[PATCH 5 of 5 STABLE RFC] dispatch: ignore --early-list-opt that might not be a flag (BC)

Yuya Nishihara yuya at tcha.org
Sat Nov 18 22:05:37 EST 2017


On Fri, 17 Nov 2017 17:39:09 -0500, Augie Fackler wrote:
> On Wed, Nov 15, 2017 at 09:54:24PM +0900, Yuya Nishihara wrote:
> > # HG changeset patch
> > # User Yuya Nishihara <yuya at tcha.org>
> > # Date 1510321154 -32400
> > #      Fri Nov 10 22:39:14 2017 +0900
> > # Branch stable
> > # Node ID f1892a7fd3fdca0401072041774be62a1774bacd
> > # Parent  c2c34cf080aefbd983320bdaca111ebbd0826ff3
> > dispatch: ignore --early-list-opt that might not be a flag (BC)
> 
> I've gone ahead and queued patches 1-3 for stable, as they seem like
> fairly straightforward correctness wins here.
> 
> I'm really conflicted on these last two. The test changes in this one
> suggest to me it'll be at least somewhat painful in the real world, so
> it should probably be opt-in. We could at least document how to opt-in
> on `hg help scripting`, which would probably be enough to catch the
> eye of people doing security-relevant things, assuming they read docs
> at all (which is probably a stretch.)

Okay, I'll make this opt-in, which will be enabled by e.g. HGPLAIN=. A config
knob wouldn't be useful as the option can't be enabled by --config argument.

Also I'll check if this can be simpler by applying more strict parsing rules.


More information about the Mercurial-devel mailing list