Re: Ugliness in the Learning Manual

[Graham Percival]

On Sun, Aug 07, 2011 at 11:22:48AM +0100, Phil Holmes wrote:
> ----- Original Message ----- From: "Jean-Charles Malahieude"
> <address@hidden>
> >(./usage.texi (/home/jcharles/GIT/Mentors/tex/texinfo.tex

That looks good.

> git grep TEXINPUTS gives:
> 
> make/lilypond-vars.make:TEXINPUTS=$(top-src-dir)/tex/::
> 
> So we're setting the location of the texinfo.tex file explicitly
> during make.
> 
> Think this makes the issue invalid, but the "feature" is probably
> worth documenting with other make stuff - Graham?

What's the '"feature"' ?

I don't think it's worth discussing this specific environment
variable on its own.  However, I'd quite like to see either:

1) change the build system to avoid using environment variables --
or rather, print the exact env vars on each line.  i.e. instead of
make(1) producing:
  export TEXINPUTS=$(top-src-dir)/tex/::
  texi2pdf blah.texi
it would produce
  TEXINPUTS=$(top-src-dir)/tex/:: texi2pdf blah.texi

Advantage: when investigating a failing command, you can actually
copy&paste the command you see make(1) running, and it should fail
in exactly the same way because there's no environment variable
garbage.
Disadvantage: I'm not 100% certain that all shells handle
environment variables in the same way.  I know that non-bash
shells don't use "export" to set an env var.  This would involve
more research about unix shell portability, and/or en email from
somebody knowledgeable about such matters.

2) list all environment variables used (both internally and
externally) in the build system in the CG.

Advantage: at least this knowledge is written down somewhere.
Disadvantage: the list will not be kept up-to-date.  (don't argue;
there's absolutely nothing you can say that will make me believe
that everybody will keep it up-to-date over the next 20 years)


Cheers,
- Graham