Node: Landing npm 4 into node 7

Created on 28 Dec 2016  Â·  4Comments  Â·  Source: nodejs/node

If you want to support landing this please react with emojis to this post rather than posting your support in this thread, thank you


npm v4 just landed into master in 9946519fba

The CTC has discussed landing npm 4 into node 7 and so far all discussion has been positive. One of the biggest reasons is to get the deprecation warnings around prepublish into as many hands as possible to avoid confusion if npm 5 ships with node 8

The only hesitation so far has been the dropping of support of partial shrinkwraps.

This did at one point break hapi, but was fixed in https://github.com/hapijs/hapi/commit/0ca8cea7898dcbc41d59778f0bb4b9d0985d7123 (thanks for pointing that out @iarna). It is worth mentioning that this will likely break older hapi builds on npm4 if the change was not backported.

It might be a good idea to scan the registry and find out what other projects are currently shipping broken shrinkwraps so we can better estimate ecosystem brekages, /cc @ChALkeR.

Please chime in if you have any concerns, we can bring this to the ctc again next week.

/cc @nodejs/npm

npm

Most helpful comment

As stated in a previous CTC meeting, I think this will be our best shot to gauge how many people may be impacted by npm changes from core. We have like no data and this is minimal enough that I think we should probably do it, if only for the sake of gathering a sort of data.

(Although coordination with @nodejs/npm will be necessary given that reports may end up on the npm issue tracker rather than ours.)

All 4 comments

I'm :+1: after npm4 has had a chance to sit in master for a couple of weeks.

@TheAlphaNerd You mean partial shrinkwraps — the ones that don't correspond to the fully resolved dependency tree for the specified versions?

As stated in a previous CTC meeting, I think this will be our best shot to gauge how many people may be impacted by npm changes from core. We have like no data and this is minimal enough that I think we should probably do it, if only for the sake of gathering a sort of data.

(Although coordination with @nodejs/npm will be necessary given that reports may end up on the npm issue tracker rather than ours.)

Closing since this already landed in v7.4.0

Was this page helpful?
0 / 5 - 0 ratings