Skip to content

Comment out [email protected] because of build errors with basic gcc stack#48478

Merged
tldahlgren merged 3 commits intospack:developfrom
climbfuji:feature/remove_pyarrow_130_due_to_build_errors
Jan 10, 2025
Merged

Comment out [email protected] because of build errors with basic gcc stack#48478
tldahlgren merged 3 commits intospack:developfrom
climbfuji:feature/remove_pyarrow_130_due_to_build_errors

Conversation

@climbfuji
Copy link
Copy Markdown
Contributor

@climbfuji climbfuji commented Jan 8, 2025

Resolves #48477

I added the package py-arrow originally, but didn't add myself as maintainer. Nobody else contributed to it so far. Since nothing in spack develop as of today depends on [email protected] (i.e. 1.2.3 suffices, and it builds ok), I am commenting out the version with a reference to the issue so that the memory isn't lost.

I am also adding myself as maintainer.

@tldahlgren
Copy link
Copy Markdown
Contributor

I am able to successfully build 1.3.0 on rhel8 using develop. I attempted a --fresh build on another system with a zen3 target but it can't seem to get past py-pip.

So I am wondering if a conflict (https://spack.readthedocs.io/en/latest/packaging_guide.html#conflicts-and-requirements) would be a better approach. What do you think?

@climbfuji
Copy link
Copy Markdown
Contributor Author

I am able to successfully build 1.3.0 on rhel8 using develop. I attempted a --fresh build on another system with a zen3 target but it can't seem to get past py-pip.

So I am wondering if a conflict (https://spack.readthedocs.io/en/latest/packaging_guide.html#conflicts-and-requirements) would be a better approach. What do you think?

I was wondering, too. I could build it on my laptop (Oracle Linux 9 with gcc@13), but it failed on an HPC RHEL 8.7 with gcc@11, and on a Ubuntu 24.04 AWS instance with gcc@11. I am wondering if the gcc version is the problem, but I didn't have time to look into it. I was just kicking the can down the road I guess!

@climbfuji
Copy link
Copy Markdown
Contributor Author

I am able to successfully build 1.3.0 on rhel8 using develop. I attempted a --fresh build on another system with a zen3 target but it can't seem to get past py-pip.
So I am wondering if a conflict (https://spack.readthedocs.io/en/latest/packaging_guide.html#conflicts-and-requirements) would be a better approach. What do you think?

I was wondering, too. I could build it on my laptop (Oracle Linux 9 with gcc@13), but it failed on an HPC RHEL 8.7 with gcc@11, and on a Ubuntu 24.04 AWS instance with gcc@11. I am wondering if the gcc version is the problem, but I didn't have time to look into it. I was just kicking the can down the road I guess!

It's not the compiler (would have been surprising anyway), since I can build 1.3.0 with [email protected] on my Oracle Linux 9 system. Since you can build on rhel8, but the build fails on the HPC with RHEL8.7, it's not a RH-based v8 vs RH-based v9 either. Hmm. Otherwise, the stacks have the same requirements across my machines, only the externals differ.

@climbfuji
Copy link
Copy Markdown
Contributor Author

@tldahlgren I can't figure out within a reasonable time what causes the build error and what allows it to build. Do you want me to leave the version in and add a blanket conflict (i.e. without when=)? Not sure if that's better, though.

@tldahlgren
Copy link
Copy Markdown
Contributor

tldahlgren commented Jan 10, 2025

@tldahlgren I can't figure out within a reasonable time what causes the build error and what allows it to build. Do you want me to leave the version in and add a blanket conflict (i.e. without when=)? Not sure if that's better, though.

Fair point under the circumstances. I've experimented with a couple of options specifying the complete architecture as the first argument and "@1.3.0" for when to trigger a failure on my system but Spack just builds with alternate target(s). I've tried with the first argument being @1.3.0 and the second the complete architecture I was hoping to result in a conflict but Spack builds with a different target(s).

If you want to simply comment out the version, perhaps also comment out lines 32, 36, and 40 with comment(s) about uncommenting once resolve the issues?

@climbfuji
Copy link
Copy Markdown
Contributor Author

@tldahlgren I can't figure out within a reasonable time what causes the build error and what allows it to build. Do you want me to leave the version in and add a blanket conflict (i.e. without when=)? Not sure if that's better, though.

Fair point under the circumstances. I've experimented with a couple of options specifying the complete architecture as the first argument and "@1.3.0" for when to trigger a failure on my system but Spack just builds with alternate target(s). I've tried with the first argument being @1.3.0 and the second the complete architecture I was hoping to result in a conflict but Spack builds with a different target(s).

If you want to simply comment out the version, perhaps also comment out lines 32, 36, and 40 with comment(s) about uncommenting once resolve the issues?

Yes, good point. Done in 675b0c7

@tldahlgren tldahlgren enabled auto-merge (squash) January 10, 2025 01:03
@tldahlgren tldahlgren disabled auto-merge January 10, 2025 01:04
@tldahlgren tldahlgren enabled auto-merge (squash) January 10, 2025 01:04
@tldahlgren tldahlgren merged commit 4643909 into spack:develop Jan 10, 2025
@climbfuji
Copy link
Copy Markdown
Contributor Author

Thanks!

@climbfuji climbfuji deleted the feature/remove_pyarrow_130_due_to_build_errors branch January 10, 2025 02:45
DingXuefeng pushed a commit to DingXuefeng/spack that referenced this pull request Jan 12, 2025
…ck (spack#48478)

* Comment out [email protected] because of build errors with basic gcc stack
* Also comment out 1.3.0-specific depends_on lines
climbfuji added a commit to climbfuji/spack that referenced this pull request Jan 13, 2025
…ck (spack#48478)

* Comment out [email protected] because of build errors with basic gcc stack
* Also comment out 1.3.0-specific depends_on lines
mtaillefumier pushed a commit to mtaillefumier/spack that referenced this pull request Jan 20, 2025
…ck (spack#48478)

* Comment out [email protected] because of build errors with basic gcc stack
* Also comment out 1.3.0-specific depends_on lines
teaguesterling pushed a commit to teaguesterling/spack that referenced this pull request Feb 5, 2025
…ck (spack#48478)

* Comment out [email protected] because of build errors with basic gcc stack
* Also comment out 1.3.0-specific depends_on lines
mrmundt pushed a commit to mrmundt/spack that referenced this pull request Feb 17, 2025
…ck (spack#48478)

* Comment out [email protected] because of build errors with basic gcc stack
* Also comment out 1.3.0-specific depends_on lines
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Installation issue: [email protected] (1.2.2 and 1.2.3 build ok)

2 participants