/archives/fontconfig/attachments/20080307/16934cfb/attachment.pgp