[PATCH 1 of 6] mercurial: implement a source transforming module loader on Python 3

Pierre-Yves David pierre-yves.david at ens-lyon.org
Fri Jun 17 12:56:38 EDT 2016


On 06/10/2016 05:53 AM, Augie Fackler wrote:
> On Mon, Jun 06, 2016 at 08:44:18AM -0700, Martijn Pieters wrote:
>> On 31 May 2016 at 22:52, Gregory Szorc <gregory.szorc at gmail.com> wrote:
>>> # HG changeset patch
>>> # User Gregory Szorc <gregory.szorc at gmail.com>
>>> # Date 1464759986 25200
>>> #      Tue May 31 22:46:26 2016 -0700
>>> # Node ID 867ebaa1ca956c76022655eb1f7a6f39da42b5a4
>>> # Parent  48b38b16a8f83ea98ebdf0b370f59fd90dc17935
>>> mercurial: implement a source transforming module loader on Python 3
> 
> [snip]
> 
>> From a Python 3 tech perspective, this looks sound to me. I like
>> hijacking the importer machinery for this over using a custom codec,
>> nice!
> 
> Yeah, after a lot (really) of reflection, this seems like the least
> awful thing, barring some sort of flag to annotate that allows
> skipping revisions in a revset (so the invasive b'' prefixing required
> could then be elided from annotate results.)
> 
> (I still think such a flag is a good idea, but am willing to not let
> the perfect be the enemy of the not-awful.)
> 
> What's the status of this series? It looks like I expect a new version
> to cope with some docstring encoding issues?

Same question here, what is the status of this?
there seems to be a consensus that this is an awful hack that solve a
massive problem the best way we currently can easily do.

As it seems like yuya spotted some issue, I'm dropping it from
patchwork, expecting a resend.

Cheers,

-- 
Pierre-Yves David


More information about the Mercurial-devel mailing list