[prev in list] [next in list] [prev in thread] [next in thread] 

List:       subversion-commits
Subject:    =?UTF-8?Q?Re:_svn_commit:_r1874634_[3/3]_-_in_/subversion/branches/decou?= =?UTF-8?Q?ple-shelving-cl
From:       "Daniel Shahaf" <d.s () daniel ! shahaf ! name>
Date:       2020-02-29 13:48:04
Message-ID: 6ec5de7e-ae13-455d-aa52-4b6a184a0d96 () www ! fastmail ! com
[Download RAW message or body]

Julian Foad wrote on Sat, 29 Feb 2020 08:12 +00:00:
> 
> Daniel Shahaf wrote:
> 
> 
> > 
> > This doesn't seem forward-compatible: if in 1.15 we add
> > SVN_EXPERIMENTAL_COMMANDS=foo, using that environment variable setting
> > on 1.14 (assuming the above branch is merged by 1.14) would disable
> > shelving. I'm aware that compatibility is a lesser concern for experimental
> > features, but still, this sort of compatibility seems achievable.
> > 
> > 
> 
> 
> I didn't want this commit to change the default, but I think it would 
> be reasonable now to change the default to a simple opt-in (each 
> feature is enabled iff the var contains its feature-name) for 1.14.
> 
> WDYT?

Off the top of my head: that'd be a sensible design, but it's not the
only conceivable design that'd be sensible.  For example, designs that
enable some or all experimental subcommands by default would be sensible
too, since those subcommands' names already have "x-" prefixes to warn
of their experimental status.

Cheers,

Daniel
[prev in list] [next in list] [prev in thread] [next in thread]