On 2026-03-08 01:47, 山卡洛 wrote:
why do you, after decades of running a non-mainstream solution (cvs),
default to git? why dont you even consider the cleaner alternatives
hg, jj, fossil?
please keep CVS if you can't come up with anything better designed
than git.
On Fri 6. Mar 2026 at 19:16, Peter G. <[email protected]> wrote:
On 03/03/2026 23:55, Nick Holland wrote>
So...give it a try at
http://cvsweb.openbsd.org
or
https://cvsweb.openbsd.org
work much faster than before
why not just ditch cvs for git and go cgit?
It works well for OpenBSD and has it has been for 29 years. There is no
point in evangelizing on git, hg, or any other version control system.
CVS works well for the use case here and if it's not broken, don't fix
it. If they had gotten on all the VCS bandwagons over the years, the
repo would have gone from CVS to SVN to Mercurial to Git by now. FreeBSD
went from CVS to SVN to Git over that same time and have had many
headaches over the years during those transitions and keeping version
control history sane.
Whenever people scream about wanting to change version control systems,
they forget all the tooling and process that is built around it.
Changing that isn't cheap or quick.
I've worked on projects that have kept very old version control systems
running because all of the tooling was built around it. I've seen
programs run for 20+ years on CMVC (just when it had been EOLed, they
eventually moved to git) and PVCS/Dimensions. The Dimensions people are
staying on that because it would cost too much money to port all of
their productivity aids and other utilities to git. That money would be
spent elsewhere on delivering actual code rather than rearranging the
furniture because someone wanted to put their mark on the project.
Simply put, if it works, it works. No need to change it.
Nobody is interested in your opinions about version control.
-Andy Wallis
