ICU-22939 Function composition and default bidi strategy #3536

catamorphism · 2025-06-28T04:15:06Z

This PR does several things:

Implements the changes to resolved values necessary to implement function composition, as described in an approved design doc.
Implements lazy/call-by-need evaluation. The previous implementation used lazy/call-by-name evaluation. Call-by-need is necessary because custom functions can potentially depend on external state and thus may return different values on different calls (for example, a function that returns the system time).
Implements the default bidi strategy as described in another approved design doc. The changes are related to the changes to resolved values, which is why these changes appear in a single PR.
Updates spec tests to the current version of the message-format-wg repo, with the exception of currency and math tests (these functions are not yet implemented).

Note 1: There are a few differences between the code and two design docs, which I detailed in an email to the icu-design list. ~~This PR is contingent on those changes being approved, but since there's a lot here to review, I wanted to go ahead and make the PR available for review.~~ As there have been no objections to the email, I'm considering the additional changes approved.

Note 2: I will be on leave from July 2-July 14 and won't see review comments during that time. I'll reply to comments when I return.

Checklist

Required: Issue filed: ICU-22939
Required: The PR title must be prefixed with a JIRA Issue number. Example: "ICU-1234 Fix xyz"
Required: Each commit message must be prefixed with a JIRA Issue number. Example: "ICU-1234 Fix xyz"
Issue accepted (done by Technical Committee after discussion)
Tests included, if applicable
API docs and/or User Guide docs changed or added, if applicable

jira-pull-request-webhook · 2025-07-01T01:12:50Z

Hooray! The files in the branch are the same across the force-push. 😃

~ Your Friendly Jira-GitHub PR Checker Bot

jira-pull-request-webhook · 2025-07-01T01:34:47Z

Notice: the branch changed across the force-push!

icu4c/source/i18n/messageformat2_function_registry.cpp is different

View Diff Across Force-Push

~ Your Friendly Jira-GitHub PR Checker Bot

srl295

a couple of starting comments

srl295 · 2025-07-22T17:03:42Z

icu4c/source/i18n/messageformat2_allocation.h

+    template<typename T>
+    inline T* create(const T& node, UErrorCode& status) {
+        if (U_FAILURE(status)) {
+            return nullptr;
+        }
+        T* result = new T(node);
+        if (result == nullptr) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+        }
+        return result;
+    }
+


I wonder if these should go into common - perhaps cmemory.h , errorcode.h or uobject.h

Could this be done in a future PR? Would be happy to open a JIRA ticket for it. This PR is already doing a lot. cmemory.h seems like the best place, or possibly a new private header, since cmemory.h is already pretty big.

icu4c/source/i18n/unicode/messageformat2.h

catamorphism · 2025-07-23T20:29:55Z

I'm guessing the Windows compilation errors might have something to do with the changes in #3521. I'll look into it.

jira-pull-request-webhook · 2025-07-23T20:38:03Z

Notice: the branch changed across the force-push!

icu4c/source/i18n/unicode/messageformat2_function_registry.h is different
icu4c/source/i18n/unicode/messageformat2.h is different

View Diff Across Force-Push

~ Your Friendly Jira-GitHub PR Checker Bot

jira-pull-request-webhook · 2025-09-09T19:19:14Z

Notice: the branch changed across the force-push!

icu4c/source/i18n/unicode/messageformat2_data_model.h is different
icu4c/source/i18n/unicode/messageformat2_formattable.h is different

View Diff Across Force-Push

~ Your Friendly Jira-GitHub PR Checker Bot

jira-pull-request-webhook · 2026-01-14T00:58:54Z

Notice: the branch changed across the force-push!

icu4c/source/i18n/unicode/messageformat2_data_model.h is different

View Diff Across Force-Push

~ Your Friendly Jira-GitHub PR Checker Bot

catamorphism · 2026-01-14T20:13:24Z

I'm not sure why the format checker is failing, as this PR doesn't change any Java files.

roubert · 2026-01-16T07:06:35Z

I'm not sure why the format checker is failing, as this PR doesn't change any Java files.

While it won't solve the mystery of why ICU4J / Format checker was being run at all for this PR, you might be able to make the error go away by rebasing on top of current main, commit 148f5ca.

mihnita · 2026-01-16T09:58:47Z

For java formatting, my PR landed, you would have to merge that into your branch
For single commit, I recommend you squash

Checking the 9 commits, there is nothing that is worth keeping separate for easier review, or better history. They are just attempts to fix problems. Happen, they are natural process, but not worth preserving.
If you don't have the right to squash from the GitHub UI, let me know and I'l do it.

Squashing will also solve the "jira ticket", which complains about the description (the squashed commit will have a description that passes the test).

And fixing all 3 complaints will also fix the general one "waiting tor required checks" and you will have a green PR.

jira-pull-request-webhook · 2026-01-19T20:39:42Z

Hooray! The files in the branch are the same across the force-push. 😃

~ Your Friendly Jira-GitHub PR Checker Bot

catamorphism · 2026-01-19T20:41:20Z

I've rebased and squashed; hoping this fixes all the CI errors 🤞🏼

…t strategy Implement the changes to resolved values necessary to implement function composition. Implement lazy/call-by-need evaluation (instead of lazy-call-by-name evaluation). Implement the default bidi strategy and APIs for controlling it. Update spec tests to those from the current version of the message-format-wg repo, except for currency and math tests (these functions are not yet implemented).

jira-pull-request-webhook · 2026-01-20T23:50:33Z

Notice: the branch changed across the force-push!

icu4c/source/i18n/messageformat2_function_registry_internal.h is different
icu4c/source/i18n/messageformat2.cpp is different
icu4c/source/i18n/unicode/messageformat2_data_model.h is different
icu4c/source/i18n/unicode/messageformat2_formattable.h is different
icu4c/source/i18n/unicode/messageformat2_function_registry.h is different
icu4c/source/i18n/unicode/messageformat2.h is different
testdata/message2/spec/u-options.json is different

View Diff Across Force-Push

~ Your Friendly Jira-GitHub PR Checker Bot

catamorphism · 2026-01-20T23:55:00Z

I've removed support for the u:locale option and force-pushed. To the best of my knowledge, this PR now reflects the draft LDML 49 spec except:

:currency, :percent, and :offset are not supported.
date/time formatting is based on an earlier version of the spec; since date/time formatting still has Draft status in the spec, I'm waiting for that to be finalized before making more changes to date/time formatting.

catamorphism force-pushed the bidi-default-strategy branch from 7373296 to 24c64c9 Compare July 1, 2025 01:12

catamorphism force-pushed the bidi-default-strategy branch from 24c64c9 to b34d0a5 Compare July 1, 2025 01:34

catamorphism changed the title ~~ICU-22939 DRAFT: Function composition and default bidi strategy~~ ICU-22939 Function composition and default bidi strategy Jul 1, 2025

catamorphism marked this pull request as ready for review July 1, 2025 20:29

catamorphism requested review from echeran, markusicu, mihnita and srl295 July 1, 2025 20:30

markusicu assigned mihnita Jul 10, 2025

srl295 reviewed Jul 22, 2025

View reviewed changes

catamorphism force-pushed the bidi-default-strategy branch from d973ccb to 87fbe3a Compare July 23, 2025 20:37

coltbergheger1997 approved these changes Aug 6, 2025

View reviewed changes

catamorphism force-pushed the bidi-default-strategy branch from 4088f35 to 267e8e3 Compare September 9, 2025 19:19

catamorphism force-pushed the bidi-default-strategy branch from 267e8e3 to df7d54a Compare January 14, 2026 00:58

mihnita requested a review from roubert January 16, 2026 00:06

catamorphism force-pushed the bidi-default-strategy branch from b244089 to 7abcb4b Compare January 19, 2026 20:39

catamorphism force-pushed the bidi-default-strategy branch from 7abcb4b to ccfe7bd Compare January 20, 2026 23:50

Uh oh!

ICU-22939 Function composition and default bidi strategy #3536

Are you sure you want to change the base?

ICU-22939 Function composition and default bidi strategy #3536

Uh oh!

Conversation

catamorphism commented Jun 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Checklist

Uh oh!

jira-pull-request-webhook bot commented Jul 1, 2025

Uh oh!

jira-pull-request-webhook bot commented Jul 1, 2025

Uh oh!

srl295 left a comment

Choose a reason for hiding this comment

Uh oh!

srl295 Jul 22, 2025

Choose a reason for hiding this comment

Uh oh!

catamorphism Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

catamorphism commented Jul 23, 2025

Uh oh!

jira-pull-request-webhook bot commented Jul 23, 2025

Uh oh!

jira-pull-request-webhook bot commented Sep 9, 2025

Uh oh!

jira-pull-request-webhook bot commented Jan 14, 2026

Uh oh!

catamorphism commented Jan 14, 2026

Uh oh!

roubert commented Jan 16, 2026

Uh oh!

mihnita commented Jan 16, 2026

Uh oh!

jira-pull-request-webhook bot commented Jan 19, 2026

Uh oh!

catamorphism commented Jan 19, 2026

Uh oh!

jira-pull-request-webhook bot commented Jan 20, 2026

Uh oh!

catamorphism commented Jan 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

catamorphism commented Jun 28, 2025 •

edited

Loading

catamorphism commented Jan 20, 2026 •

edited

Loading