| 1 |
# Reference point for ABI compliance checks |
| 2 |
# |
| 3 |
# This file lists commits on the current branch that break ABI compatibility in |
| 4 |
# ways that have been deemed acceptable (e.g., removing an extern function with |
| 5 |
# no third-party uses). The primary intent of this file is to control the ABI |
| 6 |
# compliance checks on the buildfarm, but it also serves as a central location |
| 7 |
# to document the justification for each. |
| 8 |
# |
| 9 |
# In general, entries should be added reactively after an abi-compliance-check |
| 10 |
# buildfarm failure. It is important to verify the details of the breakage |
| 11 |
# match expectations, as the first entry listed will become the updated ABI |
| 12 |
# baseline point. |
| 13 |
# |
| 14 |
# Add new entries by adding the output of the following to the top of the file: |
| 15 |
# |
| 16 |
# $ git log --pretty=format:"%H%n#%n# %s%n# %cd%n#%n# <ADD JUSTIFICATION HERE>" $ABIBREAKGITHASH -1 --date=iso |
| 17 |
# |
| 18 |
# Be sure to replace "<ADD JUSTIFICATION HERE>" with details of your change and |
| 19 |
# why it is deemed acceptable. |
| 20 |
|
| 21 |
2780538433fc3b40b4e1e7bc7cbd350e37b6d550 |
| 22 |
# |
| 23 |
# Check for USAGE privilege on types used by stored expressions. |
| 24 |
# 2026-08-10 06:38:12 -0700 |
| 25 |
# |
| 26 |
# The addition of a parameter to AlterDomainAddConstraint() breaks ABI |
| 27 |
# compatibility, but we are unaware of any impacted third-party code. |
| 28 |
|
| 29 |
2a29b607dbbba47c8e1c97458708e775bfa0e9bb |
| 30 |
# |
| 31 |
# Add an output_plugin_libraries GUC to bless trusted output plugins |
| 32 |
# 2026-08-10 06:38:12 -0700 |
| 33 |
# |
| 34 |
# Addition of a new GUC enlarges ConfigureNamesString[], |
| 35 |
# but that is not harmful. |
| 36 |
|
| 37 |
a1fa24127d6a24611e5ec8e5b00584b333a3a06c |
| 38 |
# |
| 39 |
# Preserve the owner of extended statistics rebuilt by ALTER TABLE. |
| 40 |
# 2026-08-10 06:38:11 -0700 |
| 41 |
# |
| 42 |
# CreateStatistics()'s signature changes and CreateStatsStmt gains a field, |
| 43 |
# but no known third-party code calls the former or constructs the latter. |
| 44 |
|
| 45 |
8d9a97e0bb6d820dac553848f0d5d8cc3f3e219d |
| 46 |
# |
| 47 |
# Avoid name collision with NOT NULL constraints |
| 48 |
# 2026-02-21 12:22:08 +0100 |
| 49 |
# |
| 50 |
# AddRelationNotNullConstraints() needs to receive the names of constraints |
| 51 |
# already created, so that it can avoid those names when assigning names to |
| 52 |
# not-null constraints. The purpose of this function is rather obscure, and |
| 53 |
# furthermore it's new in 18, so I don't expect any third-party code to break. |
| 54 |
# |
| 55 |
# Discussion: https://postgr.es/m/[email protected] |
| 56 |
|
| 57 |
33e3de6d77e87d6c3c6f8f878dd8de42d37c3b8f |
| 58 |
# |
| 59 |
# Add file_extend_method=posix_fallocate,write_zeros. |
| 60 |
# 2026-02-06 17:38:39 +1300 |
| 61 |
# |
| 62 |
# Modifying GUC tables isn't really an ABI break: the relevant object has |
| 63 |
# incomplete type so its layout is unknown to other C translation units. |
| 64 |
# |
| 65 |
# Discussion: https://www.postgresql.org/message-id/flat/e1f0cd3b-0164-45f5-9705-e922e59df90f%40dunslane.net#8a350b54012c0042f9869d288e978cfe |
| 66 |
|
| 67 |
492a69e1407029f8c673484f44aa719a63323d77 |
| 68 |
# |
| 69 |
# Reject ADD CONSTRAINT NOT NULL if name mismatches existing constraint |
| 70 |
# 2026-02-03 12:33:29 +0100 |
| 71 |
# |
| 72 |
# This commit added the constraint name as a parameter to |
| 73 |
# AdjustNotNullInheritance(), which it needed to verify that the new name |
| 74 |
# matches the existing one. PGXN contained no calls to that function. |
| 75 |
|
| 76 |
c6ce4dcf9d3b7a8a89aca386124c473f98fc329e |
| 77 |
# |
| 78 |
# Fix trigger transition table capture for MERGE in CTE queries. |
| 79 |
# 2026-01-24 11:30:48 +0000 |
| 80 |
# |
| 81 |
# This commit changed the TransitionCaptureState structure, replacing |
| 82 |
# the "tcs_private" field with 3 separate fields. This structure can |
| 83 |
# only be built using MakeTransitionCaptureState(), and PGXN contained |
| 84 |
# no calls to MakeTransitionCaptureState() or uses of the |
| 85 |
# TransitionCaptureState structure. |
| 86 |
|
| 87 |
bae8ca82fd00603ebafa0658640d6e4dfe20af92 |
| 88 |
# |
| 89 |
# Revisit cosmetics of "For inplace update, send nontransactional invalidations." |
| 90 |
# 2025-12-15 12:19:53 -0800 |
| 91 |
# |
| 92 |
# This removed a CacheInvalidateHeapTupleInplace() parameter. PGXN contained |
| 93 |
# no calls to that function. |
| 94 |
|
| 95 |
00eb646ea43410e5df77fed96f4a981e66811796 |
| 96 |
# |
| 97 |
# Check for CREATE privilege on the schema in CREATE STATISTICS. |
| 98 |
# 2025-11-10 09:00:00 -0600 |
| 99 |
# |
| 100 |
# This commit added a parameter to CreateStatistics(). We are unaware of any |
| 101 |
# impacted third-party code. |
| 102 |
|
| 103 |
c8af5019bee5c57502db830f8005a01cba60fee0 |
| 104 |
# |
| 105 |
# Fix lookups in pg_{clear,restore}_{attribute,relation}_stats(). |
| 106 |
# 2025-10-15 12:47:33 -0500 |
| 107 |
# |
| 108 |
# This commit replaced two functions related to lookups/privilege checks for |
| 109 |
# the new stats stuff in v18 with RangeVarGetRelidExtended(). These functions |
| 110 |
# were not intended for use elsewhere, exist in exactly one release (18.0), and |
| 111 |
# do not have any known third-party callers. |
| 112 |
|
| 113 |
9bbcec6030a2744d83311370ec92213fbd76e514 |
| 114 |
# |
| 115 |
# Translation updates |
| 116 |
# 2025-09-22 14:18:56 +0200 |
| 117 |
# |
| 118 |
# This is the original ABI baseline point for REL_18_STABLE. |