DBD::Pg Async Query Ownership and Data Preservation Fixes by oetiker · Pull Request #150 · bucardo/dbdpg

oetiker · 2025-09-19T06:16:53Z

DBD::Pg Async Query Ownership and Data Preservation Fixes

Caveat! I developed this code and the tests tests using Claude code in order to fix to #105 and mojolicious/mojo#2276

Overview

This document describes the comprehensive fixes for DBD::Pg's asynchronous query handling, addressing GitHub issue #105 and related problems with PG_OLDQUERY_WAIT data loss.

Problems Addressed

1. GitHub Issue #105: Statement Destruction Cancels Unrelated Async Queries

Problem: Destroying any statement handle would cancel the active async query of another statement, even if the destroyed statement never executed an async query.

Root Cause: The dbd_st_destroy() and dbd_st_finish() functions unconditionally cleared the global imp_dbh->async_sth pointer and called handle_old_async() without checking if the statement being destroyed actually owned the async query.

Impact: Applications using multiple statement handles with async queries would experience random query cancellations when unrelated statements were destroyed or went out of scope.

2. PG_OLDQUERY_WAIT Data Loss

Problem: Using PG_OLDQUERY_WAIT to wait for a previous async query would successfully wait but would discard the query results instead of preserving them.

Root Cause: The handle_old_async() function would consume and discard results when waiting, with no mechanism to preserve them for the owning statement.

Impact: Data loss - results from async queries would be irretrievably lost when another query used PG_OLDQUERY_WAIT.

3. Missing Ownership Verification

Problem: Any statement handle could call pg_result() and retrieve another statement's async results, violating ownership semantics.

Root Cause: The pg_db_result() function did not verify that the calling statement was the one that initiated the async query.

Impact: Race conditions, incorrect data retrieval, and unpredictable behavior in multi-statement applications.

Technical Background

PostgreSQL's async query limitations:

Only ONE async query can be active per database connection at any time
DBD::Pg provides a statement-level API despite the connection-level limitation
The module tracks ownership via imp_dbh->async_sth pointing to the statement that owns the current async query

The Fixes

Fix 1: Ownership Checking in Statement Lifecycle Functions

Location: dbdimp.c - dbd_st_finish() (lines 3977-3980)

/* Only clear async_sth if THIS statement is the current async statement */
if (imp_dbh->async_sth == imp_sth) {
    imp_dbh->async_sth = NULL;
    imp_dbh->async_status = 0;
}

Location: dbdimp.c - dbd_st_destroy() (lines 4122-4127, 4194-4197)

/* Only call handle_old_async if THIS statement owns the async query */
if (imp_dbh->async_status && imp_dbh->async_sth == imp_sth) {
    handle_old_async(aTHX_ sth, imp_dbh, PG_OLDQUERY_WAIT);
}

/* Only clear async_sth if THIS statement owns it */
if (imp_dbh->async_sth == imp_sth) {
    imp_dbh->async_sth = NULL;
}

Fix 2: Auto-Retrieve Mechanism for PG_OLDQUERY_WAIT

Location: dbdimp.c - handle_old_async() (line ~5715)

Instead of discarding results, the function now:

Retrieves the pending async query results
Stores them in the owning statement's structure
Sets async_status = 100 to indicate auto-retrieved results
Preserves row count and result data for later access

Fix 3: Ownership Verification in pg_db_result()

Location: dbdimp.c - pg_db_result() (lines 5292-5397)

Added comprehensive ownership checking:

/* Determine if called from statement or database handle */
D_imp_xxh(h);
imp_sth_t *imp_sth = NULL;
if (DBIc_TYPE(imp_xxh) == DBIt_ST) {
    imp_sth = (imp_sth_t*)imp_xxh;
}

/* Check ownership BEFORE consuming the result */
if (imp_sth && imp_dbh->async_sth && imp_sth != imp_dbh->async_sth) {
    pg_error(aTHX_ h, PGRES_FATAL_ERROR,
             "pg_result() called on wrong statement handle - "
             "this statement does not own the async query\n");
    return -2;
}

Fix 4: Auto-Retrieved Result Handling

Location: dbdimp.c - pg_db_result()

Special handling for auto-retrieved results:

/* Check if this statement had auto-retrieved results from PG_OLDQUERY_WAIT */
if (imp_sth && imp_sth->async_status == STH_ASYNC_AUTORETRIEVED) {
    /* Results were auto-retrieved - return the stored row count */
    imp_sth->async_status = STH_NO_ASYNC;  /* Clear the auto-retrieved flag */
    return imp_sth->rows;
}
else if (imp_sth && imp_sth->async_status == STH_ASYNC_AUTOERROR) {
    /* This statement has auto-retrieved error result */
    /* Report the error from the stored result */
    ...
}

New Async Status Values

The implementation uses named constants for async_status values on statement handles:

enum {
    STH_ASYNC_AUTOERROR = -2,    /* PG_OLDQUERY_WAIT auto-retrieved an error result */
    STH_ASYNC_CANCELLED = -1,
    STH_NO_ASYNC,                /* 0 - No async query */
    STH_ASYNC,                   /* 1 - Active async query */
    STH_ASYNC_PREPARE,           /* 2 - Async prepare in progress */
    STH_ASYNC_AUTORETRIEVED = 100 /* PG_OLDQUERY_WAIT auto-retrieved results */
};

Test Coverage

The fix includes comprehensive test coverage in t/08async_regression.t with 7 test scenarios:

Basic ownership - Prevents wrong statements from stealing results
$dbh->pg_result ownership - Database handle can retrieve, finished statements cannot
Statement destruction - Destroying unrelated statements doesn't affect async queries
Cross-statement interference - Multiple statements don't interfere
Interleaved operations - OLDQUERY_WAIT preserves data correctly
Error attribution - Errors are properly attributed to correct statements
Data preservation - OLDQUERY_WAIT auto-retrieves and preserves data

Behavior Changes

Before the Fix

Destroying any statement could cancel another's async query
PG_OLDQUERY_WAIT would discard previous query results
Any statement could steal another's async results
Unpredictable behavior with multiple async statements

After the Fix

Only the owning statement can affect its async query
PG_OLDQUERY_WAIT auto-retrieves and preserves results
Strict ownership verification prevents result theft
Predictable, safe multi-statement async operations

Migration Notes

The fixes are backward compatible with one exception:

Code that relied on the broken behavior of stealing async results from wrong statements will now correctly fail with an error

oetiker · 2025-10-20T11:52:55Z

ping ?

turnstep · 2025-12-10T17:18:25Z

Please pull the recent async-related changes and retest.

esabol · 2025-12-12T16:40:19Z

Also, this PR has merge conflicts. Please rebase/resolve, @oetiker.

oetiker · 2025-12-13T15:30:44Z

Also, this PR has merge conflicts. Please rebase/resolve, @oetiker.

I will have to re-evaluate after the changes ...

oetiker · 2025-12-15T09:59:54Z

The PR is rebased and the entire testsuite is passing including the new testfile.
Note I added a fix for a test which was trying to connect without password even when a password was supplied.

esabol · 2025-12-15T10:01:33Z

@turnstep : Could you approve the CI workflow to run for this PR?

esabol · 2025-12-15T10:02:12Z

The PR is rebased and the entire testsuite is passing including the new testfile. Note I added a fix for a test which was trying to connect without password even when a password was supplied.

Thank you, @oetiker ! Looks promising.

esabol · 2026-02-09T20:38:45Z

Unfortunately, the CI workflow didn't run because a maintainer didn't approve it before the request expired. 😿 I think git commit --no-edit --amend; git push -f can retrigger the CI workflow, but this PR will probably need to be rebased (again 😿) anyway after PRs #156 and #155 are merged.

oetiker · 2026-02-10T07:07:39Z

@esabol well, let me know when you are of a mind to merge this, or find issues. I do have other fish to fry in the meantime. Once you are ready I will rebase. If this were my code base it would bother me extremely to know that these bugs persist.

esabol · 2026-02-10T08:23:46Z

@esabol well, let me know when you are of a mind to merge this, or find issues. I do have other fish to fry in the meantime. Once you are ready I will rebase. If this were my code base it would bother me extremely to know that these bugs persist.

To be clear, I can't merge anything, unfortunately. I'm not a maintainer. I'm just trying to help things along where I can.

turnstep

Looks good! Will review the other sections shortly

turnstep · 2026-02-10T21:19:52Z

dbdimp.c

    STH_ASYNC,
-    STH_ASYNC_PREPARE
+    STH_ASYNC_PREPARE,
+    STH_ASYNC_AUTORETRIEVED = 100 /* PG_OLDQUERY_WAIT auto-retrieved results */


Not sure why this is being set to 100 rather than leaving it to autonumber?

turnstep · 2026-02-10T21:23:40Z

dbdimp.c

Need the new check here too?

turnstep · 2026-02-10T21:23:46Z

dbdimp.c

+       statement. This prevents finishing an unrelated statement from invalidating a pending async query. */
+    if (imp_dbh->async_sth == imp_sth) {
+        imp_dbh->async_sth = NULL;
+        imp_dbh->async_status = 0;


async_status already set above - and it should use the named constant, not 0

turnstep · 2026-02-10T21:25:32Z

dbdimp.c

+    /* Ownership will be checked below for statement-based calls when needed */
+
+    /* Check if this statement had auto-retrieved results from PG_OLDQUERY_WAIT */
+    if (imp_sth && imp_sth->async_status == STH_ASYNC_AUTORETRIEVED) {


Put constants on the left for comparisons.

turnstep · 2026-02-10T21:25:58Z

dbdimp.c

+            TRC(DBILOGFP, "%sProcessing auto-retrieved results (rows: %ld)\n",
+                THEADER_slow, imp_sth->rows);
+        }
+        /* Don't return early - let the full processing happen below */


Not needed.

turnstep · 2026-02-10T21:31:56Z

dbdimp.c

+            TRACE_PQERRORMESSAGE;
+            pg_error(aTHX_ h, status, PQerrorMessage(imp_dbh->conn));
+            rows = -2;
+            imp_sth->async_status = STH_NO_ASYNC;  /* Mark as consumed */


All three branches do this, so move to a single call.

turnstep · 2026-02-10T21:34:25Z

dbdimp.c


-            if (NULL != imp_dbh->async_sth) {
+            /* Store metadata in the appropriate statement handle */
+            /* CRITICAL: Only store results if the calling statement owns the async query */


We don't need things like upper case "critical" :)

It's all critical, ha ha ha!

turnstep · 2026-02-10T21:37:35Z

dbdimp.c

+            /* Free the last_result as needed */
+            if (imp_dbh->last_result && imp_dbh->result_clearable) {
+                TRACE_PQCLEAR;
+#if DEBUG_LAST_RESULT


Let's remove this please

turnstep · 2026-02-10T21:41:18Z

dbdimp.c

+                }
+
+                /* Mark as auto-retrieved, not discarded */
+                orig_sth->async_status = STH_ASYNC_AUTORETRIEVED;  /* Special value: auto-retrieved */


Do not need comments like this.

turnstep · 2026-02-10T21:43:51Z

dbdimp.c

+
+                imp_sth_t *orig_sth = imp_dbh->async_sth;
+
+                /* Store the error result for the original statement to discover */


This comment doesn't really match the code

t/08async_regression.t

turnstep

Wordy in the wrong places, but very thorough coverage.

turnstep · 2026-02-10T21:49:00Z

t/08async_regression.t

+use strict;
+use warnings;
+use Test::More;
+use DBI;


I know we are inconsistent, but should leave this out as we want to make sure use DBD::Pg does it for us always.

turnstep · 2026-02-10T21:50:25Z

t/08async_regression.t

+use DBI;
+use DBD::Pg qw(:async);
+
+my $dsn = $ENV{DBI_DSN} || "dbi:Pg:dbname=postgres;host=localhost";


Not sure why all of this instead of connect_database() like the other tests have?

turnstep · 2026-02-10T21:51:48Z

t/08async_regression.t

+    plan skip_all => "Cannot connect to PostgreSQL: $DBI::errstr";
+}
+
+# Test 1: Basic ownership - wrong statement cannot steal results


Wayyy too wordy. The test names should be most of the documentation. See the other tests.

turnstep · 2026-02-10T21:52:35Z

t/08async_regression.t

+# BEHAVIOR:
+# - BEFORE: sth2->pg_result() would successfully steal sth1's pending results
+# - AFTER: sth2->pg_result() fails with "wrong statement handle" error
+subtest 'Basic ownership - wrong statement' => sub {


Do not need subtests - just make the test names verbose

turnstep · 2026-02-10T21:53:06Z

t/08async_regression.t

+
+    $sth1->finish;
+    $sth2->finish;
+};


I like these tests.

turnstep · 2026-02-10T21:55:25Z

t/08async_regression.t

+
+    # Also start a good query
+    my $good = $dbh->prepare(q{SELECT 42}, { pg_async => PG_ASYNC + PG_OLDQUERY_WAIT });
+    ok($good->execute, 'good executes');


Really poor test name. Compare to a random one from 04misc.t:

Method 'private_attribute_info' is available without a database handle and returns an empty hashref

turnstep · 2026-02-10T21:56:09Z

t/08async_regression.t

+    ok($sth2->execute, 'sth2 executes with WAIT');
+
+    # sth1's results should have been auto-retrieved and preserved
+    # This is the critical test - data must not be lost


Claude sure loves the word "critical", doesn't it? 😆

oetiker · 2026-02-17T13:26:09Z

Thanks for the thorough review @turnstep! I've rebased onto current master and addressed all your feedback in a single squashed commit:

C code (dbdimp.c):

Removed = 100 from STH_ASYNC_AUTORETRIEVED — auto-numbers now
Fixed the ownership check in dbd_st_finish() at the spot you flagged — now checks imp_dbh->async_sth == imp_sth instead of just imp_sth->async_status
Replaced literal 0 with DBH_NO_ASYNC / STH_NO_ASYNC constants
Applied Yoda conditions throughout (NULL == ptr, STH_ASYNC_AUTORETRIEVED == status, etc.)
Merged the identical PGRES_TUPLES_OK / PGRES_COMMAND_OK branches into one, factored out the common async_status = STH_NO_ASYNC call
Removed all "CRITICAL" comments (Claude does indeed love that word 😅)
Cleaned up redundant/verbose comments, removed the unnecessary trace block
Simplified the auto-retrieved result handling by eliminating a duplicate check

Test file (t/08async_regression.t):

Switched to connect_database() / require 'dbdpg_test_setup.pl'
Dropped use strict; use warnings;
Removed all subtest blocks — flat structure with verbose test names
Trimmed the wordy comment blocks
Squashed the lib path fix into the test commit
57 tests, all passing against the full suite (3719 tests total)

t/08async_regression.t

turnstep · 2026-02-17T21:47:52Z

Thanks for the changes, will look those over soon. Debating releasing a 3.19.0 then making this part of 3.20.0 with a short turnaround.

turnstep

Just a few questions

dbdimp.c

t/08async_regression.t

dbdimp.c

esabol · 2026-02-18T00:50:52Z

@turnstep : Please scroll down to the bottom and approve running the CI workflow for this PR. Click on "Approve workflows to run."

- Fix double-free when imp_sth->result and imp_dbh->last_result point to the same PGresult by adding pointer equality checks - Remove redundant NULL assignments before immediate reassignment - Remove unreachable wrong-statement-handle branch in PGRES_TUPLES_OK (already caught by ownership check that returns -2) - Use PQcmdTuples for PGRES_COMMAND_OK row count in auto-retrieve path instead of hardcoding 0 - Pass imp_sth->result directly to _sqlstate instead of via local var - Add TRACE_PQERRORMESSAGE before pg_error in AUTOERROR null-check - Use -2 (error convention) for auto-error rows placeholder - Rewrite OLDQUERY_WAIT test with named variables instead of confusing array index arithmetic - Add skip_all for failed second DB connection in regression tests - Remove redundant isnt connection tests (covered by skip_all) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

turnstep · 2026-02-25T15:20:44Z

I've been tinkering with dbdimp.c a lot - can you rebase?

- Fix double-free when imp_sth->result and imp_dbh->last_result point to the same PGresult by adding pointer equality checks - Remove redundant NULL assignments before immediate reassignment - Remove unreachable wrong-statement-handle branch in PGRES_TUPLES_OK (already caught by ownership check that returns -2) - Use PQcmdTuples for PGRES_COMMAND_OK row count in auto-retrieve path instead of hardcoding 0 - Pass imp_sth->result directly to _sqlstate instead of via local var - Add TRACE_PQERRORMESSAGE before pg_error in AUTOERROR null-check - Use -2 (error convention) for auto-error rows placeholder - Rewrite OLDQUERY_WAIT test with named variables instead of confusing array index arithmetic - Add skip_all for failed second DB connection in regression tests - Remove redundant isnt connection tests (covered by skip_all) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

esabol · 2026-02-27T13:19:21Z

CI tests all failed with 3 failures in t/99_lint.t:

# Failure at line 5471: Found no matching TRACE for PQresultErrorMessage
# Failure at line 5859: Found no matching TRACE for PQntuples
# Failure at line 5862: Found no matching TRACE for PQcmdTuples
# Failure at line 5868: Found no matching TRACE for PQnfields

#   Failed test 'File "./dbdimp.c" does not have TRACE cals for each PQ function'
#   at t/99_lint.t line 126.
fatal: detected dubious ownership in repository at '/__w/dbdpg/dbdpg'
To add an exception for this directory, call:

	git config --global --add safe.directory /__w/dbdpg/dbdpg
# Looks like you failed 1 test of 728.
t/99_lint.t ............. 
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/728 subtests

~~Seems to be the result of changes made by @turnstep ?~~ No, latest CI run on master passes. https://github.com/bucardo/dbdpg/actions/runs/22487948562

detected dubious ownership in repository at '/__w/dbdpg/dbdpg' should have been fixed by the change to checkout@v4. Are you sure you rebased on latest master branch commits, @oetiker ?

esabol · 2026-02-27T14:01:48Z

Looking into this more closely, I think 99_lint.t line 153 should be changed to this:

       my %gitfiles = map { chomp; $_, 1 } qx{git -c 'safe.directory=*' ls-files};

in order to eliminate the detected dubious ownership in repository at '/__w/dbdpg/dbdpg' message.

oetiker · 2026-02-27T16:12:20Z

hmmm it looks as if pg has died during the tests ?

esabol · 2026-02-27T18:06:35Z

hmmm it looks as if pg has died during the tests ?

It's an intentional test. The problem comes from changes @turnstep made this morning in 74d0612, and it's unrelated to the work in this PR. The check @turnstep added doesn't work on Perl 5.36+. Once that's fixed, you'll need to rebase again, unfortunately, @oetiker.

turnstep · 2026-02-27T21:06:31Z

No worries, my goal this year is to do more PRs, less random pushes. Starting soon :)

esabol · 2026-02-27T22:25:40Z

@oetiker , the CI has been fixed. Please rebase one more time.

…#105) - Only clear async_sth if the statement being finished/destroyed owns it - Add ownership verification in pg_db_result() to prevent result theft - Auto-retrieve results in handle_old_async() when PG_OLDQUERY_WAIT is used, instead of discarding them - Add STH_ASYNC_AUTORETRIEVED and STH_ASYNC_AUTOERROR status values - Support both $sth->pg_result and $dbh->pg_result patterns - Add regression tests in t/08async_regression.t

- Fix double-free when imp_sth->result and imp_dbh->last_result point to the same PGresult by adding pointer equality checks - Remove redundant NULL assignments before immediate reassignment - Remove unreachable wrong-statement-handle branch in PGRES_TUPLES_OK (already caught by ownership check that returns -2) - Use PQcmdTuples for PGRES_COMMAND_OK row count in auto-retrieve path instead of hardcoding 0 - Pass imp_sth->result directly to _sqlstate instead of via local var - Add TRACE_PQERRORMESSAGE before pg_error in AUTOERROR null-check - Use -2 (error convention) for auto-error rows placeholder - Rewrite OLDQUERY_WAIT test with named variables instead of confusing array index arithmetic - Add skip_all for failed second DB connection in regression tests - Remove redundant isnt connection tests (covered by skip_all) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

oetiker · 2026-03-02T08:29:24Z

claude fortunately is unperturbed ...

esabol · 2026-03-02T10:29:16Z

All the CI tests pass! There's just one thing that seems a bit odd, and it's the message

NOTICE: table "async_test_constraints" does not exist, skipping

in the test output. It's completely harmless and expected whenever you do a DROP TABLE IF EXISTS and the table doesn't exist, of course, but it seems less than ideal. @oetiker and @turnstep, what do you think about adding the following right before the DROP TABLE IF EXISTS line?

$dbh_noerr->do(q{SET LOCAL client_min_messages = 'WARNING'});

turnstep · 2026-03-02T12:20:28Z

Few early-morning thoughts on that table:

We try to keep all artifacts with the dbd_pg_ prefix
Tables should go in t/dbdpg_test_setup.pl - and thus no need to worry about IF EXISTS
For something as simple as this, we can probably use dbd_pg_test table directly?

esabol · 2026-03-03T04:39:35Z

All of those things make sense to me. @oetiker , can you get Claude to either rename the async_test_constraints table to dbd_pg_async_constraints (or something like that) and add it to t/dbdpg_test_setup.pl or switch to using the existing dbd_pg_test table (if that's feasible)?

The table was created and dropped but never referenced by any query, producing a spurious NOTICE and using a non-standard name (missing the dbd_pg_ prefix). Addresses PR bucardo#150 review feedback. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

oetiker · 2026-03-05T08:31:22Z

done

zaucker mentioned this pull request Sep 19, 2025

PR #2268 for Mojo::Promise leads to regression with Mojo::Pg::Database mojolicious/mojo#2276

Open

oetiker changed the title ~~DBD::Pg Async Query Ownership and Data Preservation Fixes (Fix for #105 and https://github.com/mojolicious/mojo/issues/2276)~~ DBD::Pg Async Query Ownership and Data Preservation Fixes Sep 19, 2025

oetiker force-pushed the oetiker-async-regressions branch from e40cc86 to fe5c547 Compare December 15, 2025 09:14

esabol mentioned this pull request Dec 15, 2025

Fix pg db cancel missing #152

Merged

esabol approved these changes Dec 16, 2025

View reviewed changes

turnstep requested changes Feb 10, 2026

View reviewed changes

turnstep reviewed Feb 10, 2026

View reviewed changes

t/08async_regression.t Show resolved Hide resolved

turnstep requested changes Feb 10, 2026

View reviewed changes

turnstep approved these changes Feb 10, 2026

View reviewed changes

turnstep self-assigned this Feb 10, 2026

esabol mentioned this pull request Feb 11, 2026

pg_error_field rudely modifies its argument #157

Closed

oetiker force-pushed the oetiker-async-regressions branch from 82292d2 to 5c9c691 Compare February 17, 2026 13:25

oetiker force-pushed the oetiker-async-regressions branch 2 times, most recently from 898f315 to 28413ed Compare February 17, 2026 13:43

oetiker commented Feb 17, 2026

View reviewed changes

t/08async_regression.t Show resolved Hide resolved

turnstep requested changes Feb 17, 2026

View reviewed changes

oetiker force-pushed the oetiker-async-regressions branch from 8f1e267 to a86053a Compare February 27, 2026 10:36

oetiker and others added 3 commits March 2, 2026 09:28

Add missing TRACE macros for PQ function calls (fix t/99_lint.t)

5f6f84e

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

oetiker force-pushed the oetiker-async-regressions branch from c74c7f1 to 5f6f84e Compare March 2, 2026 08:29

oetiker requested a review from turnstep March 2, 2026 08:30


		imp_sth_t *orig_sth = imp_dbh->async_sth;

		/* Store the error result for the original statement to discover */

Conversation

oetiker commented Sep 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

DBD::Pg Async Query Ownership and Data Preservation Fixes

Overview

Problems Addressed

1. GitHub Issue #105: Statement Destruction Cancels Unrelated Async Queries

2. PG_OLDQUERY_WAIT Data Loss

3. Missing Ownership Verification

Technical Background

The Fixes

Fix 1: Ownership Checking in Statement Lifecycle Functions

Fix 2: Auto-Retrieve Mechanism for PG_OLDQUERY_WAIT

Fix 3: Ownership Verification in pg_db_result()

Fix 4: Auto-Retrieved Result Handling

New Async Status Values

Test Coverage

Behavior Changes

Before the Fix

After the Fix

Migration Notes

Uh oh!

oetiker commented Oct 20, 2025

Uh oh!

turnstep commented Dec 10, 2025

Uh oh!

esabol commented Dec 12, 2025

Uh oh!

oetiker commented Dec 13, 2025

Uh oh!

oetiker commented Dec 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

esabol commented Dec 15, 2025

Uh oh!

esabol commented Dec 15, 2025

Uh oh!

esabol commented Feb 9, 2026

Uh oh!

oetiker commented Feb 10, 2026

Uh oh!

esabol commented Feb 10, 2026

Uh oh!

turnstep left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

turnstep left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oetiker commented Sep 19, 2025 •

edited

Loading

oetiker commented Dec 15, 2025 •

edited

Loading

esabol commented Feb 18, 2026 •

edited

Loading

esabol commented Feb 27, 2026 •

edited

Loading

esabol commented Feb 27, 2026 •

edited

Loading

esabol commented Feb 27, 2026 •

edited

Loading