-
Notifications
You must be signed in to change notification settings - Fork 279
Patch #3090 - Fix and test S3 enabled builds #3122
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
base: main
Are you sure you want to change the base?
Conversation
I am confused by this PR. For example, what is the reason you are changing the signature of this function? |
@DennisHeimbigner if you compile the current main branch in the following way you'll see the problem:
The PR is just to show that the current main is broken when enabling S3.
Feel free to close it or add to it. |
Just so that it doesn't go noticed: that both Github actions workflows on windows, |
We should be able to enable it (using internal s3 sdk) for cygwin. Not sure about mingw. |
6f8324e
to
5d4a2c7
Compare
572c200
to
3dbbbf8
Compare
This reverts commit 8a4284a.
I am seriously confused about what is going on here. |
@mannreis I'm setting up to examine and test locally; for standardization purposes, which version of the s3 sdk are you working against? |
I'm sorry for the noise. I tried to modify the GitHub actions workflows to build with the internal S3 sdk (same as I was working with) but it was a trial and error journey. That's why I squashed commits and force pushed. |
@DennisHeimbigner So I've been digging into this, and I think I have an idea of what's going on. Mannreis isn't the one changing the function signature; part of your change with #3090 seems to have made these changes. Because we weren't testing against internal S3 implementation, we failed to catch the compilation errors. That is definitely a mistake on my part, and one we will rectify. I'm also working to see what it will take to fix this, and I think I have the issue narrowed down. The main thing I'm missing now is the implementation of whatever |
I've picked the implementation of those missing functions from https://github.com/DennisHeimbigner/netcdf-c/tree/v3plug.dmh @WardF The current state of this PR is:
|
This may be an error on my part. I believe that the HDF5 ROS3 driver uses its own |
@mannreis Thank you; I'll make the changes so that autotools and cmake are consistent. I will also take a look at the cygwin test to see what we can sort out. Thanks! |
…s behavior (in support of Unidata#3122
@mannreis I see you've picked up the implementations as you mentioned, have you pushed those yet? I've reconciled the build systems and am going to look at the Cygwin stuff now, but I want to test my fixes against your latest code if possible. Thanks! |
Let me take a look and see what happens with #3094 |
I believe I have the bulk of this reconciled, and will push to this branch if all the testing works. |
…rly reconciles them.
Rather than muddy the waters with more potential problems, I've put my attempted reconciliation of this PR, alongside #3094, into PR #3149. Fingers crossed, this will end with at most still needing to fix the cygwin tests, but we'll see. Assuming that everything passes (including the new CI you wired in @mannreis thank you very much!) I will be able to merge #3149, making #3122 (this PR) and #3094 redundant. |
…autotools behavior (in support of Unidata#3122" This reverts commit 42dbe33.
After #3090 merged I cannot build the main netcdf branch with:
cmake ../ -DNETCDF_ENABLE_NCZARR=ON -DNETCDF_ENABLE_S3_INTERNAL=ON
This PR tries to fix the most of it but there are a lot of confusion around:
NC_sortenvv
,nczm_sortenvv
andNC_freeenv
vsNCZ_freeenv
@DennisHeimbigner can you please check?
Fixes #3147