Don't overwrite unnecessary nodes #110
Labels
Comments
nonara
added a commit
that referenced
this issue
Jun 16, 2021
- Several improvements were made for speed and efficiency. - Now accommodating for new TS empty baseURL provision (closes #109) - Pre-checking necessity before overwriting paths (closes #110) - Rewrote core resolution methodology to: - Properly handle implicit indexes (closes #106) - Properly handle implicit sub-package indexes set via package.json 'main' #108) - Not follow symlinks (#107) - Resolve from output path as opposed to SourceFile path (#103)
nonara
added a commit
that referenced
this issue
Jun 16, 2021
- Several improvements were made for speed and efficiency. - Now accommodating for new TS empty baseURL provision (closes #109) - Pre-checking necessity before overwriting paths (closes #110) - Rewrote core resolution methodology to: - Properly handle implicit indexes (closes #106) - Properly handle implicit sub-package indexes set via package.json 'main' #108) - Not follow symlinks (#107) - Resolve from output path as opposed to SourceFile path (#103)
nonara
added a commit
that referenced
this issue
Jun 16, 2021
- Several improvements were made for speed and efficiency. - Now accommodating for new TS empty baseURL provision (closes #109) - Pre-checking necessity before overwriting paths (closes #110) - Rewrote core resolution methodology to: - Properly handle implicit indexes (closes #106) - Properly handle implicit sub-package indexes set via package.json 'main' #108) - Not follow symlinks (#107) - Resolve from output path as opposed to SourceFile path (#103)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Presently, the plugin resolves for every import, require, etc. call / statement, and it overwrites all of them — irrespective of there being an actual need to overwrite them.
By changing this to only process nodes which are matched by
pathsor explicitly tagged, we can make the process faster, more efficient, and ultimately reduce bug report volume.While the latter point is not necessarily a positive, the former two make this a very valuable change.
The reason this has not already been done is because there was no compiler API logic for determining paths matching. However, I've identified the logic and can replicate it by a couple of simple calls to the publicly exported underlying compiler matching functionality, which makes this possible.
The text was updated successfully, but these errors were encountered: