On 11 April 2018 at 22:37, Hans Hagen wrote:
> On 4/11/2018 10:08 PM, Brian Hunt wrote:
>> 3. Add texlua (and texluac?) to
>>    a. the setup rsync bin/ directories
>>    b. the tex/texmf-linuxmusl(|-ppc|-64)
> we don't use texluac, don't need texlua on windows and on linux texlua is
> just a symlink

This was entirely my fault which has been fixed in the meantime.

> if musl needs different bin then there should be a different initial
> download i think because basically we're now talking of: windows, osx,
> linux, linuxmusl (a different species)

Exactly. We have that. And most bits and pieces are in place to allow
that, except for the problem with mtxrun not being able to determine
the platform correctly.

>> 4. Update mtxrun
>> Here's code that ought to work for detecting musl in mtxrun.  Around line
>> 3721 of the mtxrun add:
>>      if find(architecture,"x86_64",1,true) then
>>        platform="linux-64"
>>      elseif find(architecture,"ppc",1,true) then
>>        platform="linux-ppc"
>>      else
>>        platform="linux"
>>      end
>> +   if resultof("ldd --version 2>&1"):find("musl") ~= nil then
>> +     platform = platform:gsub("linux", "linuxmusl")
>> +   end
> that's not going to happen, some system call 9each run) in order to
> determine what linux species we run ... (there are a few calls but these are
> never used becasue we use HOSTTYPE (if needed we can set one ourselves on
> these platforms)
> isn't there some magic env var that tels if musl is used? do programs and
> scripts really have to jumpo through such loops

Hans, we don't really need this code for the sake of minimals
themselves, it is part of ConTeXt code that needs it and fails.

The problem is that even if I properly detect the platform upfront,
your code (mtx-update) will refuse to respect the setting. When
    ./bin/mtxrun --script ./bin/mtx-update.lua --force --update --make
--context=beta --platform=linuxmusl-64 --texroot="/some/path"
from musl linux, the mtx-update will insist in always installing and
using linux-64 binaries even if they don't work, no matter what else
we try.

I still believe that a string like "x86_64-linuxmusl" is something
that luatex itself should be able to return when asked about what
platform it is running on (what platform it has been compiled for).
We've discussed this with Taco many years ago, but this has never been

If your question is of interest to others as well, please add an entry to the 

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://context.aanhet.net
archive  : https://bitbucket.org/phg/context-mirror/commits/
wiki     : http://contextgarden.net

Reply via email to