fix: follow symlinks on build-ideal-tree add #129
Conversation
| if (rawSpec !== '' || pkg[type][name] === undefined) { | ||
| // if we're in global mode, file specs are based on cwd, not arb path | ||
| pkg[type][name] = specType === 'file' || specType === 'directory' | ||
| pkg[type][name] = specType === 'file' |
There was a problem hiding this comment.
this was introduced in #125 as a way to compensate for the wrong parsed links and should not be necessary anymore
| spec.name = mani.name | ||
| return spec | ||
| }) | ||
| const currpath = this[_global] ? process.cwd() : this.path |
There was a problem hiding this comment.
@isaacs I decided to leave this here for now, we can revisit it later
There was a problem hiding this comment.
We can land on a separate commit, but ought to get it fixed, may as well pull in with this changeset.
There was a problem hiding this comment.
eh, let's land it in a sep commit, maybe even better sep PR, let's thing about future us trying to bisect something 😂
|
|
||
| spec.name = mani.name | ||
| return spec | ||
| } |
There was a problem hiding this comment.
This was just a small refactor for readability since when adding the _followSymlinkPath method (that is now below) things got gnarly in this part of the code
Symlinks were being tracked as invalid when loading actual trees since their realpath wouldn't match, given that build-ideal-tree was just preserving the fetchSpec read from npm-package-arg when adding new packages to the ideal tree. This change fixes it by properly parsing and following symlinks realpaths at the moment they're added to the ideal tree.
ruyadorno
force-pushed
the
fix-parse-symlink-follow-realpath
branch
from
0aaf65c to
5c50c0a
Compare
| /* istanbul ignore else - same as the ignored catch above */ | ||
| if (real) { | ||
| const { name } = spec | ||
| spec = npa(`file:${relpath(this.path, real)}`, this.path) |
There was a problem hiding this comment.
the new spec gets npa parsed relative to this.path, so that's what will end up in resolved, in the package.json, etc - also something I want to validate that makes sense in the impl 😊
| spec.name = mani.name | ||
| return spec | ||
| }) | ||
| const currpath = this[_global] ? process.cwd() : this.path |
There was a problem hiding this comment.
We can land on a separate commit, but ought to get it fixed, may as well pull in with this changeset.
ruyadorno
force-pushed
the
fix-parse-symlink-follow-realpath
branch
from
691f71b to
6ae40a7
Compare
ruyadorno
dismissed
isaacs’s stale review
Approved over pair session review
4 participants
Symlinks were being tracked as invalid when loading actual trees since
their realpath wouldn't match, given that build-ideal-tree was just
preserving the fetchSpec read from npm-package-arg when adding new
packages to the ideal tree.
This change fixes it by properly parsing and following symlinks
realpaths at the moment they're added to the ideal tree.