fix(db/migration): do migration before validation #10348
Conversation
chronolaw
changed the title
StarlightIbuki
force-pushed
the
fix/on-the-fly-mig-validation
branch
from
ae59e39 to
1acd47f
Compare
| if not routes then | ||
| -- no need to migrate | ||
| return | ||
| local function table_default(val) |
There was a problem hiding this comment.
table_or_default or get_valid_table may be a better name.
|
Perhaps we should cherry-pick this PR to the 3.2 release branch. |
We choose to only handle nesting in top-level services. Fix KAG-638 Co-authored-by: Chrono <chrono_cpp@me.com> fix lint
StarlightIbuki
force-pushed
the
fix/on-the-fly-mig-validation
branch
from
7e75cbb to
227b5ee
Compare
Co-authored-by: Chrono <chrono_cpp@me.com>
| @@ -82,6 +82,8 @@ | |||
| [#10346](https://github.com/Kong/kong/pull/10346) | |||
| - Fix an issue where control plane does not rename fields correctly for `session` for older version of data planes. | |||
| [#10352](https://github.com/Kong/kong/pull/10352) | |||
| - Fix an issue where validation to regex routes may be skipped when the old-fashioned config is used for DB-less Kong. | |||
There was a problem hiding this comment.
| - Fix an issue where validation to regex routes may be skipped when the old-fashioned config is used for DB-less Kong. | |
| - Fix an issue where migration to regex routes may be skipped when the old-fashioned config is used for DB-less Kong. |
There was a problem hiding this comment.
No. It's not the migration being skipped. The migration will happen anyway.
There was a problem hiding this comment.
So the issue is about validation being skipped? Aka it should fail validation, but now it does not fail? I tought it was about migrating also service nested routes that was not done before.
| return | ||
| local function table_default(val) | ||
| -- we cannot verify it with type(val) == "table" because | ||
| -- we may recieve indexable cdata/lightuserdata |
There was a problem hiding this comment.
-- we may recieve indexable cdata/lightuserdata
When and why would this happen?
There was a problem hiding this comment.
The caller may pass an object representing yaml/json object, which could be wrapped c object.
There was a problem hiding this comment.
For the same reason, I'm using pairs instead of pairs. But let's give it a try and revert if it fails.
Co-authored-by: Aapo Talvensaari <aapo.talvensaari@gmail.com>
Co-authored-by: Aapo Talvensaari <aapo.talvensaari@gmail.com>
Co-authored-by: Chrono <chrono_cpp@me.com> Co-authored-by: Aapo Talvensaari <aapo.talvensaari@gmail.com> Co-authored-by: Enrique García Cota <kikito@gmail.com> (cherry picked from commit 2f82ced)
Co-authored-by: Chrono <chrono_cpp@me.com> Co-authored-by: Aapo Talvensaari <aapo.talvensaari@gmail.com> Co-authored-by: Enrique García Cota <kikito@gmail.com> (cherry picked from commit 2f82ced)
Summary
We choose to only handle nesting in top-level services.
Checklist
N/A There is a user-facing docs PR against https://github.com/Kong/docs.konghq.com - PUT DOCS PR HERE
Issue reference
Fix KAG-638