Move set_blosc! methods to H5ZBlosc.jl
#1167
Conversation
As far as I can tell, this usage of Requires was entirely unnecessary since the package it was "extending" already depends on HDF5.jl Co-authored-by: Gabriel Baraldi <[email protected]>
topolarity
force-pushed
the
ct/mv-blosc
branch
2 times, most recently
from
069d059
to
60d7775
Compare
This is needed to pick up Documenter 1.4.0 which supports fully-qualified `@ref` links.
topolarity
force-pushed
the
ct/mv-blosc
branch
from
60d7775
to
d91501e
Compare
topolarity
force-pushed
the
ct/mv-blosc
branch
from
8f02318
to
330ee06
Compare
|
I need to think about the interaction with #1160. These methods were mainly there was backwards compatibility. |
|
For what it's worth, the methods are still there - I think this should behave pretty much as it did before, except that because we can't set retroactive |
|
@simonbyrne @musm any thoughts about this pull request? Does this require a version bump? |
Project.toml
Outdated
| @@ -22,6 +22,7 @@ MPIExt = "MPI" | |||
|
|
|||
| [compat] | |||
| Compat = "3.1.0, 4" | |||
| H5Zblosc = "≥ 0.1.3" | |||
There was a problem hiding this comment.
Shouldn't this be covered by the compat in H5Zblosc?
There was a problem hiding this comment.
Good point - it should be bumped / tightened whenever this falls into a release, but doing the compat on the H5Zblosc side is probably right
There's already an HDF5 compat bound in H5Zblosc - that one just needs to be kept up-to-date.
topolarity
force-pushed
the
ct/mv-blosc
branch
from
a3f5379
to
59d6e46
Compare
That's the last one gone!
As far as I can tell, Requires was not necessary here since the package it was "extending" already depends on HDF5.jl