[RFC] Better font filetype and metadata file detection for xfs initscript

[Mike A. Harris]

On Fri, 10 Oct 2003, Thomas Dodd wrote:

>> So am I, but any time something changes, I can't contact n random
>> developers and tell them what chagnes to incorporate into their
>> packages very easily, and if I make a document and put it
>> somewhere, I have no guarantee anyone will read it.  I've direct
>
>You did what you could. It's up to the otheres to do their part.
>Post the doc, add it to the XF86 rpms with a link to the current 
>version. After that it's up to others to follow.
>Just like when any other developer changes a system used by others.
>I cna tell you it's changed, but cannot force you to change.
>(I still miss the XF86 setup tools. I usually build them myself though.
>If someone wrote secondary config tool that required the XF86 versions, 
>it's not your job to fix it. They were told to change, it's up to them 
>to do so.)

The XFree86 supplied config tools are gone for 2 main reasons:

1) They really do suck.

2) In many cases, people use them simply because they don't know 
   what our tool is called, or can't find Xconfigurator, or some 
   similar issue.  They just want to configure X, and so they end 
   up using those tools, and end up with tonnes of problems
   trying to get X to work, and often file bug reports, or have 
   to hit the lists.

3) The more people using xf86config or xf86cfg, or X -configure, 
   the less people using and testing our tool, so the less bugs 
   we're aware of, and the less bugs we will be likely to fix in 
   our tool.

The biggest issue for me, is people using the broken XFree86 
supplied tools and expecting them to work.  They truely are not 
very reliable, and they open up a _lot_ of doors for people to 
shoot themselves square in the foot.

Technically inclined users who really do want to use those tools 
for whatever reason, such as yourself, tend to be able to figure 
out how to compile them themselves fairly easily, and those 
people also are responsible enough usually to figure out any 
problems they encounter, etc.


>> There is also a problem where all fonts could theoretically be 
>> used by both xfs and also fontconfig/Xft, however we only want 
>> the given fonts to be in one or the other of those 2 systems for 
>> whatever preferential reasons we have.  One example is the Luxi 
>> TTF font.  Last I checked, we only enable this in core fonts and 
>> not in fontconfig, because the Luxi Type1 font is nicer looking 
>> in fontconfig than Luxi TTF.
>
>Sounds like the fonts need fixed. The Type1 and TTF version should look 
>the same (within the limits of the format). In the above, the Luxi TTF 
>should be fixed, possibly removed untill it is fixed.

The Luxi TTF font isn't broken, so it can't be fixed.  The Type1 
font looks nicer because the Type1 font rendering technology 
isn't hindered by alleged patents like truetype rendering is.  
Also, I'm not sure if Luxi even has hints in it, and wether or 
not they're decent.

I can disable the Luxi TTF font if Owen and others think that's a
good idea.  I think keeping it out of fontconfig should be 
adequate though more or less.


-- 
Mike A. Harris     ftp://people.redhat.com/mharris
OS Systems Engineer - XFree86 maintainer - Red Hat