| Commit | Date | |
|---|---|---|
|
[r936]
by
tart
AbstractDef now shielded. The unloading is still somehow an unknown for me, so will be resolved later. |
2017-01-25 00:08:59 | Tree |
|
[r935]
by
tart
TransException is now not caught on places where FatalException wasn't being caught. This is needed for the proper working of Shielded. |
2017-01-24 23:15:06 | Tree |
|
[r934]
by
tart
Fixing FieldValue tests. The parallelism one is still red as we need to stop translating TransException to SEException. I'll try to fix that globally. |
2017-01-24 22:45:17 | Tree |
|
[r933]
by
tart
Rewriting the usage of Shield.Intransaction (yes it's far from done and I'm already rewriting :) I think it makes generally more sense to not try to make every single method "atomic" by introducing a transaction. Instead we'll just generally assert a transaction in public methods in "shielded" classes, and the actual transactions will be started on just a few "global" places - generally there where we also handle exceptions. That way the transaction-ness doesn't get out of control and at the same time asserts make sure we have transactions where we need them. |
2017-01-24 21:06:31 | Tree |
| 2017-01-24 19:57:27 | Tree | |
| 2017-01-24 19:52:59 | Tree | |
| 2017-01-24 19:50:33 | Tree | |
| 2017-01-24 19:42:39 | Tree | |
|
[r928]
by
tart
FieldValue now shielded. This is probably the prototype of how all transactional classes will look - the internal state carried by a Shielded struct. This is just the first converted class to make even just the loading transactional. Happy experimenting ahead :) |
2017-01-24 01:18:22 | Tree |
| 2017-01-14 15:28:30 | Tree |