Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Accommodate AbstractFFTs PR 26 #322

Merged
merged 1 commit into from
Nov 10, 2019
Merged

Accommodate AbstractFFTs PR 26 #322

merged 1 commit into from
Nov 10, 2019

Conversation

ararslan
Copy link
Member

@ararslan ararslan commented Nov 10, 2019

Bound FFTW to 1.1 to avoid conflicting export of Frequencies et al., set verion to 0.6.1. Carsten's authorship has been preserved.

Fixes #320
Closes #300
Closes #321

Bound FFTW to 1.1 to avoid conflicting export of Frequencies et al., set
verion to 0.6.1.
@ararslan ararslan merged commit ce1e15c into master Nov 10, 2019
@ararslan ararslan deleted the aa/cb/fftfreq branch November 10, 2019 00:38
@galenlynch
Copy link
Member

Thank you for figuring all of this stuff out!

galenlynch added a commit to galenlynch/General that referenced this pull request Nov 23, 2019
This is another attempt at JuliaRegistries#5230, which was an attempt at fixing the breakage
caused by JuliaMath/AbstractFFTs.jl#26 (e.g. JuliaDSP/DSP.jl#320,
JuliaDSP/DSP.jl#323, JuliaDSP/DSP.jl#324). The underlying problem is that
JuliaMath/AbstractFFTs.jl#26 moved functions from DSP.jl to AbstractFFTs.jl,
requiring users to have versions of both of those packages either before that
change or after it. Because FFTW reexports AbstractFFTWs, users also have to
coordinate versions of DSP.jl and FFTW.jl.

This problem was fixed for the most recent versions of DSP.jl and
FFTW.jl (v0.6.1 and v1.1, respectively) by JuliaDSP/DSP.jl#322 and
JuliaMath/FFTW.jl#124. However, because earlier versions of FFTW do not have an
upper bound for its dependency on AbstractFFTs, and neither does DSP, Pkg cannot
currently tell that some combinations of versions for AsbtractFFTs and DSP.jl
are incompatible, which is causing users to experience issues even after the
problem was fixed for the most recent versions (JuliaDSP/DSP.jl#323,
JuliaDSP/DSP.jl#324).

Fixing the problem is made slightly more complicated because DSP does not
explicitly depend on AbstractFFTs, but instead DSP implicitly depends on
AbstractFFTs through FFTW's reexport of it, and in turn DSP explicitly depends
on FFTW.

To fix this problem, I have changed the Compat.toml for FFTW to place a upper
bound on AbstractFFTs, so that packages get a consistent interface when using
FFTW. I have also changed the Compat.toml for DSP, to avoid breakage caused by
AbstractFFTs v0.5 taking some functions that were previously defined in DSP.

In addition, the Compat.toml for DSP showed versions before 0.5.1 to be
compatible with Julia 1, which was inaccurate. I have therefore changed this
TOML table to restrict the DSP versions to the range that is compatible with
Julia 1.

I have tested that these changes work for the most recent versions of DSP and
FFTW, as well as older versions.
ararslan pushed a commit to JuliaRegistries/General that referenced this pull request Nov 25, 2019
This is another attempt at #5230, which was an attempt at fixing the breakage
caused by JuliaMath/AbstractFFTs.jl#26 (e.g. JuliaDSP/DSP.jl#320,
JuliaDSP/DSP.jl#323, JuliaDSP/DSP.jl#324). The underlying problem is that
JuliaMath/AbstractFFTs.jl#26 moved functions from DSP.jl to AbstractFFTs.jl,
requiring users to have versions of both of those packages either before that
change or after it. Because FFTW reexports AbstractFFTWs, users also have to
coordinate versions of DSP.jl and FFTW.jl.

This problem was fixed for the most recent versions of DSP.jl and
FFTW.jl (v0.6.1 and v1.1, respectively) by JuliaDSP/DSP.jl#322 and
JuliaMath/FFTW.jl#124. However, because earlier versions of FFTW do not have an
upper bound for its dependency on AbstractFFTs, and neither does DSP, Pkg cannot
currently tell that some combinations of versions for AsbtractFFTs and DSP.jl
are incompatible, which is causing users to experience issues even after the
problem was fixed for the most recent versions (JuliaDSP/DSP.jl#323,
JuliaDSP/DSP.jl#324).

Fixing the problem is made slightly more complicated because DSP does not
explicitly depend on AbstractFFTs, but instead DSP implicitly depends on
AbstractFFTs through FFTW's reexport of it, and in turn DSP explicitly depends
on FFTW.

To fix this problem, I have changed the Compat.toml for FFTW to place a upper
bound on AbstractFFTs, so that packages get a consistent interface when using
FFTW. I have also changed the Compat.toml for DSP, to avoid breakage caused by
AbstractFFTs v0.5 taking some functions that were previously defined in DSP.

In addition, the Compat.toml for DSP showed versions before 0.5.1 to be
compatible with Julia 1, which was inaccurate. I have therefore changed this
TOML table to restrict the DSP versions to the range that is compatible with
Julia 1.

I have tested that these changes work for the most recent versions of DSP and
FFTW, as well as older versions.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

FFTW and DSP collision over Periodograms
3 participants