Re: keep smallquotation tests not handled by makeinfo in C?

[Patrice Dumas]

On Mon, Oct 12, 2020 at 10:34:30PM +0100, Gavin Smith wrote:
> On Mon, Oct 12, 2020 at 10:56:40PM +0200, Patrice Dumas wrote:
> > Hello,
> > 
> > In commit e3955e2c725bed949b1aecb3f57afb66832db54e 
> > * tp/tests/coverage_macro.texi: Remove some "@ifset texi2html"
> > conditions.
> > why were the @smallquotation section removed and not only the 
> > "@ifset texi2html"?  The test results looked good, and I believe
> > that they were only guarding against makeinfo in C.
> 
> I've got no problem with the @smallquotation section going back.  I
> didn't know that the test suite was once used with makeinfo in C.

In a transition period, I compared the results of makeinfo in C 
to the parser implementation in perl, especially for info, if I recall
well.  So the test suite was not used with makeinfo in C for makeinfo
in C, but for the transition. 

> I took out the texi2html conditionals themselves because they were
> causing me a lot of confusion when I was trying to accomplish something
> else, and texi2html is supposed to be an obsolete program anyway.

At some point it was usefull because we wanted compatibility with both
texi2html and makeinfo in C in texi2any but the convergence is done now
and specific texi2html feature have been included or abandonned.  I am
all for removing the texi2html @value.  I think that it is useful to
keep -c TEXI2HTML to avoid changing the tests and for the occasional
texi2html nostalgic.

As a side note, there was an heavy texi2html user which was left with
missing features with the new implementation in texi2any, it is
lilypond, and there was a bug opened in which I commented a bit but then
I stopped looking at that issue:
https://code.google.com/archive/p/lilypond/issues/1000

-- 
Pat