Add failing test with synchronous event sourced saga #112
Conversation
|
Amazing :) How important is this for you? From what I understood in the issue, you changed the flow, so no hurry with this one? |
|
We corrected the flow in one case, but we found it in another case, fixed it, and then back again after a refactoring, making the synchronous flow pretty unusable. Pretty hard to debug as we receive only a postgres "duplicate key error" from our projection (first |
|
Hey @unixslayer :) |
Sure. I'll find some time after this weekend. |
|
Great thanks @unixslayer :) |
|
👀 |
|
Model fails in two different circumstances and setups. In Behat test, initialization of SagaProjection is not triggered even it is configured and registered correctly. Although, I've noticed that I've also written unit test using @jlabedo if you grant me access to your fork I can push additional commit with changes I've made. |
|
Hello @unixslayer , i gave you access to my repository, thanks! |
unixslayer
force-pushed
the
fix-es-handler
branch
from
95467aa
to
1d5b2a9
Compare
unixslayer
force-pushed
the
fix-es-handler
branch
from
1d5b2a9
to
6e0557c
Compare
unixslayer
force-pushed
the
fix-es-handler
branch
from
6e0557c
to
bc848d2
Compare
| } else { | ||
| $namedEventHandlers[$chanelName][] = $chanelName; | ||
| $namedEventHandlers[$chanelName] = array_unique($namedEventHandlers[$chanelName]); | ||
| } |
There was a problem hiding this comment.
This is small improved to avoid registering event handler by class name routing by name, if not defined explicitly.
|
@jlabedo @unixslayer This is solved by #139 :) |
Test to reproduce #104
The test is failing because the SagaProjection is not initialized. In fact, the event handler is executed as a "normal" event handler and not in the context of a "projection" handler