[flang-commits] [flang] [flang] Add Flang Community Call Notes for July 29th meeting (PR #212810)

Wed Jul 29 09:10:20 PDT 2026

https://github.com/AlexisPerry created https://github.com/llvm/llvm-project/pull/212810

None

>From bf27b9181397dfa41250e992bc8ba1d0d7e7818b Mon Sep 17 00:00:00 2001
From: AlexisPerry <aperry at lanl.gov>
Date: Wed, 29 Jul 2026 10:09:19 -0600
Subject: [PATCH] [flang] Add Flang Community Call Notes for July 29th meeting

---
 flang/docs/MeetingNotes/2026/2026-07-29.md | 71 ++++++++++++++++++++++
 1 file changed, 71 insertions(+)
 create mode 100644 flang/docs/MeetingNotes/2026/2026-07-29.md

diff --git a/flang/docs/MeetingNotes/2026/2026-07-29.md b/flang/docs/MeetingNotes/2026/2026-07-29.md
new file mode 100644
index 0000000000000..d552237752b7d
--- /dev/null
+++ b/flang/docs/MeetingNotes/2026/2026-07-29.md
@@ -0,0 +1,71 @@
+<!--===- docs/MeetingNotes/2026/2026-07-29.md
+
+   Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+   See https://llvm.org/LICENSE.txt for license information.
+   SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+-->
+# Community Call 2026-07-29
+
+## Agenda
+
+* Design docs and/or RFCs  
+  * [\[RFC\]\[Flang\] Improving temporary and local array allocations \- unify StackArray/MemoryAllocation](https://discourse.llvm.org/t/rfc-flang-improving-temporary-and-local-array-allocations-unify-stackarray-memoryallocation/91353)	  
+    * Right now we have two passes to clean up heap vs. stack placement decisions for arrays.  Most other compilers have a more unified approach. This also relates to our default behavior of \-fno-stack-arrays  
+    * The first 3 PRs in the stack have been merged.  
+    * Need more comments from the community regarding a possible change in default behavior for automatic arrays  
+  * [\[RFC\] add support for collapsed io calls for implicit do loops](https://discourse.llvm.org/t/rfc-add-support-for-collapsed-io-calls-for-implicit-do-loops/91416)  
+    * Related PR: [https://github.com/llvm/llvm-project/pull/212646](https://github.com/llvm/llvm-project/pull/212646)   
+    * (Potentially large) performance enhancement if the existing multiple calls can be collapsed into one.  There is a question of the legality of this transformation.  
+  * [\[RFC\]\[flang\] Add Floating-point trap handling support](https://discourse.llvm.org/t/rfc-flang-add-floating-point-trap-handling-support/90544)   
+    * Github issue: [https://github.com/llvm/llvm-project/issues/198657](https://github.com/llvm/llvm-project/issues/198657)   
+    * Related PR: [https://github.com/llvm/llvm-project/pull/208828](https://github.com/llvm/llvm-project/pull/208828)   
+      * Need reviewers for the non-frontend portions  
+    * The Flang documentation should have some note about the IEEE support for this option, but that is future work.   
+  * [Changes to builtin modules](https://discourse.llvm.org/t/changes-to-builtin-modules/89072)   
+    * Has [broken apt.llvm.org packaging](https://github.com/llvm/llvm-project/issues/201304)  
+      * Needs to be fixed by the maintainer, Michael Kruse has created a [pull request](https://salsa.debian.org/pkg-llvm-team/llvm-toolchain/-/merge_requests/227) that was merged  
+      * they are currently also [otherwise broken](https://github.com/llvm/llvm-project/issues/209253)  
+    * Does this need to remain on the agenda? Are more potentially build-breaking changes coming? (last meeting we said it can be removed, so we will remove going forward)  
+    * Michael Kruse has created a merge request to fix the problem that has been approved today  
+  * [Cross-compilation of real(kind=16)](https://discourse.llvm.org/t/cross-compilation-of-real-kind-16/89161/38)  
+    * Discussion is ongoing on the RFC  
+  * [\[RFC\] The Cost of Templates when Building Flang](https://discourse.llvm.org/t/rfc-the-cost-of-templates-when-building-flang/91197)  
+    * Related PR in draft state: [https://github.com/llvm/llvm-project/pull/206907](https://github.com/llvm/llvm-project/pull/206907)  
+      * Touches a lot of files, but does not have any merge conflicts at the moment (is largely made with AI)  
+    * Michael Kruse is working to make the PR reviewable (as it is so large, and fixing up issues with the AI generated code)  
+    * A first non-draft patch will be available soon  
+    * This work has potentially huge compile-time performance impact  
+  * [Choosing the default summation algorithm for Flang's SUM intrinsic](https://discourse.llvm.org/t/choosing-the-default-summation-algorithm-for-flangs-sum-intrinsic/91300)  
+    * The first implementation of Kahan summation did not go as planned, RFC comment will be coming soon  
+  * [\[RFC\] Support for \-std=f2023](https://discourse.llvm.org/t/rfc-support-for-std-f2023/89608)  
+    * New PR: [https://github.com/llvm/llvm-project/pull/205938](https://github.com/llvm/llvm-project/pull/205938)   
+    * Related Issue: [https://github.com/llvm/llvm-project/issues/196409](https://github.com/llvm/llvm-project/issues/196409)   
+    * **Reviews needed**, as the PR has been up for a few weeks already and there appears to not yet be consensus on the Discourse post  
+    * Current sticking points: concerns were expressed in the Discourse post, Chris Earl worked to address them in the new PR, but he has not had feedback either way since  
+* PRs of Note  
+  * *Reviewers needed:* [https://github.com/llvm/llvm-project/pull/203030](https://github.com/llvm/llvm-project/pull/203030)  
+    * This PR is part 2 of 3 for a stack PR implementing rank-1 integer arrays with constant extent being used as explicit shape bounds.   
+    * PR 3 of 3: [https://github.com/ivanrodriguez3753/llvm-project/pull/2](https://github.com/ivanrodriguez3753/llvm-project/pull/2)  
+    * Associated llvm-test-suite PR: [https://github.com/llvm/llvm-test-suite/pull/411](https://github.com/llvm/llvm-test-suite/pull/411)   
+  * *Needs input:* Avoid full source path in remarks: [https://github.com/llvm/llvm-project/pull/212669](https://github.com/llvm/llvm-project/pull/212669)   
+    * Inconsistencies between Flang and Clang, are these intentional or coincidental and should be corrected  
+    * Tarun: we should be consistent in with Clang on all fronts with respect to file paths in remarks  
+    * Vineet: for now, but the behavior change behind a flag and then later we can change the default and remove it.  Would like to avoid breaking user scripts.  
+      * Tarun: change the default now, but provide a way to get the old behavior if users need it, be sure to put something in the Release Notes about the change  
+    * Please provide additional comments on the related Discourse post: [https://discourse.llvm.org/t/flang-should-flang-remarks-include-the-full-path-to-the-source-file/91422](https://discourse.llvm.org/t/flang-should-flang-remarks-include-the-full-path-to-the-source-file/91422) 
+
+* FYI   
+  * LLVM 23.1.0-rc1 and 23.1.0-rc2 Released\!  
+    * [https://discourse.llvm.org/t/llvm-23-1-0-rc1-released/91369](https://discourse.llvm.org/t/llvm-23-1-0-rc1-released/91369)  
+    * [https://discourse.llvm.org/t/llvm-23-1-0-rc2-released/91409](https://discourse.llvm.org/t/llvm-23-1-0-rc2-released/91409)   
+    * rc3 will be released on August 11th  
+  * OpenMP 6.1 Public Comment Draft: [https://www.openmp.org/wp-content/uploads/openmp-TR15.pdf](https://www.openmp.org/wp-content/uploads/openmp-TR15.pdf)  
+    * Email [info at openmp.org](mailto:info at openmp.org) or [michael.klemm at openmp.org](mailto:michael.klemm at openmp.org) with any feedback  
+    * This will be up for comment until the end of September  
+* Other topics as time allows  
+
+## Details
+
+* Consists of over **792,000** lines of code, documentation, build files, and test  
+* To date, over **12,807** commits have been made to Flang

