On Thu, Oct 28, 2010 at 10:36:40PM +0200, Gianluca Montecchi wrote:
> On Thursday 28 October 2010 21:11:26 W. Trevor King wrote:
> > On Thu, Oct 28, 2010 at 03:13:21PM +0200, Gianluca Montecchi wrote:
> > > I have not a way to sort the targets in the order I want, since when
> > > I read the bug list, they are reported in creation order, with the
> > > newer first.
> > 
> > Cheap answer:
> > 
> > I think targets can be ordered by with dependency trees, and sorted by
> >   $ ./be depend -t 0 /773
> >   bea/773 blocked by:
> >     bea/275:fm: Fix Unicode handling.
> > Although of course you'd want to reformat the output for HTML.
> 
> Ok, but this order all the bugs of this target, that is interesting but it is 
> not what I need (for this time).

Some of the dependencies are earlier targets.  Perhaps a better
example of would have been:

  wk...@thialfi be $ ./be depend --severity target --status all -t 0 /47c
  bea/47c blocks:
   bea/4fc:ct: patch-52
    bea/bd0:ft: 0.2
   bea/bd0:ft: 0.2

> Bonus question: why  'be depend /773' output
> 
> bea/773 blocked by:
> bea/275
> 
> while 'be depend -t 0 /773' output
> 
> bea/773 blocked by:
>  bea/275:fm: Fix Unicode handling.
> 
> ? 
> 
> Should not be the output coherent ?

Good point.  Fixed in my branch.

> > Mid-range answer:
> > 
> > I suggest people just pick a sane target naming scheme that can be
> > sorted correctly by distutils.version.LooseVersion (see PEP 386 [1]),
> > and we add a LooseVersion-based summary sorter.
> 
> Right, but still, no way of saying that v2.1 will be released before release 
> 1.3 and after 1.2. There is always the possibility that I miss something...
> PEP 386 is usefull if you have only one tree, not it you have two tree 
> because 
> you have a stable tree and a devel tree.
> 
> > Luxury answer:
> > > Also the depend field it is not usefull in this occasion, since while it
> > > is still true that a 1.1 version will be released after a 1.0 version,
> > > it is not always true that a 1.2 version will be released before a 2.0
> > > version, so
> > 
> > If the order of 1.2 vs 2.0 is not defined, we can list them in
> > whatever order we want. 
> 
> Right, but add a 1.3. 
> The output should be
> 1.2
> 2.0
> 1.3
> 
> since 1.3 depend on 1.2, but 2.0 is indipendent from any 1.x version

If 2.0 is independent, I think the target list would be fine as:
  1.2
  1.3
  2.0

Also, if they were independent, they would probably be in independent
VCS branches so the 1.3 target would not exist in the .be/ repository
of the 2.0 branch ;).

> > > I cannot set that V2 depend on V1.2, but even if I do it, I still get
> > > that V2 will be showed before V1.2 since it is newer.
> > 
> > This is because of the default sort (libbe.bug.cmp_full) sorts by
> > time.  The list sorting code (libbe.command.list.List._sort_bugs)
> > shows how you can sort a list of bugs by any parameter you like.
> 
> I find it, but it seems that it does not work.

Can you post an example of it not working?

> Of course I can develop some way to force a order in the "be html" command, 
> like a reverse order or something else, at least passing explicitally a 
> string 
> to the html command that specify the order.

Again, I think LooseVersion will do a good enough job.  How many open
targets do we expect to be dealing with here?

> I think that also for the command "be list" can be usefull to display the 
> bugs 
> sorted by target (and sort the target reflecting the programmed release plan)

From the command line, I would do that with:
  for bug in $(be list --severity target --ids); do
    echo $bug;
    be depend --show-status --show-summary $bug;
  done
If you have a nice interface bundle that into a single `be list`
call, go ahead ;).

> Anyway, thinking about it a little more, actually, what I need is more 
> something like "what is the next target I should complete ?" than "what is 
> order of the tags ?"

That sounds like libbe.bugdir.BugDir.target.  Perhaps it needs a
better interface than `be set` and `be target --resolve`.

-- 
This email may be signed or encrypted with GPG (http://www.gnupg.org).
The GPG signature (if present) will be attached as 'signature.asc'.
For more information, see http://en.wikipedia.org/wiki/Pretty_Good_Privacy

My public key is at http://www.physics.drexel.edu/~wking/pubkey.txt

Attachment: pgpxjrfeC9vPC.pgp
Description: PGP signature

_______________________________________________
Be-devel mailing list
[email protected]
http://void.printf.net/cgi-bin/mailman/listinfo/be-devel

Reply via email to