merge-upstream in an empty branch doesn't specify --create

[James Westby]

Doing merge-upstream in an empty branch does three odd things.

Firstly it tries to add a changelog entry at the end, but it doesn't use --create.

Secondly, you get negative revision numbers for the upstream revisions.

Thirdly, it removes any .bzrignore.

Perhaps the first two are both ok, or perhaps merge-upstream shouldn't be allowed
in this case, with something else being done instead.

Thanks,

James

[Muharem Hrnjadovic (al-maisan)]

An additional glitch needs fixing. The following error occurs when calling "bzr merge-upstream" in an empty branch.

{{{
Traceback (most recent call last):
  File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 835, in exception_to_return_code
    return the_callable(*args, **kwargs)
  File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 1030, in run_bzr
    ret = run(*run_argv)
  File "/usr/lib/python2.6/dist-packages/bzrlib/commands.py", line 647, in run_argv_aliases
    return self.run(**all_cmd_args)
  File "/home/mhr/.bazaar/plugins/builddeb/cmds.py", line 617, in run
    merge_type=merge_type)
  File "/home/mhr/.bazaar/plugins/builddeb/import_dsc.py", line 1617, in merge_upstream
    previous_upstream_revision = get_snapshot_revision(previous_version.upstream_version)
AttributeError: 'NoneType' object has no attribute 'upstream_version'
}}}

[Muharem Hrnjadovic (al-maisan)]

re. "it removes any .bzrignore":

I did not find anything specific to merge-upstream that removes .bzrignore files. This is rather a general filter feature of imports (see import_dsc.import_archive() and should_ignore() for more detail).

Is this still to be considered buggy behaviour? I guess not.

[James Westby (james-w)]

On Sat Sep 19 06:06:04 UTC 2009 Muharem Hrnjadovic wrote:
> re. "it removes any .bzrignore":
>
> I did not find anything specific to merge-upstream that removes
> .bzrignore files. This is rather a general filter feature of imports
> (see import_dsc.import_archive() and should_ignore() for more detail).
>
> Is this still to be considered buggy behaviour? I guess not.

I think it is. I am not sure we should be removing .bzrignore
at any time, but I think certainly not in merge-upstream with
a tarball and a branch.

Thanks,

James

[Muharem Hrnjadovic (al-maisan)]

James Westby wrote:
> On Sat Sep 19 06:06:04 UTC 2009 Muharem Hrnjadovic wrote:
>> re. "it removes any .bzrignore":
>>
>> I did not find anything specific to merge-upstream that removes
>> .bzrignore files. This is rather a general filter feature of imports
>> (see import_dsc.import_archive() and should_ignore() for more detail).
>>
>> Is this still to be considered buggy behaviour? I guess not.
>
> I think it is. I am not sure we should be removing .bzrignore
> at any time, but I think certainly not in merge-upstream with
> a tarball and a branch.

Hello James,

thank you very much for commenting on this. I have disabled the
filtering of ".bzrignore" files altogether (please see the enclosed
incremental diff).

Should you want to disable ".bzrignore" filtering for merge-upstream
operations specifically, please let me know.

Best regards

--
Muharem Hrnjadovic <email address hidden>
Public key id : B2BBFCFC
Key fingerprint : A5A3 CC67 2B87 D641 103F 5602 219F 6B60 B2BB FCFC