fix(core): correctly skip asset bundling

[plumdog]

Fixes #21925.

cxapi.BUNDLING_STACKS refers to stacks by their logical path within the app, so use this.node.path when comparing.

Previous behaviour only worked because:

• For "top level" stacks, the default name == id == path
• For stacks nested one layer deep in a stage, default name == ${stage}-${id} == path.replace('/', '-')

I've added tests that fail when:

• The stack has a custom name, not the id
• The stack is not nested one layer deep in a stage, so .replace('/', '-') doesn't fudge the path into the name

Question: while I have verified the above regarding cxapi.BUNDLING_STACKS experimentally, I haven't found anything that documents this.

---

All Submissions:

• Have you followed the guidelines in our Contributing guide?

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache-2.0 license

[gitpod-io]

[plumdog]

I would like to add a link here to something that documents cxapi.BUNDLING_STACKS.

[mrgrain]

Hey @plumdog thanks for looking into this! Unfortunately your PR is basically a revert of this revert: #21174
Whatever we do, we need to make sure we don't break bundling inside a pipeline stack again.

PS: I have little context on the issue, just remembered the revert. cc @kaizencc @corymhall

[aws-cdk-automation]

AWS CodeBuild CI Report

• CodeBuild project: AutoBuildv2Project1C6BFA3F-wQm2hXv2jqQv
• Commit ID: 7aac474
• Result: SUCCEEDED
• Build Logs (available for 30 days)

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository

[plumdog]

@mrgrain aha.

My remaining question is, I think "what is cxapi.BUNDLING_STACKS and - crucially - how can a stack tell from that whether it should bundle?"

I know it was a revert, but I'd be curious to see the broken behaviour that #21174 fixed - in other words, what would a failing for that PR look like?

[mrgrain]

@plumdog The issue surfaced as the error described in #18459 when bundling in a Pipeline stage (But not necessarily the same scenarios the issue describes. However the issue feels also very relevant tbh).

There is also an other active PR attempting to fix this: #21248

[plumdog]

@mrgrain Gotcha! So yeah, my findings match with those of #21248.

Thank you for your help, but I'm happy to close this PR and wait for #21248.