bzr fast-import from git fails on utf8 conversion

[Jan Rovins]

on Ubuntu 10.04.3 LTS

I am running an import from a local git repository.

it dies with the following message, after running for a few minutes:
----------------------------
pruning empty directory Development/PassaicXmlRpcConnector/PASSAICXMLRPCCONNECTOR (:40878)
12:35:05 8200/18664 commits processed at 1922/minute (40921)
pruning empty directory Development/SysMgmt/ssShare/Inc (:41391)
ABORT: exception occurred processing commit :41410
bzr: ERROR: exceptions.UnicodeDecodeError: 'utf8' codec can't decode byte 0x92 in position 49: unexp ected code byte

*** Bazaar has encountered an internal error. This probably indicates a
    bug in Bazaar. You can help us fix it by filing a bug report at
        https://bugs.launchpad.net/bzr/+filebug
    attaching the crash file
        /root/.cache/crash/bzr-20120127173508-6194.crash
    and including a description of the problem.

    The crash file is plain text and you can inspect or edit it to remove
    private information.
---------------------------------

Crash file is attached.

[Jelmer Vernooij (jelmer)]

The input file isn't strict utf8, which is tripping bzr-fastimport up. Git doesn't have an explicit encoding, but bzr does.

bzr-fastimport should probably give a better error message in this case, but it can't really import this file without resorting to heuristics.

It should be possible to manually fix up the file by editing the fastimport file.

[Jan Rovins (janr-f)]

Thanks for the info,

How would I go about the Manual fixup of the import file? It appears to be in binary format.
Are there any documntation and reccomended tools or utilities for working with the fast-import/export file formats?

[Jelmer Vernooij (jelmer)]

I think the main tools for manipulating fastimport/fastexport files are available in the Git community. bzr-fastimport also has some (see "bzr help commands" and look for 'fastimport'), but they're probably not as well-polished.