fix(compiler) restore old copy behaviour, with new ignore option #6020
Conversation
|
we encounter copy problems with |
|
Hey, is there any update or traction on this? This is currently acting as a blocker from us updating. |
|
@George-Payne it seems like the copy task test I added is failing, mind checking? |
George-Payne
force-pushed
the
old-copy-behaviour
branch
from
5c3b35f
to
d21408f
Compare
@christian-bromann Didn't realise there were multiple test commands, so missed that one. I moved it to use a glob and amended the commit. |
|
Hey, is there anything blocking this from being added? We're very excited about this fix to make it in 😃 |
|
@George-Payne thanks for the contribution, we will release this next week on Monday. |
|
Hi @christian-bromann sorry to bother, there is any news on the release of this? |
|
This patch has been released in https://github.com/ionic-team/stencil/releases/tag/v4.22.3, sorry for the delay! |
|
@christian-bromann in the v4.22.3 changelog isn't row of this merge, also in the commits section of this tag. edit: |
What is the current behavior?
#5899 introduced a lot of regressions:
A breaking change in copy behavior: #5983
file to file copying broken: #5956
EBUSY: resource busy: #5966
Directories from above the root dir are copied outside of the output dir.
What is the new behavior?
This PR restores the old copy behavior, but uses the passed filter list, rather than the hardcoded one.
This isn't an optimum fix, as it changes the copy back to
dirtodir(orfiletofile), so ignore patterns aren't matched against child files. Only glob matches are ignored (as was the previous behavior). IMO it's better than the large number of regressions, especially as this can now be worked around (e.g. in the case of #5781) by the user moving to a glob copy, with the new ignore option.Documentation
none
Does this introduce a breaking change?
Testing
Tested on own repo.
Other information
fixes: #5983
fixes: #5956
fixes: #5966