Paul Hammant asked:
> Not answering your question, but could you explain more about your
> 'serverless' use - sounds intriguing
The idea is to do the simplest SVN example that could possibly work.
Since TortoiseSVN on Windows can work without a server for a single user
and local repo, I assumed it could be done on Linux with plain SVN.
My main focus is on showing how a domain-specific modelling tool can be
integrated with version control systems, on multiple platforms at once,
and possibly with multiple simultaneous users. Since the modelling tool
is the focus, I wanted the SVN side of things to "just happen", with
a minimum of fuss for the user.
I'd built integration for TortoiseSVN on Windows (here's a video if
it makes it clearer: https://www.youtube.com/watch?v=Uxnh8Hs77gY ),
Git on both Windows and Linux, and was checking out SVN on Linux. I'd
already taken care of the multi-user case, with Windows and Linux
editing the same models and versioning to the same remote SVN. Now
I was adding the simplest single user case: one user on Linux with a
local SVN repository. Oddly, there's not much on the net about that,
but it does seem to work fine with file: URLs without a server process.
All the best,
Steve
PS I'm a big fan of trunk-based development - not for everyone, maybe,
but a boon if you're working with model files, for which merge really
doesn't work. The multi-user modelling tool repository takes care of
integrating users' work, and the version control system just accepts
the results with thanks as the next trunk version. Details here if you
are interested: https://modeling-languages.com/smart-model-versioning/
Received on 2018-08-16 13:36:41 CEST