[Feature] Add a SQL Database #14
Comments
dlatypov
added a commit
to dlatypov/lkml-gerrit-bridge
that referenced
this issue
Nov 18, 2020
Problem: Non-deterministic change ids mean we get duplicate changes uploaded to Gerrit for the same patch if the Bridge gets restarted. When issue google#14 is implemented, this will not matter as much, but there's still a chance for updates not to be persisted in the DB. Solution: Use a deterministic change id, i.e. "I" + <commit sha of v1>. If we try and create a Gerrit change for the same patch, I'm assuming the collision with the pre-existing change id will prevent us from doing so. Alternative: As noted in the file, we could consider using a `applypatch-msg` hook. But we can do more if we edit the commit directly from Python, since we have access to more information, e.g. the lore link. Add in a "Lore-Link" footer while we're here to demonstrate this. This iteration of the code is essentially a direct translation of what I had originally prototyped in bash for the hook.
dlatypov
added a commit
to dlatypov/lkml-gerrit-bridge
that referenced
this issue
Nov 18, 2020
Problem: Non-deterministic change ids mean we get duplicate changes uploaded to Gerrit for the same patch if the Bridge gets restarted. When issue google#14 is implemented, this will not matter as much, but there's still a chance for updates not to be persisted in the DB. Solution: Use a deterministic change id, i.e. "I" + <commit sha of v1>. If we try and create a Gerrit change for the same patch, I'm assuming the collision with the pre-existing change id will prevent us from doing so. Alternative: As noted in the file, we could consider using a `applypatch-msg` hook. But we can do more if we edit the commit directly from Python, since we have access to more information, e.g. the lore link. Add in a "Lore-Link" footer while we're here to demonstrate this. This iteration of the code is essentially a direct translation of what I had originally prototyped in bash for the hook.
dlatypov
added a commit
to dlatypov/lkml-gerrit-bridge
that referenced
this issue
Nov 18, 2020
Problem: Non-deterministic change ids mean we get duplicate changes uploaded to Gerrit for the same patch if the Bridge gets restarted. When issue google#14 is implemented, this will not matter as much, but there's still a chance for updates not to be persisted in the DB. Solution: Use a deterministic change id, i.e. "I" + <commit sha of v1>. If we try and create a Gerrit change for the same patch, I'm assuming the collision with the pre-existing change id will prevent us from doing so. Alternative: As noted in the file, we could consider using a `applypatch-msg` hook. But we can do more if we edit the commit directly from Python, since we have access to more information, e.g. the lore link. Add in a "Lore-Link" footer while we're here to demonstrate this. This iteration of the code is essentially a direct translation of what I had originally prototyped in bash for the hook.
bjh83
pushed a commit
that referenced
this issue
Nov 18, 2020
Problem: Non-deterministic change ids mean we get duplicate changes uploaded to Gerrit for the same patch if the Bridge gets restarted. When issue #14 is implemented, this will not matter as much, but there's still a chance for updates not to be persisted in the DB. Solution: Use a deterministic change id, i.e. "I" + <commit sha of v1>. If we try and create a Gerrit change for the same patch, I'm assuming the collision with the pre-existing change id will prevent us from doing so. Alternative: As noted in the file, we could consider using a `applypatch-msg` hook. But we can do more if we edit the commit directly from Python, since we have access to more information, e.g. the lore link. Add in a "Lore-Link" footer while we're here to demonstrate this. This iteration of the code is essentially a direct translation of what I had originally prototyped in bash for the hook.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Currently, message_dao.py is just a wrapper for a dictionary. This should be changed to use a SQL Database like PostgreSQL.
The text was updated successfully, but these errors were encountered: