[PATCH v3 rebased] pager: migrate heavily-used extension into core

[Sean Farley]

Augie Fackler <raf at durin42.com> writes:

>> On Feb 6, 2017, at 14:47, Bryan O'Sullivan <bos at serpentine.com> wrote:
>> 
>> # HG changeset patch
>> # User Bryan O'Sullivan <bryano at fb.com>
>> # Date 1486160890 28800
>> #      Fri Feb 03 14:28:10 2017 -0800
>> # Node ID ae22925dafd4a270cb80a7bb54c9d70bce49a633
>> # Parent  1f51b4658f21bbb797e922d155c1046eddccf91d
>> pager: migrate heavily-used extension into core
>
> Here's my counterproposal, not quite done enough to mail (mostly on account of it being late in my timezone):
>
> https://hg.durin42.com/hg-wip/log?rev=%40%3A%3A7c0170f0420f%20-%20%40&revcount=50
>
> It's 32 patches (but most of them are one line, turning on paging of a command), and probably breaks some hairy edges of the pager extension in the name of moving the behavior to core behind a more usable API. I only touched commands in core or that were mentioned in pager.attend, so there's another round of changes to make in hgext to get things fully on the pager bandwagon. That said: it should work! Tests pass modulo small things you'd sort of expect: lots of output churn from the new global variable, and one `help -k` test fails because I marked the pager extension as deprecated.
>
> I'll try and coordinate with someone on IRC tomorrow to go over this with me and get it mailed, since it sounded like everyone on the other thread was largely enthusiastic about this direction.

Wow, impressive!