a few pointers needed for solving a bug

John Arbash Meinel john at arbash-meinel.com
Thu Mar 22 16:12:37 GMT 2007

Sabin Iacob wrote:
> John Arbash Meinel wrote:
>>> iacobs at moria ~/tmp/tst $ bzr init --dirstate-tags
>> I would guess that --dirstate or --dirstate-tags should perform the same
>> here. (since they use the same working tree format).
> looks like you are right :); I was trying to reproduce the exact
> conditions under which I first ran into problems
> reverted to 2365, and still getting the same:
> sequence of operations:
> * move everything from a directory somewhere else
> * delete the directory (forgetting to unversion it)
> * try to commit
> I get:
> missing some/test
> renamed some/test/dir => some/dir
> bzr: ERROR: The file id file-20070322150508-p1rbbllqejwguoi3-8 is not
> present in the tree <WorkingTree4 of /home/iacobs/tmp/tst2>.

^- well, this would seem that we are still having a bug. but at least it
v- this error.

> do bzr revert; results in assertion error:
> +N  some/test/
> R   some/dir/ => some/test/dir/
> R   some/filee => some/test/filee


> bzr: ERROR: exceptions.TypeError: unsubscriptable object

>  File "/usr/lib/python2.4/site-packages/bzrlib/workingtree_4.py", line
> 2128, in _iter_changes
>    for result in _process_entry(current_entry, None):
>  File "/usr/lib/python2.4/site-packages/bzrlib/workingtree_4.py", line
> 1909, in _process_entry
>    target_parent_id = target_parent_entry[0][2]
> TypeError: unsubscriptable object
> bzr 0.16.0dev0 on python 2.4.4.final.0 (linux2)
> arguments: ['/usr/bin/bzr', 'st']

If you want, send me the contents of '.bzr/checkout/dirstate'. That is
the file which tracks the current state of the workingtree. So I can use
it to figure out what it thinks is going on.

> However, revert actually does what it's supposed to do, the file tree
> gets restored to the previous state.
> I attached a script containing the sequence of operations (I attempted
> to build a proper blackbox test, but got lost :( )

I'll take a look at it. But I'd like to make sure that you end up
running with the latest bzr.dev. Since it seems to have at least a
partial fix for this bug.


More information about the bazaar mailing list