* Skipping serialization for ignored objects
Serialization now evaluates the ignore list and skips matching objects. Note: The match is using the ABAP `cp` operator.
Example:
```
*ztest.prog.* Ignore individual program
*.ectd.* Ignore all objects of a given type
*/subpack/* Ignore all objects in a sub-package
```
In case an objects originates from a different system (`<> sy-sysid`), an icon will be shown after the object. On hover, you can see the id of the original system.
Closes https://github.com/abapGit/abapGit/issues/3865
When serializing objects in parallel threads, it can happen that there are not sufficient system resources available (free dialog processes). In this case, abapGit can run into error "RFC task <task name> already open". This is caused by reusing the same task name in subsequent tries to serialize an object.
The correction appends a sequential number to the previous task name so each try will result in a unique task name.
Closes#5224.
PS: This does not fix any resource issues but should prevent the error. It will lead to wait situations and longer processing, however.
There are scenarios where you want to serialize objects differently than abapGit. This exit is called at the end of the serialize process and gives an opportunity to change the content of the serialzed files.
One such scenario would be to automatically pretty-print code. This can be achieved by looping over all `.abap` files and processing them accordingly in this exit.
Closes#5050
Co-authored-by: Lars Hvam <larshp@hotmail.com>
* Refactor serializer class
- Consolidate i18n parameters
- Move determination of translation languages into serializer class
- Pass translation language setting when parallel processing (was a bug)
* Lint
* Add unit tests
Co-authored-by: Lars Hvam <larshp@hotmail.com>
the old get_instance() method in ZCL_ABAPGIT_PERSIST_SETTINGS has been moved to a method in zcl_abapgit_persist_factory, plus interface added. Added to persist injector
* Terminology: Inclusive Language - Part 6
Change remaining variables and switch setting to main_language.
Finally closes#4043
* Add todo
Co-authored-by: Lars Hvam <larshp@hotmail.com>
* Release serialization resources (fixes#4774)
Release serialization resources to prevent dump #4774 for export of large packages
* Free files after parallelization
* transfer files one by one
* Update src/objects/core/zcl_abapgit_serialize.clas.abap
Co-authored-by: abaplint[bot] <24845621+abaplint[bot]@users.noreply.github.com>
* fix lint, missing fs declaration
* index access on delete lt_found
* Update src/objects/core/zcl_abapgit_serialize.clas.abap
Co-authored-by: abaplint[bot] <24845621+abaplint[bot]@users.noreply.github.com>
* cleanup: index variable not required
* rollback to original APPEND LINES
Co-authored-by: abaplint[bot] <24845621+abaplint[bot]@users.noreply.github.com>
Co-authored-by: Lars Hvam <larshp@hotmail.com>
Combines `zcl_abapgit_file_status=>identify_object` and `zcl_abapgit_object_files=>filename` into new class `zcl_abapgit_filename_logic`. This class provides `file_to_object` and `object_to_file` conversion logic similar to what `zcl_abapgit_folder_logic` does for packages and git folders.
Also includes unit tests which were missing for the old methods.
Co-authored-by: Lars Hvam <larshp@hotmail.com>
* Terminology: Inclusive Language - Part 5
Change of terminology (#4043)
- Replace `get_master_language` with `get_main_language` and remove it
* File Status
* Jump
* Language
* Main language
* Main language
* Show "data" by object in repo view
Continuing on #3441
- "data" is now shown as `TABU` `<table>` in the repo view
* Remove get_item
* Add cmnt
* Update src/objects/core/zcl_abapgit_serialize.clas.abap
Co-authored-by: abaplint[bot] <24845621+abaplint[bot]@users.noreply.github.com>
* Update zcl_abapgit_serialize.clas.abap
Co-authored-by: Lars Hvam <larshp@hotmail.com>
Co-authored-by: abaplint[bot] <24845621+abaplint[bot]@users.noreply.github.com>
* Group unsupported messages per object type
When number of unsupported objects increase in a package, user needs to scroll through the long list of messages. This change groups messages per object type reducing the number of messages. Also, the filtering before serialization is good in terms of effective work process utilization( in parallel mode ).
* Replace LINE_EXISTS with READ TABLE
* Cache the supported object types
* Show object name
If there is only one object name for the unsupported object type, show the name.
* Spelling: "object(s)" to "objects"
Co-authored-by: Marc Bernard <59966492+mbtools@users.noreply.github.com>
Co-authored-by: Lars Hvam <larshp@hotmail.com>
