[GNC-dev] gnucash maint: Multiple changes pushed

[John Ralls]

> On Apr 10, 2018, at 1:22 AM, Geert Janssens <geert.gnucash at kobaltwit.be> wrote:
> 
> Op dinsdag 10 april 2018 00:18:07 CEST schreef John Ralls:
>> Updated	 via  https://github.com/Gnucash/gnucash/commit/c3f528b8 (commit)
>> 	 via  https://github.com/Gnucash/gnucash/commit/aea33ca5 (commit)
>> 	from  https://github.com/Gnucash/gnucash/commit/972647d2 (commit)
>> 
>> 
>> 
>> commit c3f528b80c6cccdf9c139dc0f63a96ef513a0e62
>> Author: John Ralls <jralls at ceridwen.us>
>> Date:   Mon Apr 9 15:17:32 2018 -0700
>> 
>>    Bug 795068 - src/optional/python-bindings missing several files
>> 
>> diff --git a/src/optional/python-bindings/CMakeLists.txt
>> b/src/optional/python-bindings/CMakeLists.txt index e87c194..96d92a1 100644
>> --- a/src/optional/python-bindings/CMakeLists.txt
>> +++ b/src/optional/python-bindings/CMakeLists.txt
>> @@ -1,7 +1,9 @@
>> ADD_SUBDIRECTORY(example_scripts)
>> ADD_SUBDIRECTORY(tests)
>> 
>> -IF (BUILDING_FROM_VCS)
>> +  SET(PYEXEC_FILES  __init__.py function_class.py gnucash_business.py
>> gnucash_core.py) +
>> +  IF (BUILDING_FROM_VCS)
>>   SET(SWIG_FILES ${CMAKE_CURRENT_SOURCE_DIR}/gnucash_core.i
>> ${CMAKE_CURRENT_SOURCE_DIR}/timespec.i) SET(GNUCASH_CORE_C_INCLUDES
>>     ${CMAKE_BINARY_DIR}/src/config.h
>> @@ -63,8 +65,6 @@ IF(WITH_PYTHON)
>>     ${PYTHON_INCLUDE_DIRS}
>>   )
>> 
>> -  SET(PYEXEC_FILES  __init__.py function_class.py gnucash_business.py
>> gnucash_core.py) -
>>   ADD_LIBRARY(gnucash_core_c MODULE ${SWIG_GNUCASH_CORE_C})
>>   TARGET_INCLUDE_DIRECTORIES(gnucash_core_c PRIVATE
>> ${gnucash_core_c_INCLUDE_DIRS})
>> 
>> 
>> commit aea33ca515aceb4f63a5cbc8165d4ec93f245d0c
>> Author: John Ralls <jralls at ceridwen.us>
>> Date:   Sun Apr 8 15:20:55 2018 -0700
>> 
>>    Bug 795049 - GnuCash 2.6.20-1 (Fedora Linux package) is unable to
>> open...
>> 
>>    MariaDB database.
>> 
>> diff --git a/src/backend/dbi/test/test-backend-dbi-basic.c
>> b/src/backend/dbi/test/test-backend-dbi-basic.c index 386a2c3..4a55a62
>> 100644
>> --- a/src/backend/dbi/test/test-backend-dbi-basic.c
>> +++ b/src/backend/dbi/test/test-backend-dbi-basic.c
>> @@ -666,15 +666,15 @@ test_suite_gnc_backend_dbi (void)
>>     }
>>     if (g_list_find_custom (drivers, "sqlite3", (GCompareFunc)g_strcmp0))
>>         create_dbi_test_suite ("sqlite3", "sqlite3");
>> -    if (strlen (TEST_MYSQL_URL) > 0 &&
>> -            g_list_find_custom (drivers, "mysql", (GCompareFunc)g_strcmp0))
>> -        create_dbi_test_suite ("mysql", TEST_MYSQL_URL);
>>     if (strlen (TEST_PGSQL_URL) > 0 &&
>>             g_list_find_custom (drivers, "pgsql", (GCompareFunc)g_strcmp0))
>> {
>>         g_setenv ("PGOPTIONS", "-c client_min_messages=WARNING", FALSE);
>>         create_dbi_test_suite ("postgres", TEST_PGSQL_URL);
>>     }
>> +    if (strlen (TEST_MYSQL_URL) > 0 &&
>> +            g_list_find_custom (drivers, "mysql", (GCompareFunc)g_strcmp0))
>> +        create_dbi_test_suite ("mysql", TEST_MYSQL_URL);
>> 
>>     GNC_TEST_ADD_FUNC( suitename, "adjust sql options string localtime",
>>         test_adjust_sql_options_string );
> 
> This is surprising: this commit suggests only the order of the DB test would 
> make a difference on whether a mariadb based book can be opened or not; 
> testing for mysql after testing for postgresql would fix this.
> I understood from our IRC conversation yesterday the issue was that there were 
> two mysql statements in one dbi connection query. Did you forget to commit 
> that part of the fix or did we misunderstand the problem ?


Rats. Yes, the actual fix isn’t in the commit, so I’ll have to fix that and re-tag 2.6.21. Thanks for catching it!

The change to the order in the test program was to get the test to run the postgres test before getting stopped by the mysql one failing to see if postgres would work with the two-query statement.

Regards,
John Ralls