Conversation
…flow, mark pkg-config as not buildable
…2d2, ewok. Pin versions for fiat and ectrans
climbfuji
force-pushed
the
feature/updates_for_rc2
branch
from
3c57673 to
f0d74e0
Compare
climbfuji
force-pushed
the
feature/updates_for_rc2
branch
from
f0d74e0 to
a25219c
Compare
climbfuji
force-pushed
the
feature/updates_for_rc2
branch
6 times, most recently
from
fd67ee4 to
9d533ac
Compare
climbfuji
force-pushed
the
feature/updates_for_rc2
branch
from
9d533ac to
c08b0f1
Compare
|
@kgerheiser CI tests all pass, finally (2934b24) - ready for merging from my side (first spack, then spack-stack). Update. Looks like my last cleanup of the macOS CI build re-introduced the curl build error. I want to ignore this for the PR, as long as all the others pass, and fix it in a next PR, which has a few minor documentation updates that are not included here. |
| gmake -j4 2>&1 | tee log.gmake | ||
| gmake install 2>&1 | tee log.install | ||
|
|
||
| Create modulefile ``/lustre/f2/pdata/esrl/gsd/spack-stack/modulefiles/qt/5.15.3`` from template ``doc/modulefile_templates/qt`` and update ``QT_PATH`` in this file. |
There was a problem hiding this comment.
This documentation will be updated and completed over the next week
|
All other CI tests passed. This is ready to merge once the spack PR was merged, the submodule pointer updated, and the |
| ## Note: This is unreliable ... | ||
| ## Speed up installation by building in parallel |
There was a problem hiding this comment.
What makes it unreliable? I used this technique myself after seeing it here and it worked fine.
There was a problem hiding this comment.
What happened a few times was that the install step was skipped, as if it ran the query if the jobs are still running faster than they appeared in the ps output. I thought this shouldn't be possible, but simply repeating the same run worked. Could probably be solved by adding a little sleep call in between?
| version: [3.0.3] | ||
| py-setuptools: | ||
| version: [57.4.0] | ||
| # Try without |
There was a problem hiding this comment.
I think we can remove that comment as well.
… configs/sites/discover/packages.yaml: there is no variant python in ecflow
configs/common/packages.yaml
Outdated
| pkgconf: | ||
| buildable: False | ||
| pkg-config: | ||
| buildable: False |
There was a problem hiding this comment.
One of them should be buildable. An external pkg-config is generally available on these HPC systems so that's why I say go for that and leave pkgconf as not buildable.
|
If we think that pkg-config is generally available, then it won’t hurt to turn building off for both. I’ll try that for rc2, and if it doesn’t work across all platforms I’ll remove the `pkg-config` entry.
… On Jun 22, 2022, at 4:05 PM, Kyle Gerheiser ***@***.***> wrote:
@kgerheiser commented on this pull request.
In configs/common/packages.yaml <#214 (comment)>:
> + pkgconf:
+ buildable: False
pkg-config:
buildable: False
One of them should be buildable. An external pkg-config is generally available on these HPC systems. So that's why I say go for that.
—
Reply to this email directly, view it on GitHub <#214 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5C2RJIVWN6OULWGJROYFDVQOE3BANCNFSM5ZDL6JWA>.
You are receiving this because you were assigned.
|
|
macOS does not have a |
Ok, so let me change it right away then. Thanks! |
2 participants
A number of updates for spack-stack-1.0.0 release candidate 2:
jedi-ufs-alltemplatecreate-spack-mirrorCI testyaksamodulefile to blacklistpy-setuptoolstcl/tk) instead oflmod/luaon AWS Ubuntu and Redhatjedi-ufs-alltemplateskylab-1.0.0templateqt5andecflowecflowFixes #162 (see https://github.com/climbfuji/spack-stack/runs/6943697862?check_suite_focus=true)
Fixes #211
Fixes #213
Fixes #123
Waiting on JCSDA/spack#106
CI tests all passed except for macos-apple-clang build (for some weird reason, the same test passed before - rerunning now without parallel spack installs). Rerunning macos-apple-clang now manually, check https://github.com/climbfuji/spack-stack/runs/6977275667?check_suite_focus=true for the progress.