# Combined Call 2026-04-08 ## Agenda * Design docs and/or RFCs * [\[RFC\] Helping the new contributors to create better contributions](https://discourse.llvm.org/t/rfc-helping-the-new-contributors-to-create-better-contributions/90213) * No discussion at this time * [\[RFC\] Support for More Than 15 Array Dimensions in Flang](https://discourse.llvm.org/t/rfc-support-for-more-than-15-array-dimensions-in-flang/90423) * It does not appear that there is a real-world use case for this functionality at this time that would justify the potential downsides. * [\[RFC\] Alignment of global arrays](https://discourse.llvm.org/t/rfc-alignment-of-global-arrays/90397) * Related issue: [\[Flang\] Bad alignment of globals in modules · Issue \#189069](https://github.com/llvm/llvm-project/issues/189069) * Related PR: [https://github.com/llvm/llvm-project/pull/189709](https://github.com/llvm/llvm-project/pull/189709) (likely to be merged soon?) * Should there be a command-line flag for this? Could specify different alignments this way (ex. 32 or 64\) * Should it be determined by the user providing an architecture flag? * We don’t want to set a default alignment that conflicts with the architecture * How to handle the case of conflicting command line args? * Consider library compatibility * [\[RFC\] Flang lowering for OpenMP metadirective](https://discourse.llvm.org/t/rfc-flang-lowering-for-openmp-metadirective/90338) * No discussion at this time * [\[RFC\] preserve logical operations in the IR for reductions](https://discourse.llvm.org/t/rfc-flang-preserve-logical-operations-in-the-ir-for-reductions/90470) * No discussion at this time * [\[RFC\] Fortran 2023 ENUMERATION TYPEs](https://discourse.llvm.org/t/rfc-fortran-2023-enumeration-types/90255) * Related to Issue [\#183789](https://github.com/llvm/llvm-project/issues/183789) * Had initial prototype, refactoring now after RFC discussion and PR feedback * PRs (\~10) should start coming in later this week * Creating a new document "OpenMP-extensions"? * [https://github.com/llvm/llvm-project/pull/186696\#discussion\_r2939508874](https://github.com/llvm/llvm-project/pull/186696#discussion_r2939508874) * Has also been discussed in the OpenMP call, general consensus is to move forward with this * Will this include “not yet implemented” features? * [\[OpenMP\] Misspelling of \-Wopen-mp](https://discourse.llvm.org/t/openmp-misspelling-of-wopen-mp/90196) * PR: [https://github.com/llvm/llvm-project/pull/188434](https://github.com/llvm/llvm-project/pull/188434) * There do not seem to be objections, but what is the deprecation period? * 2 releases, is it enough? Yes * [\[RFC\] Adding conditional expressions in Flang (F2023)](https://discourse.llvm.org/t/rfc-adding-conditional-expressions-in-flang-f2023/89869) * Has been split into two PRs * Parsing/Semantics: [https://github.com/llvm/llvm-project/pull/186489](https://github.com/llvm/llvm-project/pull/186489) **(has 4 approvals, needs merge)** * Lowering: [https://github.com/llvm/llvm-project/pull/186490](https://github.com/llvm/llvm-project/pull/186490) (has 2 approvals, awaiting 1 review) * Tests added to test suite: [https://github.com/llvm/llvm-test-suite/pull/369](https://github.com/llvm/llvm-test-suite/pull/369) * [\[Flang\]\[Affine\] Linearized array access in \-promote-to-affine causes false dependence and blocks loop tiling](https://discourse.llvm.org/t/flang-affine-linearized-array-access-in-promote-to-affine-causes-false-dependence-and-blocks-loop-tiling/89927) * Several paths forward have been suggested, **please comment on the post to help guide this new contributor** * [Changes to builtin modules](https://discourse.llvm.org/t/changes-to-builtin-modules/89072) * Community could help by commenting and highlighting how important this change is. Would relieve a lot of annoyance and this is holding up Flang community progress. * Initial PR has been merged. Additional PR: [https://github.com/llvm/llvm-project/pull/171610](https://github.com/llvm/llvm-project/pull/171610) * Main PR (from which the other two were extracted): [https://github.com/llvm/llvm-project/pull/171515](https://github.com/llvm/llvm-project/pull/171515) * Still needs some working with reviewers, will contact directly at EuroLLVM next week * PRs of Note * Support for rank-1 integer array expressions in declarations (assumed-shape and explicit-shape) and allocate statements * Has been broken up, reviewers needed. * These should now be self-contained parser support PRs. Short design description is given in the GitHub Issue, [https://github.com/llvm/llvm-project/issues/178071](https://github.com/llvm/llvm-project/issues/178071) * [https://github.com/llvm/llvm-project/pull/188445](https://github.com/llvm/llvm-project/pull/188445) * [https://github.com/llvm/llvm-project/pull/188446](https://github.com/llvm/llvm-project/pull/188446) * [https://github.com/llvm/llvm-project/pull/188447](https://github.com/llvm/llvm-project/pull/188447) * PR related to pointer assignment coming * Document RFC best practices \- [PR \#190836](https://github.com/llvm/llvm-project/pull/190836) * \[flang\]\[OpenMP\] Support user-defined declare reduction with derived types [https://github.com/llvm/llvm-project/pull/190288](https://github.com/llvm/llvm-project/pull/190288) \- could use AMDGPU code reviews * Michael Klemm asked to be tagged to review * Issues of Note * None at this time * FYI * [LLVM 22.1.3 Released\!](https://discourse.llvm.org/t/llvm-22-1-3-released/90467) * Other topics as time allows * Next J3 meeting is scheduled for the week of June 15th in Coventry, UK * Best practice for PR that is dependent on someone else’s (unmerged) PR? * Could make it a part of a PR stack in some cases (LLVM user branches) * Otherwise, suggest making a draft PR and linking to the dependent one * Peter Klausler has retired. Thank you for all your hard work getting Flang started and ensuring it’s quality\!\! * Reviews that would normally have gone to him should be directed to either Eugene Epshteyn or Jean Perier * AMD and NVIDIA are hiring for Fortran compiler people ## Details * Consists of over **744,000** lines of code, documentation, build files, and test * To date, over **12,082** commits have been made to Flang