SVGHMI: Work in progress. C side mostly implemented, neither built nor tested.

SVGHMI: More strict formating of active HMI labels : @ separator only used for paths

SVGHMI: new test svg to drive next implementation

Non significant changes in generated xslt, side effect of changes in yslt_noindent.yml2

Merge default in SVGHMI branch

SVGHMI: Starting to define JS side more in details.

SVGHMI: added transformation of code given in inkscape's description fields into javascript functions

SVGHMI: gui.svg -> svghmi.svg

SVGHMI: moved static JS code to a separate file included at xhtml generation time

SVGHMI: Work In Progress : fixed pointer types in ctypes interface, cleaned up server startup and cleanup code, changed document type to XHTML, cleaner JS script : encapsulated in a function and in CDATA.

PLCobject: Call "stop" and "cleanup" methods from python runtime files in reverse order compared to "init" and "start".

SVGHMI: WIP for python<->C data exchange : message from browser hit the C side.

SVGHMI: WIP for python<->C data exchange

SVGHMI: WIP for python<->C data exchange

SVGHMI: WIP for python<->C data exchange

SVGHMI: single session for now

SVGHMI: few fixes on serving

WIP on svghmi, now builds and runs. HTTP serving + WS transport ready, missing actual data to transmit and thread to collect it.

Non significant changes to mute GCC warnings about unused variables.

Fixed sequel of 5f79b194fa63 'SVGHMI: filter out temporary variables created while generating ST code out of FBD.'. Some ref to temporary variables are still not using the _TMP_ prefix. They are quite hard to find, so to be continued...

SVGHMI: Intermediate state while working on svghmi.c

svghmi.c : deduplicated variable access code borrowed from plc_debug.c. Added targets/var_access.c.

SVGHMI: Added iterators in svghmi.c copy-pasted form plc_debug.c

SVGHMI: svghmi.c now has mutex, iterator, and read/write buffer.

SVGHMI: draft for svghmi.c. It has all PLC variables pointed in HMI tree in an array.

Pass HMITree to SVG transform. It seems it could really help to reduce JS tree binding logic in the end.

Add special nodes at HMI Tree root, fix code to handle special node (no path).

Swap CTN code gen and Libraries code gen, because SVGHMI use both, and need the library part to produce hmitree before CTN part.

yslt_noindent : Avoid YSLT's indent logic to interfere when using linequote.

SVGHMI: make a reasonably basic test

SVGHMI: filter out temporary variables created while generating ST code out of FBD.

SVGHMI: added deduction of HMI tree from list of HMI_* instances.

SVGHMI: added extraction of SVG bounding boxes, obtained from "inkscape -S", and passed to XSLT transform as variable.

Allow using docutil.get_inkscape_path() on Linux also.

SVGHMI: little fixes to make the build go through

SVGHMI: Added XSLT transformation, Makefile to get XSLT from ysl2 (copy of plcopen/Makefile) and a minimal stylesheet to start with.

Moved XSLT model query python code so that XSLT part can be reused for other transformations (i.e. in SVGHMI)

Complete SVGUI test minimal skeleton so that it builds.

intermediate commit, work in progress

Use a POU Library's Generate_C to collect all variables in SVGHMI.

merged

Enable svghmi plugin in features, and small fixes to skeleton.

Merged default in svghmi

Skeleton for svghmi extension

IDE: Allow POU Libraries to inject PLC Configurations globals during build.

WIP: initial setup.

IDE: Fixed variable traces graphs RingBuffers. Removed an apparently useless wxCallAfter in trend graph that was leading to pydeadobject exception on wxGTK when double-clicking.

IDE: Cleaned up some useless tests in variable trace data handling code, changed from bare numpy arrays to RingBuffers inorder to avoid RAM outage and crash after long tracing session.