Difference between revisions of "EDT:The Differences between EDT 0.8.2 and EDT 0.8.1"

Revision as of 11:38, 20 November 2012

EDT 0.8.2 contains some language changes, so existing code might need to be modified.

It is recommended that you use a new workspace after upgrading to 0.8.2 M3 or later. You can import your existing projects into the new workspace. You may see some odd issues if you do not do this, such as EGL Search queries not finding all matches.

• Only the package "eglx.lang" from the system eglars is automatically imported. All other types from the system eglars must be imported explicitly in the code.
  • Fix: Right-click on your projects >  EGL Source > Organize Imports
• Annotation syntax is now more strict. For simple annotations with a single field we used to allow it as: myAnnotation = "some value". This is no longer supported.
  • Fix: Manually change the syntax to: @myAnnotation{"some value"} (or you can explicitly name the field, @myAnnotation{myField = "some value"})
• Timestamp patterns are now case-sensitive. Month must be uppercase 'M', minute must be lowercase 'm', hour must be uppercase 'H'. The full pattern is: yyyyMMddHHmmssffffff
  • Fix: Manually change invalid patterns to the correct case
• The field named "class" in Widget has been renamed to "cssClass".
  • Fix: Manually change references to this field to the new name.
• New versions of the widget projects need to be used. The previous versions of the projects will contain compile errors.
  • Fix: Delete the existing widget projects, create a new project to import the new version of the projects, and update the EGL Path of the projects that referenced the old widget projects. 

The following changes were made between 0.8.2 M3 and 0.8.2 M4.

• The precision and decimals functions were moved from MathLib to the numeric data types.
  • Fix: Change MathLib.precision(x) to x.precision(). Change MathLib.decimals(x) to x.decimals().
• The second parameter of MathLib.modf changed from bigint to float.
  • Fix: Existing code will continue to work since bigint and float are compatible, but the value assigned to the parameter may be slightly different than before.