Re: grep: fix multibyte regex handling under macOS (1819ad327b7a1f19540a819813b70a0e8a7f798f)

From: Jeff King <peff@peff.net>
Date: Feb 7, 2023, 18:19 UTC
Message-ID: <Y+KWFyjTa8t5d5Q8@coredump.intra.peff.net>
In-Reply-To: <CANgJU+UJ-p54MDmQWYTEsdh_ss-q8TUEcqHFm568Lpfsy2dU+w@mail.gmail.com>

On Sat, Feb 04, 2023 at 12:32:09PM +0100, demerphq wrote:

Show 12 quoted lines

> > I think this is a perennial question for portable software: is it better
> > to be consistent across platforms (by shipping our own regex engine), or
> > consistent with other programs on the same platform (by using the system
> > regex).
> 
> Personally I think that while this seems to be an impartial reading of
> the question at hand I think it frames the debate in a way that has
> the potential to bias[1] the discussion in favour of a particular
> policy outcome[2]. It implies that all other things are equal between
> the two options presented, and frames the question as something that
> comes down to personal preference between one form of consistency and
> another.

Sorry, I didn't mean to imply one side or the other, or even that they were equal. I more just meant: you can make arguments either way.

_I_ don't have a strong opinion there, but you have made many of those arguments in your email. And everything you said was pretty sensible.

I did want to mention one thing, though:

> It also assumes that being consistent with other programs on
> the same platform is inherently beneficial, when that doesn't seem to
> be clearly established[2].

This isn't consistency per se, but one problem that arises when a portable program ships its own implementation is that its implementation sucks more than the system one. And that was the case with 1819ad327; our compat/regex was worse than the system with respect to multi-byte characters.

But yes, if your point is: "well, don't ship a crappy implementation", then I agree that is a way to mitigate that problem. ;)

-Peff