Trial git-based task list

Daniel Shahaf danielsh at apache.org
Mon Oct 24 21:04:54 UTC 2016


Ximin Luo wrote on Mon, Oct 24, 2016 at 19:54:00 +0000:
> $ git pull
> $ ./task add "Some stuff"
> $ ./task add priority:H "Some more important stuff"
> $ git push

And suppose the push gives an error because somebody else had just
pushed something, i.e. a race condition / conflict, what does one do
then?

I suppose the answer is "fetch, reset --hard origin/master, and replay
the last few task commands from the shell history".

> - raw data format is relatively simple, we can re-import it into something else if needed

There's «task export» that emits json.  I would use that in preference
to hacking together a parser.

> It's very lightweight, e.g. it doesn't support adding comments onto tasks.

It does; see «task annotate».  Annotations are one line each, but can
have any number of them.

> 

Note that by default, tasks are renumbered whenever a task is closed.
That means the small integer identifiers are volatile.  The permanent
identifier of a task is its uuid.

(There's a way to avoid renumbering with a «gc=off» config knob, but
upstream considers that "unsupported".)

Cheers,

Daniel



More information about the Reproducible-builds mailing list