You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This will redesign resolver to work with
multiple edge case scenarios.
Previously pac resolver was filtering
pipelinerun based on annotations but it was
resolving all pipelines in .tekton dir
leading to resolving unnecessary pipelines
and other issue was it was storing task
based on task name, instead of annotation name
and version, so different version of task
were not used across pipelineruns in .tekton dir
Now with new design we are first filtering
pipelinerun based on annotations, and then
processing all pipelineruns one by one
and only resolving pipeline related to these
pipelineruns. Also we are now maintaining map
of tasks with name and version at event level
to not re fetch the task and now inline spec
replacement in pipelinerun is done one by one
so respective task as mentioned in annotation
with name and version is used
Also before filtering the pipelineruns, we should
make sure that no two pipelineruns exists with
same name in .tekton dir
Added tests for the three scenarios
0 commit comments