[lltx] [luaotfload] Footnotes

[Will Robertson]

On 11/05/2010, at 12:42 PM, Khaled Hosny wrote:

> On Fri, Apr 09, 2010 at 09:09:23AM +0200, Khaled Hosny wrote:
>> On Fri, Apr 09, 2010 at 10:09:47AM +0930, Will Robertson wrote:
>>> The following defaults will happen, I think:
>>> 
>>> \fontspec{Latin Modern Roman} % <- name
>>> \fontspec{ec-lmr10}           % <- name
>> ec-lmr10 is a tfm file, but anyway the name mechanism falls back to
>> filename if no font is found (but going directly to filename search is
>> faster).
>> 
>>> \fontspec{ec-lmr10.otf}       % <- file (from the extension)
>>> \fontspec[Path]{ec-lmr10}     % <- file
>>> 
>>> ([Path] is the same as [ExternalLocation].)
>>> 
>>> This is the same as currently happens, so I can't see any problems.
>> 
>> OK, once the change is made in fontspec, I'll change luaotfload
>> defaults.
> 
> Will, any updates on this? I want to make the change before uploading to
> CTAN, but I'm not sure how broken fontspec will be then.


To confirm: this means that you'll no longer be compatible with XeTeX if people use the plain \font syntax. Is this actually okay? I'm starting to think it might not be the best idea.

From the point of view of fontspec it doesn't make any difference, but I just want to make sure we're not raising another problem in trying to solve an edge case.

As a counterpoint, what would happen instead if the filename database was queried first and only if a font was not found in the search path a system (or named) font was used?

\font\tenrm=ec-lmr10        % file
\font\tenrm=file:ec-lmr10   % file
\font\tenrm=sys-ec-lmr10    % name if not found as a file
\font\tenrm={name:Latin Modern Roman} % name

This would still solve the problem from the beginning of this thread (I think), and looks to me a little more simple. Have I missed a problem with this approach, though?

-- Will