Ar an seachtú lá déag de mí Iúil, scríobh Stephen J. Turnbull:
[...] This introduces a serious backward incompatibility with GNU
Emacs
and XEmacs 21.4, without adding functionality beyond an attractive
convenience.
A forward compatibility problem rather than one of backward compatibility,
surely? #r"your text here" gives a syntax error in 21.4 and GNU Emacs, so
there’s no code in the wild for those platforms that uses that syntax, old
code works fine.
Compatibility problems will arise when people choose to use raw string
syntax in code they release, which I agree should not be done for a few
releases. (Anyway, when you’re developing your code, you can use raw strings
for your regexes, and then hit C-u C-x C-e to insert the normal string
representations of same, then comment out the raw strings. Though that is
ugly and error-prone, I suspect)
And attractive convenience _is_ added functionality!
There is a policy question here, and it should be discussed
(including
whether we should be more interested in compatibility with SXEmacs than
GNU Emacs!) before committing IMO.
We share packages with SXEmacs, so compatibility there is relevant. I do not
mean to suggest that it’s more relevant than compatibility with GNU. (And
the posted version of the patch is incompatible with SXEmacs if you use
Unicode escapes; I didn’t post the patch primarily because of SXEmacs
compatibility, I posted it mostly because I think it’s a really useful
feature and I’d like to see it in XEmacs.)
(Tangent; oh Jesus fuck does anyone have any suggestions as to where to
start when looking for the dropping-characters-with-Linux-and-ncurses 21.5
bug!?! This is driving me insane.)
I'm not going to veto (I like raw strings!), but please give it
some
time (a couple weeks, rather than the usual 48 hours) before you
conclude that lack of discussion == tacit approval.
Sure.
--
Santa Maradona, priez pour moi!