Le mardi 05 décembre 2006 à 13:02 -0800, Keith Packard a écrit :
> On Tue, 2006-12-05 at 21:46 +0100, Loïc Minier wrote:
> > On Tue, Dec 05, 2006, Josselin Mouette wrote:
> > > The fix is to have these packages register their fonts to defoma. This
> > > will automatically run fc-cache -f in the def
On Tue, 2006-12-05 at 21:46 +0100, Loïc Minier wrote:
> On Tue, Dec 05, 2006, Josselin Mouette wrote:
> > The fix is to have these packages register their fonts to defoma. This
> > will automatically run fc-cache -f in the defoma directory.
>
> Either defoma is borken, or dh_installdefoma doesn't
On Tue, 2006-12-05 at 20:53 +0100, Loïc Minier wrote:
> On Tue, Dec 05, 2006, Keith Packard wrote:
> > I think I could fix upstream fontconfig to do a more careful check when
> > fc-cache is run and finds font files newer than the cache for their
> > directory. I don't do that at library initializa
On Tue, Dec 05, 2006, Josselin Mouette wrote:
> The fix is to have these packages register their fonts to defoma. This
> will automatically run fc-cache -f in the defoma directory.
Either defoma is borken, or dh_installdefoma doesn't generate the
appropriate snippet for what you describe. ttf-d
On Tue, Dec 05, 2006, Keith Packard wrote:
> I think I could fix upstream fontconfig to do a more careful check when
> fc-cache is run and finds font files newer than the cache for their
> directory. I don't do that at library initialization time for
> performance reasons, as font files are general
On Tue, 2006-12-05 at 20:08 +0100, Loïc Minier wrote:
> On Tue, Dec 05, 2006, Keith Packard wrote:
> > Right, the current DejaVu font package has a broken version of condensed
> > which does not correctly report the setwidth value in the OS/2 header of
> > the file. I think this will cause the inco
On Tue, Dec 05, 2006, Keith Packard wrote:
> Right, the current DejaVu font package has a broken version of condensed
> which does not correctly report the setwidth value in the OS/2 header of
> the file. I think this will cause the incorrect font selection error
> that we've seen in this case.
I
On Tue, 2006-12-05 at 15:29 +0100, Loïc Minier wrote:
> I don't see how this relates to my analysis. I upgraded fontconfig,
> and the fonts were ugly. I removed DejaVu Condensed, and the fonts
> were nice again, and the result of fc-match changed as well; certainly
> you can explain what par
On Mon, Dec 04, 2006, Josselin Mouette wrote:
> I think Loïc's analysis is wrong here
(But certainly Cc:ing him would help getting his feedback.)
>as XUL isn't using the condensed
> version of the DejaVu (or Vera) fonts. For most pages, it is using the
> N
Le lundi 04 décembre 2006 à 12:37 +0100, Mike Hommey a écrit :
> > I first thought of a configuration issue, and I can confirm that the
> > configuration files are exactly the same. Reverting the library brings
> > back Bitstream Vera in XUL applications.
> >
> > This doesn't seem to be a problem
On Mon, Dec 04, 2006 at 10:36:00AM +0100, Josselin Mouette <[EMAIL PROTECTED]>
wrote:
> Le dimanche 03 décembre 2006 à 18:27 -0800, Keith Packard a écrit :
> > On Mon, 2006-12-04 at 00:23 +0100, Josselin Mouette wrote:
> > > I see the same situation while my fonts are Bitstream Vera.
> > >
> > >
Le dimanche 03 décembre 2006 à 18:27 -0800, Keith Packard a écrit :
> On Mon, 2006-12-04 at 00:23 +0100, Josselin Mouette wrote:
> > I see the same situation while my fonts are Bitstream Vera.
> >
> > I think Loïc's analysis is wrong here, as XUL isn't using the condensed
> > version of the DejaVu
On Mon, 2006-12-04 at 00:23 +0100, Josselin Mouette wrote:
> I see the same situation while my fonts are Bitstream Vera.
>
> I think Loïc's analysis is wrong here, as XUL isn't using the condensed
> version of the DejaVu (or Vera) fonts. For most pages, it is using the
> Nimbus Sans fonts. Which m
I see the same situation while my fonts are Bitstream Vera.
I think Loïc's analysis is wrong here, as XUL isn't using the condensed
version of the DejaVu (or Vera) fonts. For most pages, it is using the
Nimbus Sans fonts. Which means there is a serious regression somewhere,
as this bug had been fi
14 matches
Mail list logo
