- make end indent regexp symmetrical to being regexp allowing for selective
change (e.g., disactivation via non-matching regexp) what is leads to
indentation
Signed-off-by: Michael Steiner <michael.steiner@intel.com>
Giving a default value to `plantuml-exec-mode` local var at startup time
is wrong. This change keeps the local var to `nil` and resorts to the
customizable `plantuml-default-exec-mode`. The local var can be set by users
at will on a buffer local level, but without such an explicit opt-in then the
default value is always used
It is now possible to support many execution modes, currently available are:
- `jar`: the traditional way of running PlantUML from `plantuml-mode`
- `server`: ask a [`plantuml-server`](https://github.com/plantuml/plantuml-server) instance to run the preview (defaults to https://www.plantuml.com/plantuml)
* Added capability to automatically download the PlantUML jar path
As an effort to ease the first user setup, a new function is added to download the latest
PlantUML JAR release into `plantuml-jar-path` from Maven central.
Fixes#73
Now the latest version of PlantUML is downloaded as part of the CI script. This has two consequences:
we always test against the latest and greatest PlantUML version.
PlantUML evolves at a much higher pace than plantuml-mode. This allows for some sanity check of our code with the upstream project.
builds are not reproducible
As a side effect of always downloading a (potentially) new PlantUML version, re-running the build for an old commit can lead to different outcomes. This is usually a big no-no but given the limited efforts currently spent on plantuml-mode I find it more convenient to immediately raise a flag if a new version of PlantUML causes issues.
* Improving preview functionality. No buffer switch.
- Using temp buffer instead of new buffer.
- Windows layout is retained. No need to switch buffer after every preview
* fix version
* split indentation rexgexs
Split the regexs as a preparation for improvement and enhancement.
No explicit enhancement was intended, although the
`plantuml-indent-regexp-block-start' covers more blocks now.
* enhance and test regexs for groups
Changed `plantuml-indent-regexp-group-start' to cover all cases
mentioned on sequence diagram page of the plantuml website. Added
tests for all elements-with and without label (text following the
keyword).
* add support and test for box/end box blocks
* indentation test for a nested interface block
* correction of () at end of file
* indentation test for deployment diagram
* file “Cask”: added dependency to dash
* indentation tests for diagram commons
* add indentation tests for component diagrams
* regex corrections and added indentation tests for class diagrams
correct indentation regexs found by tests for class diagrams
* add indentation tests for object diagrams
* add indentation tests for state diagrams
* enhance indentation class diagram test: abstract class and interface
* move indentation tests to diagram test files
* mv test/plantuml-indentation-test.el -> test/plantuml-indentation-basics-test.el
* add indentation sequence diagram tests and mv basics tests to diagram tests
* add indentation for ref-groups
* improved indentation for ref-groups
* support indentation for activity diagrams - current/old style
* indentation support for notes
* rm not used regexs
* indentation support for plantuml macros
* correction: … → etc.
* fixed indentation code and added de-/activate to indentation
* added tests for indentation changes
* changed indentation code and tests
- more changes to plantuml-indent-regexs
- added more test
- changed function to determine indentation level and tests accordingly
- changed plantuml-indent-line and tests accordingly
- changed test names and doc strings to be more consistent
* more indentation tests, corrected regex for par block
- the par block usually does not have any label
- added database to start regex and added a test
- added block tests
Wrapping the defcustoms in eval-and-compile made the code compileable
but not customizable for the user.
Rewriting the start-process into a function instead of a macro makes it
possible to remove the eval-and-compiles.
