fix: don't write pnpm-lock.yaml if it has no changes
#6260
Conversation
kenrick95
changed the title
zkochan
changed the title
pnpm-lock.yaml if it has no changes
pkg-manager/headless/src/index.ts
Outdated
| @@ -545,7 +546,7 @@ export async function headlessInstall (opts: HeadlessOptions) { | |||
| storeDir: opts.storeDir, | |||
| virtualStoreDir, | |||
| }) | |||
| if (opts.useLockfile) { | |||
| if (opts.useLockfile && JSON.stringify(wantedLockfile) !== JSON.stringify(opts.wantedLockfileUnmodified)) { | |||
There was a problem hiding this comment.
nit: It might be worth following up with a deepEqual(...) function that early exits as soon as it sees the lockfiles are different for performance. I've seen pnpm-lock.yaml files grow to be 4MB, and serializing 2 * 4MB could add a few seconds to each pnpm install.
There was a problem hiding this comment.
I have changed it. There was actually no need in the copy.
| @@ -128,6 +129,7 @@ export async function readLockfiles ( | |||
| existsCurrentLockfile: files[1] != null, | |||
| existsWantedLockfile: files[0] != null && !isEmptyLockfile(wantedLockfile), | |||
| wantedLockfile, | |||
| wantedLockfileUnmodified: JSON.parse(JSON.stringify(wantedLockfile)), | |||
There was a problem hiding this comment.
nit: It might be worth a comment that we have to recreate this object since the consumer is expected to mutate wantedLockfile. But maybe that's fairly obvious and isn't worth commenting.
There was a problem hiding this comment.
I was also considering just to save the hash of the object and then comparing hashes. That would use less memory but probably would do more processing
There was a problem hiding this comment.
I have changed it. There was actually no need in the copy.
close #6158