Skip to content

Conversation

KristofferC
Copy link
Member

@KristofferC KristofferC commented Sep 17, 2025

Backported PRs:

Need manual backport:

Contains multiple commits, manual intervention needed:

Non-merged PRs with backport label:

oscardssmith and others added 7 commits September 17, 2025 12:25
Successor to #56808 which enables
an opaque closure to be constructed with a specified world age (if, for
example, the current world age isn't valid due to being in a generated
function). Required for EnzymeAD/Reactant.jl#365

(cherry picked from commit e46cc1a)
(cherry picked from commit 95e132e)
fixes #59397

(cherry picked from commit c2fd42e)
Addresses the same CI failure that
#59141 was intended to address.
The issue here is twofold. First, in a source build, we failed to
install openssl header files. Second, even if we had done this, libssh
needs to be explicitly told where to find openssl, otherwise it might
prefer systme files.

(cherry picked from commit 5ddd721)
This adds one new section (h2) and a nested subsection (h3) to the
landing page of the Julia documentation website and associated PDF file.
The new section provides a quick overview of the differences between
`Core`, `Base` and the standard library, as well as a list of links to
standard library packages.

The list of standard library packages is thus duplicated in the website
sidebar, but I don't think this is an issue because the sidebar is
primarily intended for navigation, whereas this list is primarily
intended to be used for referencing purposes.

Partially addresses #28712, although I don't provide short descriptions
of each standard library package. I think the appropriate place for that
would be in module-level docstrings of the packages themselves, which
are not currently available.

---------

Co-authored-by: Viral B. Shah <[email protected]>
(cherry picked from commit abb104c)
This is as an alternative to changing all allocation functions to
mutating all memory and returning an aliasing pointer. This operates
usually late in the pipeline, so either should not make much difference,
but I think it should suffice to mark this volatile to prevent any
de-duplication of this load, and this should also be most conservative
for GC but more liberal for other optimizations.

Fixes #59547

Produced with dubious help by Claude.

(cherry picked from commit 218f691)
)

Fixes #59569

Thanks @N5N3 for finding the bug!

For 1.13, we could re-enable this optimization and propagate the
information to the sub-alg or have the sub-alg's embedded indices be
relative to `kw.lo` rather than absolute. For backports, this minimal
change is more appropriate.

(cherry picked from commit 067b013)
@KristofferC KristofferC changed the title Make version of opaque closure constructor in world (#56823) Backports for 1.12.0-rc3 (or 1.12.0) Sep 17, 2025
@topolarity topolarity force-pushed the backports-release-1.12 branch from 6888606 to 6be73b4 Compare September 17, 2025 18:57
KristofferC and others added 2 commits September 19, 2025 10:45
…59548)

- Use svec instead of tuple for arguments (better match for ABI which
will require boxes)
- Directly forward single svec argument, both runtime and codegen,
without copying.
- Optimize all consistant builtin functions of constant arguments, not
just ones with special tfuncs. Reducing code duplication and divergence.
- Codegen for `svec()` directly, so optimizer can see each store (and
doesn't have to build the whole thing on the stack first).

Written with help by Claude
@KristofferC KristofferC merged commit a4940e7 into release-1.12 Sep 19, 2025
8 checks passed
@KristofferC KristofferC deleted the backports-release-1.12 branch September 19, 2025 13:50
@KristofferC KristofferC restored the backports-release-1.12 branch September 22, 2025 08:45
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.

9 participants