This commit adds an indirection for function module GUI_IS_AVAILABLE
via interface ZIF_ABAPGIT_GUI_FUNCTIONS and factory ZCL_ABAPGIT_UI_FACTORY.
For tests the implementation can be replaced with
ZCL_ABAPGIT_UI_INJECTOR=>SET_GUI_FUNCTIONS
* mark inactive objects in repo list
Inactive object are marked in the repo list
The sort order for changed files is not changed
unchanged inactive files come right after the changed files
* move acitve check to object interface
* add is_active for objects up to F
* add is_active for objects up to P
* add is_active for objects up to T
* add is_active for objects up to X
all core object implement (except SSFO) the standard active check via FM RS_INACTIVE_OBJECTS_WARNING now
* add is_active to plugins bridge - always active
* finish object interface implementation + objects
* add is_active for offline with object interface
* rename returning to RV_
* use returning for serialize
* cleanup, PP, UT
* add code line attribute for leading spaces
Use attribute abapgit-leadig-spaces to store leading spaces of coding lines at serialization. This is used in deserialization to restoreformatting after XML normalization.
* add attribute also for global types and forms
* fix code formatting
Before this commit is applied the FIX_IDS methods was errorneous
and lead to subsequent dump. This commit fixes the algorithm to
replace IDs and IDREFs
* pretty print
* definitions to top of routine
* space before .
* omit parameter names where possible
* double space after if
* remove pseudo comments where possible
* clear PRESP and PWORK
clear username fields PRESP and PWORK
* remove unused FM exceptions
if sy-subrc is analysed (only <> 0) no detailed FM exceptions are needed, only OTHERS
* set PRESP and PWORK on deserialization
PRESP and PWORK is not set to SY-UNAME by default so set it explicitly
* Object Type UDMO Data Model #1905
Support for Object Type UDMO Data Model #1905
* Object Type UDMO Data Model #1905
* Object Type UDMO Data Model #1905
* Object Type UDMO Data Model #1905 Bad formatting correction
* Object Type UDMO Data Model #1905 Bad formatting
* Object Type UDMO Data Model #1905 Reformulation
* Update package.json
* pretty print
Before this commit is applied DDLX source code was recorded
in the xml file. We changed this so that now a separte .asddls
file is created. The change is downward compatible. Previously
recorded DDLX objects will show a diff the next time and the new
.asddls file has to be committed.
* removed popup saving msag
* fixed capitalization
* refactored with variable prefixes
* refactored exceptions and split methods
also removed S messages.
Previous code was mostly copy %26 paste of the original
* Cleaned up
* more cleanup
* enable ADT jump from CI result
* enable ADT jump from CI result for all objects
* cleanup + PP
* cleanup + PP + save/activate in ADT
* remove commented code
* objects: use a different constant for 'R3TR'
The constant rs_c_pgmid_r3tr comes from the package RS which is not
available in pure BASIS systems.
The constant seok_pgmid_r3tr comes from the package SEOK and abapGit is
already using another constants from this package - seok_access_*.
Related to #1826
* objects: replace rs_c_true with abap_true
The constant rs_c_true comes from the package RS which is no available
in pure BASIS systems.
I believe that the constant rs_c_true was used by mistake.
The constant was assigned to a variable which is of the type abap_bool
and the variable is initialized with abap_false.
Related to #1826
This commit adds three new methods to ZCL_ABAPGIT_OBJECTS_SUPER:
- SERIALIZE_LONGTEXTS
- DESERIALIZE_LONGTEXTS
- DELETE_LONGTEXTS
They can be called on demand in concrete object type serializers.
In this commit they were called from DTEL, DOMA and TABL
* correct xml errors of CUA interfaces
Issue error RS_CUA_INTERNAL_WRITE while pulling an old abapgit repository #1807. Same issue as #562 (error RS_CUA_INTERNAL_WRITE) which is solved by correcting manually the XML of the CUA interfaces (fastidious), or by regenerating the XML (but unfortunately sometimes the abapGit repository is from someone else).
Automatically fix this before RS_CUA_INTERNAL_WRITE is called. It's useless having this as a configuration option.
* refactor: extract code to a new method
* Better refactoring of auto correction CUA
https://github.com/larshp/abapGit/issues/1575
Deletion of PARA caused a popup in FM RS_PARAMETER_DELETE.
To avoid that, the logic of the FM had to be implemented directly in the delete method.
TADIR performance #1756
All usages of TADIR aligned to use the same structure.
DELFLAG added in the structure.
2 identical structure definitions deleted.
todo: test this PR works after abapmerge has run
After this commit is applied all calls of "SET PARAMETER 'EUK'"
are replaced with the call of objects_super->set_default_package.
The method uses the ABAP memory which is in this case more reliable.
* CHAR: fix package popup during pull
* Remove superfluous ABAP memory clearance
memory is already cleared in RS_CORR_INSERT
* Introduce objects_super set_default_package
Refactor set default package logic to super class
zcl_abapgit_objects_super method set_default_package.
If the short text of the function group has changed in the remote repo,
it's currently not possible to pull that change.
With this commit the short text is pulled properly.
The data in table ICFAPPLCUST (Application Customizing Data) isn't
deleted by cl_icf_tree=>if_icf_tree~delete_node( ). Before this commit
is applied it's not possible to delete ICF nodes with data in that table.
With this commit we delete the data in the table. It's the same as
transaction SICF does.
Before this commit VCLS jump calls VIEWCLUSTER_MAINTENANCE_CALL
which opens inplace. But that's not compatible with abapGit UI
and nothing is visible.
And the semantics changed. The above mentioned FM call the view
cluster for data maintenance and not the development object.
To be aligned with abapGit logic now the transaction SE54 is called.
Unfortunately, we are versioning AVAS objects that can be instantiated
only in SAP systems but we use the repo also in sandbox Customer
systems.
Whenever I open the repository I get an exception of the type
cx_pak_invalid_data.
Without this commit I see the error message:
"AVAS error".
With this commit I see the error message:
"AVAS $GUID: invalid data: The functionality is not supported in
customer systems"
DDLS objects are not cheked in a transport in their deserialize method.
The parallel activation does not call RS_WORKING_OBJECTS_ACTIVATE which
checks in the activated object in a transport.
Therefor, DDLS objecst are never checked in a transport when
the parallel activation is in place (the option experimental features
is enabled - which is enabled by default on all our systems).
I wanted to use TRINT_OBJECTS_CHECK_AND_INSERT because this function
is used by ADT:
CL_WB_ADT_REST_RESOURCE~post
CL_WB_ADT_REST_RESOURCE_DATA
~if_wb_adt_rest_resource_data~get_transport_key
CL_WB_OBJECT~get_transport_key -> PGMID%3DR3TR, OBJTYPE%3DDDLS
CL_WB_ADT_REST_RESOURCE~lock
CL_WB_ADT_REST_RESOURCE~check_obj_in_transport_request
CL_ADT_CTS_MANAGEMENT~check_objects_in_request
CL_ADT_CTS_MANAGEMENT~LCL_DB_PERSISTENCE~check_objects
CALL FUNCTION CTS_WBO_API_CHECK_OBJECTS
CALL FUNCTION TRINT_WBO_IMPL_CHECK_OBJECTS
SAPLCTS_WBO_API_IMPL /
LCTS_WBO_API_IMPLF06 /
FORM CALL_OBJECTS_CHECK_AND_INSERT
CALL FUNCTION
TRINT_OBJECTS_CHECK_AND_INSERT
But it is not enough to call this function, hence, I use RS_CORR_INSERT
which was mentioned by @larshp in our slack conversation. The function
internally use calls TRINT_OBJECTS_CHECK_AND_INSERT.
I tested this patch with modification of a CDS view and also with
adding a new CDS view.
Closes#1466
The short description contains "(unused)".
For some strange reason, I saw that field set to D in one of our sandbox
systems.
Addresses the issue #1427.
Old code generate error of duplicity because SPRAS is case sensitive, but the name of files not.
New code use the 2-caracteres code for language.
Examples (T002):
SPRAS: 'C' -> LAISO: CS -> 'Czech'
SPRAS: 'c' -> LAISO: CA -> 'Catalan'
For some reason, we are using experimental features of abapGit.
Some time ago, all class were re-serialized with the serializer
LCL_OBJECT_CLASS_NEW.
After upgrade to 1.60, we started getting an error telling that CLAS
is not supported.
The CLASS NEW serializer became standard and the class has been renamed.
Unfortunately, our xml files still includes reference to the CLASS NEW.
I would just use a sed script, if knew all the affected repositories -
we have plenty of repositories versioned by abapGit.
This patch revives the CLASS NEW. It reports itself as the standard
class, so the CLASS NEW can be removed in year or two when all files
are re-serialized.
See commit 21d22d93f2
* ECTD: Ecatt test data container serialization
* add changed_by | clear SAPRL and VAR_EXT_PATH
* fix changed by
* fix conditional logic in changed by
* fix linter errors
* Add T100 support to ZCX_ABAPGIT_EXCEPTION
* Use T100 error messages in DEVC
* Replace zcx_abapgit_exception->text with get_text
* Use SE24 generated constructor in T100 exception
