1 |
joko |
1.1 |
|
2 |
|
|
propagating tangram v2 schema changes - proposal |
3 |
|
|
|
4 |
|
|
what we want is: |
5 |
|
|
- being able to propagate changes in database schemas between different development- and production areas |
6 |
|
|
- this should occour neither without interfering with the used relational schema nor should affect any data |
7 |
|
|
- it should apply to changes on table and field level |
8 |
|
|
- the default "safe" mode should be to propagate additions and warn about other changes which might be: |
9 |
|
|
- field removals |
10 |
|
|
- field additions |
11 |
|
|
- field type changes |
12 |
|
|
|
13 |
|
|
what we know about tangram v2 is: |
14 |
|
|
- we should use explicit class ids in tangram schema, take care to properly increase them for each new class! |
15 |
|
|
- new classes get automatically introduced by a native tangram schema-deploy |
16 |
|
|
|
17 |
|
|
what remains is: |
18 |
|
|
DBI-Error [Tangram]: DBD::mysql::st execute failed: Unknown column 'UserData.account' in 'field list' |
19 |
|
|
DBI-Error [Tangram]: DBD::mysql::st fetchrow_array failed: fetch() without execute() |
20 |
|
|
|
21 |
|
|
this could be solved through: |
22 |
|
|
MySQL::Diff by Adam Spiers ( http://adamspiers.org/computing/mysqldiff/ ) |
23 |
|
|
|