On 2019-04-17 02:37, Andrew Cagney wrote:
On Tue, 16 Apr 2019 at 15:05, Sad Clouds <[email protected]> wrote:
Does it actually need to be distributed? If no, then what's wrong with
Subversion? Personally, I can't stand Git.
Subversion fails on two counts:
- it isn't ACID (I'm told that's the correct DB term)
In subversion parallel pushes are magically merged, maybe. For
instance: developer #1's deletes a .h macro and developer #2 adds a .c
use. ACID guarantees that #1 XOR #2 wins, and the other gets to sort
out the mess.
I'm not sure I get it.
Are you saying that subversion would interleave two commits? Commits in
subversion are supposed to be atomic. And each commit gets a
monotonically increasing commit number. Which also gives you in which
order the commits happened.
The fact that the two commits would be dealing with separate files means
there are absolutely no conflicts here, so I could/would expect that
both commits would go through. The fact that one deletes some things
that the other make use of is not something any VCS really can
understand or deal with. It would be no different from if I were to
commit some code that didn't work at all (for example make use of some
macro that never existed in the first place).
If the commits would be hitting the same file, I really expect that the
second commit would fail, with a requirement that the person first
update to head before committing. At which point any conflicts or merges
between the two commits should be sorted.
- it doesn't have true branches
Instead it has conventions; lots of conventions; and based on
experience, projects end up discovering that developers have been
following all of them
That is my main problem/issue with subversion. It also don't have tags,
which I also don't like.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: [email protected] || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol