Re: What's the deal with the module system?

[David Kastrup]

John Mandereau <address@hidden> writes:

> Le mercredi 25 novembre 2009 à 10:48 +0100, David Kastrup a écrit :
>> Sounds like a dependency impacting developers rather severely.
>
> Are you joking?

I do not know the matter enough to tell funny from serious suggestions.
And "Sounds like..." is a statement how it appears to me, not
necessarily how it is, given more information.

> Is requiring an officially released version and not supporting a
> version from CVS excessive?

When the code requires meddling with the internals of Texi2HTML, like I
understood Reinhold, it is in some parts linked to the particular
version.  In this case it makes sense to distribute the required version
alongside since no other version can be known to work, anyway.

> We have no control on Texi2HTML development, and in particular on its
> customization interface.

Again: I have no expertise here but just interpreted Reinhold's
statement.

>> Maybe it would make sense to place a copy of texi2html into the
>> Lilypond source tree as long as Lilypond can't avoid meddling with
>> internals of texi2html?  Otherwise arbitrary breakage with versions
>> out of the control of Lilypond will remain to be expected.
>
> As Texi2HTML weighs 500 KB (more than 22000 lines of code, which is
> roughly as much as Makeinfo in C!!!), so it's been decided not to
> include it.

Optional Makefile targets for downloading and compiling a particular
version in-tree when configure finds that the current one will not work?

Or any configure or error messages or error catching that will give a
useful information linking this failure of the test suite with the
Texi2HTML version?

Or Makefile targets (and reference to them in configure/error messages)
which don't produce the HTML information and thus don't fail when
Texi2HTML is not of the right version?

If the situation is unavoidable, at least I think one should put
information about it right at the place where things _will_ explode.

It will save you lots of annoying messages from people like me.  And it
might cause some more people to actually run the checks.

-- 
David Kastrup