Sherpa Prototype

output & tmp variables have now their init values from the header model used for evaluation in evalProgram, more specifically varsExtract. The header's view makes these editable. Presentation of the header's view has been modified a bit.

oops forgot to make the tests descriptions editable too

Tests are editable now. Bit or refactoring in prepareEditables, still much work needed there. In addition prepareEditables & genSimpleHandler need to be factored since they are being used in both TestsView.js & HeaderView.js

programs headers now have a doc property for the author to explain what the program is about. it's editable which led to some refactoring of prepareEditables. Still some improvements needed there but for now it's working
new logError function in org.sherpa to make sure the errors we want to deliverto the programmer will be redirected later to a specific div

Improving JS quality; removing the indexEA index in renderHeader & renderVarList by using editableArray[editableArray.length] to keep adding new elements to the end of this array

fixed missing of 2nd parameter to editable in prepareEditables from HeaderView.js which led to some layout issues when editing

adding some documentation

Essentially updating the header / tests / programs models when variable names are modified. Updating the header model when variables roles are modified.
Functions are in sherpa.js for now; get it working now, refactor later
FIXED - nasty bug in evalProgram / evalStatement from edits.js which led to a JS infinite loop apparently due to so local variables not defined with var

Refactoring but nothing really big

Some refactoring;
- get rid of page_initialize non sense by merging it with the document ready JQuery function
- start migration to unobtrusive JS programming techniques; e.g. global variables / global functions
- break down edit.js into files to host the views for headers, programs & tests
- add a new tab to host the tests results which are still rendered by renderEvaluatedTests in edit.js until it's moved into the model parts of the program. The "tests" tab is using a partial version of this function which does not evaluate the program. Trying to keep edit separate from run / get tests results so it might be easier later on to log each separate run attempts; this will most likely evolve quickly