| Title | Too time-consuming to find all problems with migration |
| Status | closed |
| Priority | essential |
| Assigned user | Gareth Rees |
| Organization | Ravenbrook |
| Description | With our current process, migrating jobs involves repeatedly running migrate.py, finding a problem with a particular job not being translated for some reason, adjusting the translate_jobspec and prepare_issue functions to take account of the exceptional case, and running migrate.py again. Since migration starts with the first job each time, this procedure is potentially O(n^2) in the number of jobs. |
| Analysis | There should be some way of getting migration to start again from where it failed last time. However, this shouldn't be the only behaviour, because you do often want to start again from the beginning (if you missed a field, or if a field got translated wrongly). So perhaps a command line option: python migrate.py -s job000123. |
| How found | manual_test |
| Evidence | <http://www.ravenbrook.com/project/p4dt...c/2001-11-27/release-1.3.0-test-report/>, item 5. |
| Observed in | 1.3.0 |
| Introduced in | 1.3.0 |
| Created by | Gareth Rees |
| Created on | 2001-11-27 16:04:22 |
| Last modified by | Gareth Rees |
| Last modified on | 2001-12-10 19:59:59 |
| History | 2001-11-27 GDR Created. |
| Change | Effect | Date | User | Description |
|---|---|---|---|---|
| 24465 | closed | 2001-11-27 18:49:51 | Gareth Rees | Added default value for migrated_user_groups and documented migrated_user_{groups,password}. Added -s option to migrate script to start migrating from a particular job. Documented a processmail-related error that can come up when migrating. Added test configuration for migrating Ravenbrook jobs to Bugzilla. |