Disconnected operations

[Michael V. Savchuk]

Hello.

As far as i understand true disconnected operations are
possible (mean without instatnt commiting). Is it
technically possible to use prefix to each issue depending
on working copy?
For example, we have 2 working copies, one has prefix "a",
second has prefix "b".

After adding some issues directory structure will be look
like:
issues/
ia1
ia2
ib1
ia3
ib2

or something similar. That will remove problems preventing
using DiTrack as a backend to another program when using
disconnected operations.
Didn't look at the code, so please let me know is it
possible and very difficult or not to implement.

Yours,
Mickhail Savchuk.

[Luke Kenneth Casson Leighton]

hi michael,

there's some code in ditrack which uniquify's individual messages
added to issues, but not individual issues themselves.

the code creates and keeps a "local" reference to the individual
message until such time as it has been synchronised.

personally i believe that a better approach would be to commit the
message regardless, but with a UUID (128-bit random identifier) and
also to commit the issue regardless (again with a UUID), and "leave a
blank space" to be filled in with the "global" number 1, 2, 3, 4....
etc.

then when a global sync is performed, the globally unique reference
number is filled in.

you could also have, to make the local [offline] committer's time
slightly easier, a "local" number <username>.1, <username>.2 or some
other locally-understood but globally unique identifier.

l.

Post by Michael V. Savchuk
Hello.
As far as i understand true disconnected operations are
possible (mean without instatnt commiting). Is it
technically possible to use prefix to each issue depending
on working copy?
For example, we have 2 working copies, one has prefix "a",
second has prefix "b".
After adding some issues directory structure will be look
issues/
ia1
ia2
ib1
ia3
ib2
or something similar. That will remove problems preventing
using DiTrack as a backend to another program when using
disconnected operations.
Didn't look at the code, so please let me know is it
possible and very difficult or not to implement.
Yours,
Mickhail Savchuk.
_______________________________________________
Dev mailing list
http://lists.ditrack.org/mailman/listinfo/dev