release-1.6: Backports for 1.6.7 #45319
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Discovered while running the syntax tests with lines 25-28 of `jlfrontend.scm` commented out. Turned out we didn't handle `.op` correctly in neither `check-dotop` nor in `deparse`. This meant we just got `error: malformed expression` instead of an actually useful error message. (cherry picked from commit 9112135)
(cherry picked from commit c589e0d)
This prior conditional was a fixed constant branch, so this seems more like the intent. (cherry picked from commit ac51add)
(cherry picked from commit b9d8280)
If multiple processes attempt to clean up cache files at the same time, a race condition can result, e.g. https://buildkite.com/clima/climaatmos-ci/builds/812#6a961e99-e2a1-488b-a116-2a45dee26d38/102-104 (cherry picked from commit d4acead)
KristofferC
changed the title
Co-authored-by: Jeff Bezanson <jeff.bezanson@gmail.com>
|
f31ef76 should also be backported |
There's no performance change, if the `indices`s of `Rdest` and `Rsrc` are all `NTuple{N,<:AbstractUnitRange}`.
(cherry picked from commit a91be39)
(cherry picked from commit eed2dba)
Otherwise, overrides do not trigger when using `artifact"..."` inside a submodule. (cherry picked from commit 9b106ad)
|
@fxcoudert, could you take this branch for a spin to make sure it works ok for your use case? |
|
@KristofferC I won't have time cause I'm travelling right now (and I've never used the 1.6 branch so far). Sorry. But I appreciate the heads-up. |
* Apply patch for GMP CVE-2021-43618 * Update checksums (cherry picked from commit dea9805)
(cherry picked from commit 3d6731b)
(cherry picked from commit 6be86a3)
Note: this is the first build of the real upstream version 1.2.12 which was released a few days ago. (cherry picked from commit 81e7cfc)
(cherry picked from commit 1c10a9e)
Acquiring this lock in many implementations could result in deadlock, even with an existing reader. Add a TLS check for reentry before, instead of relying on the implementation specifics, to avoid any issues. (cherry picked from commit 7df454b)
KristofferC
force-pushed
the
backports-release-1.6
branch
from
6e9251a to
b6eda9d
Compare
KristofferC
force-pushed
the
backports-release-1.6
branch
2 times, most recently
from
789dcc6 to
9561ec5
Compare
|
@KristofferC please don't forget #39369 🙂 |
|
@nanosoldier |
|
Your package evaluation job has completed - possible new issues were detected. A full report can be found here. |
|
It would be nice but not critical to add JuliaLang/Pkg.jl#3073 |
…sue, since 12.1 Later, we should probably switch to using mach_exc_server generated from `mig mach_exc.defs`. (cherry picked from commit 267b124)
* avoid using `@sync_add` on remotecalls It seems like @sync_add adds the Futures to a queue (Channel) for @sync, which in turn calls wait() for all the futures synchronously. Not only that is slightly detrimental for network operations (latencies add up), but in case of Distributed the call to wait() may actually cause some compilation on remote processes, which is also wait()ed for. In result, some operations took a great amount of "serial" processing time if executed on many workers at once. For me, this closes #44645. The major change can be illustrated as follows: First add some workers: ``` using Distributed addprocs(10) ``` and then trigger something that, for example, causes package imports on the workers: ``` using SomeTinyPackage ``` In my case (importing UnicodePlots on 10 workers), this improves the loading time over 10 workers from ~11s to ~5.5s. This is a far bigger issue when worker count gets high. The time of the processing on each worker is usually around 0.3s, so triggering this problem even on a relatively small cluster (64 workers) causes a really annoying delay, and running `@everywhere` for the first time on reasonable clusters (I tested with 1024 workers, see #44645) usually takes more than 5 minutes. Which sucks. Anyway, on 64 workers this reduces the "first import" time from ~30s to ~6s, and on 1024 workers this seems to reduce the time from over 5 minutes (I didn't bother to measure that precisely now, sorry) to ~11s. Related issues: - Probably fixes #39291. - #42156 is a kinda complementary -- it removes the most painful source of slowness (the 0.3s precompilation on the workers), but the fact that the wait()ing is serial remains a problem if the network latencies are high. May help with #38931 Co-authored-by: Valentin Churavy <vchuravy@users.noreply.github.com> (cherry picked from commit 62e0729)
* Fix USE_BINARYBUILDER_OPENBLAS=0 on macOS We didn't properly move the OpenBLAS -> Objconv dependency chain over to manifest files. This was not visible until now because we rarely build OpenBLAS from source without also building Objconv from source. Fixes #42519 * Also update path to `objconv` binary (cherry picked from commit 6af7584)
(cherry picked from commit 434d340)
Co-authored-by: KristofferC <kristoffer.carlsson@juliacomputing.com> (cherry picked from commit 86f5501)
Distributed closes and destroys stdin, but some tests attempted to explicitly use it, leading to test problems. We previously interpreted this as passing devnull, but this is better to be explicit. (cherry picked from commit 64a86f9)
This should fix the `Exec format error` problems we've had recently due to us adding some non-x86_64 linux workers to the pool.
(cherry picked from commit 4f1c68e)
(cherry picked from commit 9b83dd8)
* [release-1.6] Backport MbedTLS security patches A few MbedTLS security patches should be backported to our LTS release. In the future, we should lock ourselves to MbedTLS LTS releases so that we can take advantage of their backporting as well. * Bump mbedtls version and checksums
KristofferC
force-pushed
the
backports-release-1.6
branch
from
07de466 to
443c321
Compare
DilumAluthge
removed
the
merge me
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Backported PRs:
base/version_git.sh: fix the default value forBase.GIT_VERSION_INFO(for e.g. when Git is not available) #43717init.defaultBranch#44629round(Integer, big(Inf))#44676.opin lowering #44770copyto!(dest, Rdest::CartesianIndices, src, Rsrc::CartesianIndices)with different step. #45289iterate(::Dict)#44534@sync_addon remotecalls #44671===when encountering null pointer #44749Need manual backport:
jl_setjmpon windows #42267Contains multiple commits, manual intervention needed:
pinv#45009Non-merged PRs with backport label:
transposeinLinearAlgebra.copy_transpose!to handle matmul of isbits array-elements with non-identity transpose #42715