Error: Not implemented: Disallowing partial commit with structural changes

[Chuck Tuffli]

I'm getting the error:

Error: Not implemented: Disallowing partial commit with structural changes

Below is a terminal log showing the steps, but I'm not sure why this is a
partial commit as both files are explicitly listed. What am I goofing up?

[2000] vv init vv_partial vv_partial
[2001] cd vv_partial
[2002] vv user create ctuffli
[2003] vv user set ctuffli
[2004] echo foo > bar
[2005] ln -s bar baz
[2006] vv add bar baz
[2007] vv commit -m 'Add initial files'


revision: 2:e65072e58b2d754a849c803efbd642098dd5b5f7
branch: master
who: ctuffli
when: 2012/12/05 13:32:19.330 -0800
comment: Add initial files
parent: 1:b8ee751a5c1ba86f50f099ad5b856869f7f0cee8
[2008] echo flubber >> bar
[2009] vv rm baz
[2010] vv status
Modified: @/bar
Removed: @/baz
[2011] vv commit -m'Update + removal' /home/ctuffli/dev/vv_partial/bar
/home/ctuffli/dev/vv_partial/baz
Error: Not implemented: Disallowing partial commit with structural changes.

[2012] vv commit -m'Update + removal'


revision: 3:b54a32adfaf1c3bbe0056b726c35d211934c521c
branch: master
who: ctuffli
when: 2012/12/05 13:37:59.759 -0800
comment: Update + removal
parent: 2:e65072e58b2d754a849c803efbd642098dd5b5f7
[2013] vv version
2.1.0.10979 [64-bit LINUX]
[2014] uname -srvm
Linux 2.6.32-45-generic #100-Ubuntu SMP Wed Nov 14 10:47:49 UTC 2012 x86_64

[Jeff Hostetler]

Yes, you're right. In this case the list of args
is the complete set of the things that changed so
it isn't really a partial commit.

We found a problem with partial commits and disabled
it for the current release. (Consider the edge case
where you commit a move, but don't also commit *both*
parent directories.) I put a quick/dirty/hack test
in the current code to complain if *any* file/folder
args are given when there are structural changes
(moves/renames/add/deletes) -- rather than try to
actually fix the underlying problem (or to determine
if the given args makes a complete set). That was
too difficult/risky in the time permitting.

This has been fixed in the (yet to be released) 2.5
version.

jeff

Post by Chuck Tuffli
Error: Not implemented: Disallowing partial commit with structural changes
Below is a terminal log showing the steps, but I'm not sure why this is a partial commit as both files are explicitly listed. What am I goofing up?
[2000] vv init vv_partial vv_partial
[2001] cd vv_partial
[2002] vv user create ctuffli
[2003] vv user set ctuffli
[2004] echo foo > bar
[2005] ln -s bar baz
[2006] vv add bar baz
[2007] vv commit -m 'Add initial files'
revision: 2:e65072e58b2d754a849c803efbd642098dd5b5f7
branch: master
who: ctuffli
when: 2012/12/05 13:32:19.330 -0800
comment: Add initial files
parent: 1:b8ee751a5c1ba86f50f099ad5b856869f7f0cee8
[2008] echo flubber >> bar
[2009] vv rm baz
[2010] vv status
[2011] vv commit -m'Update + removal' /home/ctuffli/dev/vv_partial/bar /home/ctuffli/dev/vv_partial/baz
Error: Not implemented: Disallowing partial commit with structural changes.
[2012] vv commit -m'Update + removal'
revision: 3:b54a32adfaf1c3bbe0056b726c35d211934c521c
branch: master
who: ctuffli
when: 2012/12/05 13:37:59.759 -0800
comment: Update + removal
parent: 2:e65072e58b2d754a849c803efbd642098dd5b5f7
[2013] vv version
2.1.0.10979 [64-bit LINUX]
[2014] uname -srvm
Linux 2.6.32-45-generic #100-Ubuntu SMP Wed Nov 14 10:47:49 UTC 2012 x86_64
_______________________________________________
Veracity-users mailing list
http://lists.sourcegear.com/cgi-bin/mailman/listinfo/veracity-users

[Chuck Tuffli]

Post by Jeff Hostetler
Yes, you're right. In this case the list of args
is the complete set of the things that changed so
it isn't really a partial commit.
We found a problem with partial commits and disabled
it for the current release. (Consider the edge case
where you commit a move, but don't also commit *both*
parent directories.) I put a quick/dirty/hack test
in the current code to complain if *any* file/folder
args are given when there are structural changes
(moves/renames/add/deletes) -- rather than try to
actually fix the underlying problem (or to determine
if the given args makes a complete set). That was
too difficult/risky in the time permitting.
This has been fixed in the (yet to be released) 2.5
version.
jeff

Ok, I can live with that. Thanks Jeff!

---chuck