On 3/22/17 8:29 AM, Gregory Szorc wrote:
On Mar 22, 2017, at 04:51, Ryan McElroy <r...@fb.com> wrote:
On 3/22/17 7:29 AM, Sean Farley wrote:
Yuya Nishihara <y...@tcha.org> writes:
On Sun, 12 Mar 2017 21:38:00 -0700, Gregory Szorc wrote:
# HG changeset patch
# User Gregory Szorc <gregory.sz...@gmail.com>
# Date 1489378362 25200
# Sun Mar 12 21:12:42 2017 -0700
# Node ID d30057d358076cbe7d632cd573095af97543f932
# Parent 1c3352d7eaf24533ad52d4b8a024211e9189fb0b
show: new extension for displaying various repository data
The idea sounds nice to me. I just checked minor implementation details
about formatter.
Just a quick reply (as I whittle down my backlog), but a lot of people
(including myself) have a 'show' alias (inspired by 'git show').
That may or may not be a factor in this.
Greg called this out specifically in his excellent summary.
FB also has a "show" extension that replaced our "show" alias:
https://urldefense.proofpoint.com/v2/url?u=https-3A__bitbucket.org_facebook_hg-2Dexperimental_src_default_hgext3rd_show.py&d=DwIGaQ&c=5VD0RTtNlTh3ycd41b3MUw&r=nuarHzhP1wi1T9iURRCj1A&m=KZNXs3QMg2M7YtCvXYwgouTxwTN2WFhsYLOLHtLfE2Y&s=qWD0RFF73TrSukNubnCJl-EdsGcgjpl5olyQNX88XQ4&e=
At Paris, people from Facebook mildly objected to "show" because of the
conflict.
At Mountain View, Durham said to just use "show" and FB would deal with it.
I myself was convinced to use "display" after Paris and that's what I initially
implemented. Justification for "display" was summarized in that RFC patch's commit
message. However, enough people said the reasons weren't strong enough and Durham gave me a green
light, so I changed to my preference: show.
Anyway, the command and extension name is a bikeshed and can be changed before
4.2. I'd prefer review focus on the behavior so we can get something landed and
iterate on the feature. I *really* want a smartlog/wip feature in the core
distribution and this extension/command is where it will live.
I'd just pick what makes the most sense for Mercurial users, and not
worry about what people have aliased, etc. The conflict with 'git show'
is a valid concern though.
For 'hg view', you could argue that hgk is just one implementation of
this new 'view' paradigm, and maybe it could change it's command line to
accept the same bookmark/smartlog/etc inputs as the proposed 'hg
view/show'. That might let us justify using 'hg view'
_______________________________________________
Mercurial-devel mailing list
Mercurial-devel@mercurial-scm.org
https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
