[YSQL] Segmentation fault on return ROW from FUNCTION #1676
Comments
d-uspenskiy
changed the title
d-uspenskiy
changed the title
|
Stack trace for: select composrec(); |
|
Slightly different stack for the 2nd issue (temptable) issue reported above. |
|
This |
jaki
pushed a commit
to jaki/yugabyte-db
that referenced
this issue
Aug 22, 2019
Fix another issue around records where a heap tuple does not set its `t_ybctid` invalid. Resolve the yugabyte#1676 TODOs, mainly involving uncommenting `SELECT` statements. Fix some whitespace diff issues. The `yb_plpgsql` test should continue to pass.
jaki
pushed a commit
that referenced
this issue
Aug 26, 2019
Summary: Enable `CREATE TYPE` and `DROP TYPE` for user-defined types, closing issue #1152. This includes user-defined composite types, enum types, range types, base types, and shell types. Also extend support to record types, particularly those returning from functions, and close issues #742 and #1676. * Defer `ALTER TYPE` to issue #1893 * Map Postgres user-defined types either to existing Postgres base types or directly to Yugabyte types * Apply fixes related to issue #1156 in regress test `yb_pg_plpgsql` (formerly named `yb_plpgsql`) * Invalidate some `t_ybctid` attributes of `HeapTuple`s to get records working * Discover and create a potentially related issue #1975 * Update several Postgres regress tests * Create new Postgres regress tests * Rename related tests from `yb_foo` to `yb_pg_foo` if they are derivatives of the vanilla Postgres `foo` test * Port related tests from `foo` to `yb_pg_foo` by copying then modifying * Fix an issue with the expected output of `TestPgDump` that is **not** related to UDTs and records Assuming it still exists, see fine-grained commit history at jj-kim/yugabyte-db/tree/support-user-defined-type. Otherwise, here are the commit message subjects (most recent to least recent). * Comment out more of an unrelated issue * Add more DISCARD TEMP just in case * Add DISCARD TEMP to the end of some tests * Set UDT type entity OIDs to invalid * Combine some DROP TABLE statements * Fix yb_pg_foreign_key expected output from 0294a9c * Rename schedule collections to arrays * Add yb_pg_rowtypes regress test * Order tests in schedules; split yb_create_index * Dynamically create type entities for certain UDTs * Improve base type test with custom C func type * Add base type test and handle more base type cases * Increase fixed length type entity size to 8 bytes * Reduce large diffs for yb_pg_jsonb test * Add testjsonb table for yb_pg_jsonb test * Reduce diff for yb_pg_jsonb Postgres regress tests * Rename more Postgres regress tests * Port case regress test to yb_pg_case * Add missing feature tests to yb_feature_types * Fix TestPgDump expected output discrepancy * Uncomment/fix more statements in yb_arrays test * Fix TestPgRegressBetaFeatures java test from tenk1 * Fix TestPgRegressPgMisc java test due to tenk1 * Remove WITH OIDS when creating tenk1 table * Finish cleaning off yb_privileges test * Resolve part of yb_privileges regress test * Remap UDTs without mucking around sys attributes * Increase timeout of TestPgRegressLargeTable * Uncomment TODOs in yb_arrays test * Remove TestPgTypes java test * Uncomment more statements in yb_pg_json test * Uncomment some statements in yb_pg_json test * Remove intermediate UDT mapping for direct mapping * Uncomment TODOs in yb_pg_enum regress test * Update utils to have new UDTOIDs * Add small test for record types * Change function name in TestPgRegressTypesUDT.java * Add test for creating and dropping types * Rename several regress tests * Redirect ALTER TYPE to issue #1893 * Create custom primitives for UDTs * Fix and address #1676 comments in yb_plpgsql * Address last half of #1156 comments in yb_plpgsql * Address first half of #1156 comments in yb_plpgsql * Update yb_plpgsql test expected output * Support functions returning record type * Create new debug utility YBCPgDataTypeToStr * Don't error on domain type * Update some existing yb_foo regress tests * Add TODO comment for BYTEAARRAY to BYTEA * Add tests for creating enums and ranges * Rename yb_enum to yb_pg_enum * Copy over CREATE TYPE tests * Support user-defined enum and base types * Remove IsYugaByteEnabled check for tup init * Set base type of user-defined types to BYTEAARRAY * Enable CREATE TYPE and DROP TYPE Test Plan: * `TestPgRegressArrays` java test, particularly `yb_pg_arrays` regress test * `TestPgRegressAuthorization` java test, particularly `yb_pg_privileges` regress test * `TestPgRegressBetaFeatures` java test, particularly `yb_pg_plpgsql`, `yb_plpgsql` regress tests * `TestPgRegressFeature` java test, particularly `yb_feature_types` regress test * `TestPgRegressPgMisc` java test, particularly `yb_pg_case`, `yb_pg_identity` regress tests * `TestPgRegressTypesString` java test, particularly `yb_pg_json`, `yb_pg_jsonb` regress tests * `TestPgRegressTypesUDT` java test, particularly `yb_base_type`, `yb_create_type`, `yb_pg_create_type`, `yb_pg_enum`, `yb_pg_rangetypes`, `yb_pg_rowtypes` Unrelated tests that got added or changed: * `yb_create_function` regress test, involving `TestPgRegressTypesUDT` java test * `yb_create_index`, `yb_pg_create_index` regress tests, involving `TestPgRegressMisc`, `TestPgRegressTypesUDT` java tests * `yb_create_table`, `yb_drop_table` regress tests, involving `TestPgRegressTable` java test * `yb_feature_temp` regress test, involving `TestPgRegressFeature` java test * `yb_pg_copy`, `yb_pg_create_table` regress tests, involving `TestPgRegressArrays`, `TestPgRegressBetaFeatures`, `TestPgRegressMisc`, `TestPgRegressPgMisc`, `TestPgRegressTypesString`, `TestPgRegressTypesUDT` java tests * `yb_pg_foreign_key` regress test, involving `TestPgForeignKey` java test * `yb_pg_int8` regress test, involving `TestPgRegressAuthorization`, `TestPgRegressTypesNumeric`, `TestPgRegressTypesUDT` java tests * `yb_pg_triggers` regress test, involving `TestPgRegressBetaFeatures` java test * `yb_select`, `yb_sequence` regress tests, involving `TestPgRegressPgMisc` java test * `yb_triggers` regress test, involving `TestPgRegressBetaFeatures` java test Reviewers: mihnea, neha Reviewed By: neha Subscribers: neil, kannan, yql Differential Revision: https://phabricator.dev.yugabyte.com/D6900
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The following sequence of commands causes
Segmentation faultin postgresql processAlso SELECT from TEMP table causes
Segmentation faultThe text was updated successfully, but these errors were encountered: