[PATCH 6 of 6 V2] test-pattern: actually update tests using the patterns

Augie Fackler raf at durin42.com
Fri Nov 17 13:02:39 EST 2017


On Wed, Nov 15, 2017 at 05:07:06PM +0100, Boris Feld wrote:
> On Mon, 2017-11-13 at 18:09 -0500, Augie Fackler wrote:
> > On Mon, Nov 13, 2017 at 12:11:20PM +0100, Boris Feld wrote:
> > > # HG changeset patch
> > > # User Boris Feld <boris.feld at octobus.net>
> > > # Date 1509866592 -3600
> > > #      Sun Nov 05 08:23:12 2017 +0100
> > > # Node ID 715948f2256fa181c4f3730825492f9da0f74b2e
> > > # Parent  c06ed80fc1063905dc6a711fc32c88d5913b0db6
> > > # EXP-Topic better-substitute
> > > # Available At https://bitbucket.org/octobus/mercurial-devel/
> > > #              hg pull https://bitbucket.org/octobus/mercurial-deve
> > > l/ -r 715948f2256f
> > > test-pattern: actually update tests using the patterns
> >
> >
> > I love this feature. Buuuuut, could we get away with a non-.py file
> > for configuring the substitutions? Maybe just use a configparser file
> > for now?
> >
> > I'm hesitant to use Python files for config languages, it's been a
> > bad
> > experience in the past.
>
> We give a try to the INI format but it doesn't seem to handle well
> multi-lines and special characters which are painful for defining 200
> characters regexp.

For the record, this argument was not convincing.

>
> One argument in favor Python files is that we can have dynamic
> definition that computes values from the environment. run-test.py
> already has to do this for LOCALIP, so we expect more case to show up.

But this one seems unavoidable and hopelessly regrettable. Queued with regret.


More information about the Mercurial-devel mailing list