skeema / tengo

93%

main: 94%

LAST BUILD ON BRANCH flavor-patch

branch: flavor-patch

CHANGE BRANCH

x

Reset

• flavor-patch
• bulkdropopts
• ci-golint-fix
• col-reorder-algo
• conn-refactor
• dep-golang-19
• fix-gomod-err
• fk-order-mysql55
• foreign-keys
• generated-cols
• ghactions
• index-ordering
• introspection-params
• key-col-name-only
• main
• mariadb106
• master
• partitioning
• refs/tags/v0.10.0
• refs/tags/v0.10.1
• refs/tags/v0.10.3
• refs/tags/v0.9.4
• refs/tags/v0.9.5
• refs/tags/v0.9.6
• refs/tags/v0.9.7
• remove-caching
• reuse-containers
• routines
• routines-rebased
• v0.8.1
• v0.8.10
• v0.8.11
• v0.8.12
• v0.8.13
• v0.8.14
• v0.8.15
• v0.8.16
• v0.8.17
• v0.8.18
• v0.8.19
• v0.8.2
• v0.8.3
• v0.8.4
• v0.8.5
• v0.8.7
• v0.8.8
• v0.8.9
• v0.9.0
• v0.9.1
• v0.9.2
• v0.9.3

Build # 206

Build Type

push

travis-ci

Committed by evanelias

Commit Message

Flavor: include patch version; misc fixes for MySQL 8.0.19

This commit adds a Patch field to the Flavor struct, and adjusts several
Flavor-related methods to now account for its presence.

This change adds complexity and may affect backwards-compatibility in subtle
ways. However, this commit is fundamentally necessary to account for MySQL 8
now regularly including functionality changes in post-GA patch releases. This
package's Flavor abstraction predates MySQL 8, and previously the patch
release version was not relevant for conditional logic anywhere in this
package.

The specific motivating changes here were MySQL 8.0.19's omission of integer
display widths from information_schema and SHOW CREATE TABLE, and ordering
changes for foreign keys in SHOW CREATE TABLE. The former did not break the
package's internal behavior, but it did cause a massive number of
incompatibilities in the integration test suite. The latter prevented tables
with multiple FKs from being supported for diff. This commit fixes both of
these problems.

One other obscure change: since 8.0.19 retains display widths specifically
for signed tinyint(1), for simplicity's sake this commit changes a couple test
fixtures to instead use tinyint(1) unsigned, which does still get stripped.

Run Details

47 of 47 new or added lines in 2 files covered. (100.0%)

3002 of 3243 relevant lines covered (92.57%)

240.56 hits per line