Closed rhiadj closed 6 years ago
Thanks @rhiadj, very happy to include this π I'll pick it up when I next time I update shrinkpack.
Brilliant! Thanks @JamieMason - much appreciated π
Sorry for the wait, this will make it in eventually :+1:
When finished, the new version that supports npm 5 should account for this: https://github.com/JamieMason/shrinkpack/blob/dev/src/lib/get-unused-tars.js#L6. Anything that isn't a tar or tgz will be ignored.
Great stuff - thanks for the update @JamieMason
Don't worry too much re the wait - I've been using a cloned/patched version from my repo for now but will def move back to your mainline release once this is in.
As always - much appreciated and keep up the fantastic work! This tool has saved me countless hours already!
Because #83 is blocked I thought I'd look at releasing this and some other small changes while we wait.
Unfortunately though the build is having a strange issue running npm install -g npm@4.4.3
in Travis and I'm now on holiday for a week. I will pick this back up when I return, thanks.
@JamieMason no worries at all - as it happens I've jumped ship to yarn now (which has similar functionality built in) so this isn't as important as it was. Obviously feel free to continue this as is/was. I'm very grateful for all your help - best of luck for the future!
As part of a project I'm working on we'd like to use shrinkpack - however we'd like to use a git submodule in order to store the node packages in a separate repository.
This has several advantages - the main two being:
With this in mind we set up the repo as follows:
However we can't do this currently as whenever we run
shrinkpack
it removes the.git
andREADME.md
(and any other file) insidenode_shrinkwrap
as part of the deletion logic:In order to work around this I've tweaked the bundle logic slightly so that it only considers files ending in
.tar
and.tgz
during various operations.Interested to get your thoughts on this?
Notes: I tried to run the tests (on a vanilla clone) but got cryptic errors so wasn't able to add any additional tests for this new logic - or indeed check if any were failing (sorry about this I did try!).
Thanks