Re: fail on regression test

[Ikumi Keita]

>>>>> Arash Esbati <arash@gnu.org> writes:
> Sounds good :-)  I see that currently we don't parse the second optional
> date argument after the package name.

I skimmed over v2 and v3 documents of siunitx briefly. If I don't miss
important points, the difference between them is that v3 adds some new
macros (e.g. \qty), depreciates some macros (e.g. \si), and drops
support for "s" type of column specification of tabular environment.

Then why don't we just add support for new macros and retain supports
for old macros as well as "s" column? In that way, single siunitx.el
would serve both v2 and v3 of siunitx.sty.

In addition, v3 document says, in "8 Compatibility with other packages",
that there are cases where the macros "\SI" and "\si" are needed to
avoid macro name clash with other packages. So siunitx.el should provide
support for at least those two even for siunitx.sty v3.

(Maybe siunitx.el should refrain from adding "\qty" support when
physics.sty is loaded and "\unit" support when units.sty is loaded in
the document.)

Regards,
Ikumi Keita