pidgin/pidgin

Fix several leaks in tests

14 months ago, Elliott Sales de Andrade
4e1bf25f5575
Fix several leaks in tests

All of these are specific to tests, not the library code.

For the moment, `protocol_xfer` still leaks connections (and anything they hold on to) because it is very difficult to disentangle them from the connection manager in the partially implemented state they are in.

This fixes leaks of options in the account option test (these two leaks occur for every test since they all leak the option):
```
61 (48 direct, 13 indirect) bytes in 1 blocks are definitely lost in loss record 133 of 276
at 0x4848464: calloc (vg_replace_malloc.c:1340)
by 0x49F75F0: g_malloc0 (gmem.c:163)
by 0x48C3B2E: purple_account_option_new (purpleaccountoption.c:78)
by 0x4014AF: test_purple_account_option_copy_int (test_account_option.c:67)
by 0x4A1CC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A1CC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1D181: g_test_run_suite (gtestutils.c:3115)
by 0x4A156EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A156EC: g_test_run (gtestutils.c:2221)
by 0x401721: main (test_account_option.c:157)

61 (48 direct, 13 indirect) bytes in 1 blocks are definitely lost in loss record 134 of 276
at 0x4848464: calloc (vg_replace_malloc.c:1340)
by 0x49F75F0: g_malloc0 (gmem.c:163)
by 0x48C3BC7: purple_account_option_copy (purpleaccountoption.c:93)
by 0x4014BF: test_purple_account_option_copy_int (test_account_option.c:68)
by 0x4A1CC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A1CC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1D181: g_test_run_suite (gtestutils.c:3115)
by 0x4A156EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A156EC: g_test_run (gtestutils.c:2221)
by 0x401721: main (test_account_option.c:157)
```
leaks in the credential manager test (times 3 for read/write/cancel tests):
```
69 (16 direct, 53 indirect) bytes in 1 blocks are definitely lost in loss record 2,427 of 3,503
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x4A58168: g_malloc (gmem.c:130)
by 0x4A6FAB5: g_slice_alloc (gslice.c:1074)
by 0x4A700EC: g_slice_alloc0 (gslice.c:1100)
by 0x4A3BECB: g_error_allocate (gerror.c:710)
by 0x4A3C93F: UnknownInlinedFun (gerror.c:724)
by 0x4A3C93F: g_error_new_valist (gerror.c:766)
by 0x4BEE558: g_task_return_new_error (gtask.c:1941)
by 0x48D82C0: purple_credential_manager_read_password_async (purplecredentialmanager.c:492)
by 0x403634: test_purple_credential_manager_no_provider_read_password_idle (test_credential_manager.c:329)
by 0x4A4ECB1: g_idle_dispatch (gmain.c:6124)
by 0x4A4FCBE: UnknownInlinedFun (gmain.c:3444)
by 0x4A4FCBE: g_main_context_dispatch (gmain.c:4162)
by 0x4AA5597: g_main_context_iterate.constprop.0 (gmain.c:4238)
by 0x4A4F28E: g_main_loop_run (gmain.c:4438)
by 0x40369F: test_purple_credential_manager_no_provider_read_password_async (test_credential_manager.c:345)
by 0x4A7DC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A7DC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A7D9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A7D9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A7E181: g_test_run_suite (gtestutils.c:3115)
by 0x4A766EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A766EC: g_test_run (gtestutils.c:2221)
by 0x4048F6: main (test_credential_manager.c:695)
```
a leak in the image test:
```
161 bytes in 1 blocks are definitely lost in loss record 260 of 274
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x4A55363: g_try_malloc (gmem.c:286)
by 0x4A3D630: UnknownInlinedFun (gfileutils.c:819)
by 0x4A3D630: UnknownInlinedFun (gfileutils.c:924)
by 0x4A3D630: g_file_get_contents (gfileutils.c:1027)
by 0x401890: test_image_new_from_file (test_image.c:144)
by 0x4A7DC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A7DC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A7D9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A7E181: g_test_run_suite (gtestutils.c:3115)
by 0x4A766EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A766EC: g_test_run (gtestutils.c:2221)
by 0x40195D: main (test_image.c:172)
```
a leak in queued output stream test:
```
72 (40 direct, 32 indirect) bytes in 1 blocks are definitely lost in loss record 219 of 396
at 0x49D51EF: g_type_create_instance (gtype.c:1909)
by 0x49BAC1F: g_object_new_internal (gobject.c:2228)
by 0x49BC247: g_object_new_with_properties (gobject.c:2391)
by 0x49BCFF0: g_object_new (gobject.c:2037)
by 0x402003: test_queued_output_stream_push_bytes_async_error (test_queued_output_stream.c:219)
by 0x4A7DC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A7DC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A7D9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A7E181: g_test_run_suite (gtestutils.c:3115)
by 0x4A766EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A766EC: g_test_run (gtestutils.c:2221)
by 0x402429: main (test_queued_output_stream.c:280)
```
and protocol xfer tests (times 3 for each test that creates a test protocol object):
```
112 (48 direct, 64 indirect) bytes in 1 blocks are definitely lost in loss record 3,430 of 3,698
at 0x49D51EF: g_type_create_instance (gtype.c:1909)
by 0x49BAC1F: g_object_new_internal (gobject.c:2228)
by 0x49BC247: g_object_new_with_properties (gobject.c:2391)
by 0x49BCFF0: g_object_new (gobject.c:2037)
by 0x40291C: test_purple_protocol_xfer_send_file_func (test_protocol_xfer.c:146)
by 0x4A7DC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A7DC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A7D9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A7E181: g_test_run_suite (gtestutils.c:3115)
by 0x4A766EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A766EC: g_test_run (gtestutils.c:2221)
by 0x402B64: main (test_protocol_xfer.c:195)
```
and util tests (times 3 for each call to `purple_text_strip_mnemonic` in the test):
```
5 bytes in 1 blocks are definitely lost in loss record 5 of 247
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x49F7168: g_malloc (gmem.c:130)
by 0x491975B: purple_text_strip_mnemonic (util.c:895)
by 0x4015B0: test_util_text_strip_mnemonic (test_util.c:49)
by 0x4A1CC7D: UnknownInlinedFun (gtestutils.c:2933)
by 0x4A1CC7D: g_test_run_suite_internal (gtestutils.c:3021)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1C9E4: g_test_run_suite_internal (gtestutils.c:3038)
by 0x4A1D181: g_test_run_suite (gtestutils.c:3115)
by 0x4A156EC: UnknownInlinedFun (gtestutils.c:2234)
by 0x4A156EC: g_test_run (gtestutils.c:2221)
by 0x401901: main (test_util.c:224)
```
and these leaks in any test that initializes the test UI:
```
4,104 bytes in 1 blocks are possibly lost in loss record 3,451 of 3,457
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x5235B67: sqlite3MemMalloc.lto_priv.0 (sqlite3.c:25493)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29181)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29227)
by 0x5232797: sqlite3Malloc.lto_priv.0 (sqlite3.c:29221)
by 0x523BD8B: pcache1Alloc.lto_priv.0 (sqlite3.c:53546)
by 0x5249A8B: UnknownInlinedFun (sqlite3.c:53675)
by 0x5249A8B: allocateTempSpace (sqlite3.c:70848)
by 0x52625A6: sqlite3VdbeExec.lto_priv.0 (sqlite3.c:93857)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:87995)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:88056)
by 0x525CBEE: sqlite3_step (sqlite3.c:88045)
by 0x529B324: sqlite3_exec (sqlite3.c:131002)
by 0x48FD558: purple_sqlite3_run_migration (purplesqlite3.c:37)
by 0x48FDBB4: purple_sqlite3_run_migrations_from_resources (purplesqlite3.c:195)
by 0x48FDED9: purple_sqlite_history_adapter_run_migrations (purplesqlitehistoryadapter.c:69)
by 0x48FE7F0: purple_sqlite_history_adapter_activate (purplesqlitehistoryadapter.c:287)
by 0x48DB656: purple_history_adapter_activate (purplehistoryadapter.c:181)
by 0x48DC9BC: purple_history_manager_set_active (purplehistorymanager.c:308)
by 0x402BA8: test_ui_init_history (test_ui.c:132)
by 0x402C80: test_ui_purple_init (test_ui.c:167)
by 0x4027BB: main (test_contact.c:88)

4,368 bytes in 1 blocks are possibly lost in loss record 3,453 of 3,457
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x5235B67: sqlite3MemMalloc.lto_priv.0 (sqlite3.c:25493)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29181)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29227)
by 0x5232797: sqlite3Malloc.lto_priv.0 (sqlite3.c:29221)
by 0x523BD8B: pcache1Alloc.lto_priv.0 (sqlite3.c:53546)
by 0x5240077: UnknownInlinedFun (sqlite3.c:53634)
by 0x5240077: pcache1FetchStage2 (sqlite3.c:54104)
by 0x5243E9C: UnknownInlinedFun (sqlite3.c:52671)
by 0x5243E9C: getPageNormal.lto_priv.0 (sqlite3.c:60628)
by 0x524A510: UnknownInlinedFun (sqlite3.c:60805)
by 0x524A510: btreeGetPage.lto_priv.0 (sqlite3.c:70289)
by 0x524C2F6: UnknownInlinedFun (sqlite3.c:71257)
by 0x524C2F6: sqlite3BtreeBeginTrans.lto_priv.0 (sqlite3.c:71647)
by 0x5266B3A: sqlite3VdbeExec.lto_priv.0 (sqlite3.c:93532)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:87995)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:88056)
by 0x525CBEE: sqlite3_step (sqlite3.c:88045)
by 0x48FD715: purple_sqlite3_get_schema_version (purplesqlite3.c:79)
by 0x48FD9DD: purple_sqlite3_run_migrations_from_resources (purplesqlite3.c:146)
by 0x48FDED9: purple_sqlite_history_adapter_run_migrations (purplesqlitehistoryadapter.c:69)
by 0x48FE7F0: purple_sqlite_history_adapter_activate (purplesqlitehistoryadapter.c:287)
by 0x48DB656: purple_history_adapter_activate (purplehistoryadapter.c:181)
by 0x48DC9BC: purple_history_manager_set_active (purplehistorymanager.c:308)
by 0x402BA8: test_ui_init_history (test_ui.c:132)
by 0x402C80: test_ui_purple_init (test_ui.c:167)
by 0x4027BB: main (test_contact.c:88)

4,368 bytes in 1 blocks are possibly lost in loss record 3,454 of 3,457
at 0x484386F: malloc (vg_replace_malloc.c:393)
by 0x5235B67: sqlite3MemMalloc.lto_priv.0 (sqlite3.c:25493)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29181)
by 0x5232797: UnknownInlinedFun (sqlite3.c:29227)
by 0x5232797: sqlite3Malloc.lto_priv.0 (sqlite3.c:29221)
by 0x523BD8B: pcache1Alloc.lto_priv.0 (sqlite3.c:53546)
by 0x5240077: UnknownInlinedFun (sqlite3.c:53634)
by 0x5240077: pcache1FetchStage2 (sqlite3.c:54104)
by 0x5243E9C: UnknownInlinedFun (sqlite3.c:52671)
by 0x5243E9C: getPageNormal.lto_priv.0 (sqlite3.c:60628)
by 0x52499BC: UnknownInlinedFun (sqlite3.c:60805)
by 0x52499BC: UnknownInlinedFun (sqlite3.c:70289)
by 0x52499BC: btreeGetUnusedPage (sqlite3.c:70432)
by 0x524F504: allocateBtreePage.lto_priv.0 (sqlite3.c:74604)
by 0x5256209: btreeCreateTable.lto_priv.0 (sqlite3.c:77830)
by 0x5265EB2: UnknownInlinedFun (sqlite3.c:77849)
by 0x5265EB2: sqlite3VdbeExec.lto_priv.0 (sqlite3.c:96382)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:87995)
by 0x525CBEE: UnknownInlinedFun (sqlite3.c:88056)
by 0x525CBEE: sqlite3_step (sqlite3.c:88045)
by 0x529B324: sqlite3_exec (sqlite3.c:131002)
by 0x48FD558: purple_sqlite3_run_migration (purplesqlite3.c:37)
by 0x48FDBB4: purple_sqlite3_run_migrations_from_resources (purplesqlite3.c:195)
by 0x48FDED9: purple_sqlite_history_adapter_run_migrations (purplesqlitehistoryadapter.c:69)
by 0x48FE7F0: purple_sqlite_history_adapter_activate (purplesqlitehistoryadapter.c:287)
by 0x48DB656: purple_history_adapter_activate (purplehistoryadapter.c:181)
by 0x48DC9BC: purple_history_manager_set_active (purplehistorymanager.c:308)
by 0x402BA8: test_ui_init_history (test_ui.c:132)
by 0x402C80: test_ui_purple_init (test_ui.c:167)
by 0x4027BB: main (test_contact.c:88)
```

Testing Done:
Ran tests in valgrind, and all above leaks were gone except for noted `protocol_xfer` issues, and a bunch of leaks of the `PurpleBuddy`-`PurpleContact` compatibility bindings, which will go away in their entirety eventually.

Reviewed at https://reviews.imfreedom.org/r/2385/
/*
* Purple - Internet Messaging Library
* Copyright (C) Pidgin Developers <devel@pidgin.im>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, see <https://www.gnu.org/licenses/>.
*/
#if !defined(PURPLE_GLOBAL_HEADER_INSIDE) && !defined(PURPLE_COMPILATION)
# error "only <pidgin.h> may be included directly"
#endif
#ifndef PURPLE_SQLITE3_H
#define PURPLE_SQLITE3_H
#include <glib.h>
#include <gio/gio.h>
#include <sqlite3.h>
G_BEGIN_DECLS
/**
* PURPLE_SQLITE3_ERROR:
*
* An error domain for sqlite3 errors.
*
* Since: 3.0.0
*/
#define PURPLE_SQLITE3_DOMAIN (g_quark_from_static_string("sqlite3"))
/**
* PurpleSqlite3:
*
* A sqlite3 connection.
*
* This type alias exists for introspection purposes, and is no different from
* the `sqlite3` type.
*
* Since: 3.0.0
*/
#ifdef __GI_SCANNER__
typedef gpointer PurpleSqlite3;
#else
typedef sqlite3 PurpleSqlite3;
#endif
/**
* purple_sqlite3_get_schema_version:
* @db: The sqlite3 connection.
* @error: Return address for a #GError, or %NULL.
*
* Attempts to read the result of `PRAGMA user_version` which this API uses to
* store the schema version.
*
* Returns: %TRUE on success, or %FALSE on error with @error set.
*
* Since: 3.0.0
*/
int purple_sqlite3_get_schema_version(PurpleSqlite3 *db, GError **error);
/**
* purple_sqlite3_run_migrations_from_strings:
* @db: The sqlite3 connection.
* @migrations: (array zero-terminated=1): A list of SQL statements, each item
* being its own migration.
* @error: Return address for a #GError, or %NULL.
*
* Runs the given migrations in the order they are given. The index of each
* migration plus 1 is assumed is to be the version number of the migration,
* which means that you can not change the order of the migrations. The
* reasoning for the addition of 1 is because `PRAGMA user_version` defaults to
* 0.
*
* This expects each string in @migrations to be a complete migration. That is,
* each string in the array should contain all of the SQL for that migration.
* For example, if you're expecting to have 2 migrations, the initial creating
* two tables, and then adding a column to one of the existing tables, you
* would have something like the following code.
*
* ```c
* const char *migrations[] = {
* // Our initial migration that creates user and session tables.
* "CREATE TABLE user(id INTEGER PRIMARY KEY, name TEXT);"
* "CREATE TABLE session(user INTEGER, token TEXT) FOREIGN KEY(user) REFERENCES user(id);",
* // Begin our second migration that will add a display name to the user
* // table. Note the ',' at the end of the previous line.
* "ALTER TABLE user ADD COLUMN(display_name TEXT);",
* NULL
* };
* ```
*
* Also, this function will run each migration in its own transaction so you
* don't need to worry about them. This is done to make sure that the database
* stays at a known version and an incomplete migration will not be saved.
*
* Returns: %TRUE on success, or %FALSE on error potentially with @error set.
*
* Since: 3.0.0
*/
gboolean purple_sqlite3_run_migrations_from_strings(PurpleSqlite3 *db, const char *migrations[], GError **error);
/**
* purple_sqlite3_run_migrations_from_resources:
* @db: The sqlite3 connection.
* @path: The base path in @resource to use.
* @migrations: (array zero-terminated=1): The list of migrations in the order
* to run them.
* @error: Return address for a #GError, or %NULL.
*
* Runs the given migrations in the order they are given. The index of each
* migration plus 1 is assumed to be the version number of the migration, which
* means that you can not change the order of the migrations. The reasoning for
* the addition of 1 is because `PRAGMA user_version` defaults to 0.
*
* This will attempt to load the migrations via
* [func@Gio.resources_open_stream] by concatenating @path and the individual
* items of @migrations. Each migration will be ran in a transaction that
* includes updating the schema version, which is stored in
* `PRAGMA user_version`. This means you can't use `PRAGMA user_version` for
* other things.
*
* Returns: %TRUE on success, or %FALSE on error potentially with @error set.
*
* Since: 3.0.0
*/
gboolean purple_sqlite3_run_migrations_from_resources(PurpleSqlite3 *db, const char *path, const char *migrations[], GError **error);
G_END_DECLS
#endif /* PURPLE_SQLITE3_H */