branch warnings (was Re: Hg update should prefer heads without bookmarks)

[cowwoc]

On 24/06/2015 11:23 AM, Bryan Murdock wrote:
> On Tue, Jun 23, 2015 at 11:19 AM, Matt Mackall <mpm at selenic.com> wrote:
>> We should similarly
>> issue a note or a warning for people who try to make their first commit
>> on a branch not named default.
> Do you mean in addition to the "branches are permanent and global, did
> you want a bookmark?" warning that hg gives when people run hg branch?
>   I still don't understand the need for that warning, let alone an
> additional warning when someone goes to commit on that branch.  Are
> branches really that dangerous and scary?  Especially when they are
> (still at this point) local to a clone?  I agree that there should be
> some friction when someone goes to *push* a new branch (and
> --new-branch takes care of that just fine), but not for creating and
> committing to a branch in a local clone.

The problem is that by the time you'd get the warning (at push time) you 
might have already constructed multiple commits and moving them to a 
bookmark would be a pain in the behind. I support the existing timing of 
the warning.

Gili