Closed legobeat closed 1 year ago
To address thew question about size, when measured by yarn install --prod
: This commit is an increase of 3284 b (27368 - 24084).
The main contributor here is object.assign
clocking in at 1284b. It seems the maintainer is optimizing heavily for backwards-compatibility: https://github.com/inspect-js/node-deep-equal/pull/100
This package is using babel to address that. I think you're right @mcmire that this one isn't worth it. Thank you for catching that!
Is object.assign et al worth filling in for still? Might be worth replacing deep-equal
alltogether for something more appropriate.
Alternatively, 2.0.2
is the last version not depending on object.assign
.
(then on the other hand, in context of metamask-extension
, we already have a runtime dependency on the same version of object.assign
:shrug: )
@mcmire I resurrected this in #94, with the ambition of getting out a non-breaking maintenance release before deprecating nodejs v12, which would also unlock bringing this in line with metamask-module-template
.
Valid thing to check for sure - looking at the deep-equal changelog and version history, these do serve their role in addressing more cases (fixing several bugs, though nothing stands out as obviously affecting us here) and some performance improvements in form of optimization.
Seems worth it to me, though not super clear-cut.
FWIW, as for dev,
tape
has also upgraded this dependency in #89, so the combination of these two is a larger net-reduction in lockfile entries than it may first appear