[Dltk-dev] Bug or inconsistent implementation: IInterpreterInstallType.v

[Gabriel Petrovay <gabriel.petrovay@xxxxxxxxxx>]

Hi,

There is a problem with the way the method
IInterpreterInstallType.validateInstallLocation(IFileHandle
installLocation) is used in DLTK.

On one hand, the AddScriptInterpreterDialog class calls this with a
file reference (when you browse for an interpreter). Therefore all the
implementation (Tcl, Ruby or my plugin) use the following code to
validate the installLocation:
		if (!installLocation.exists() || !installLocation.isFile()) {
			return createStatus(IStatus.ERROR,
					InterpreterMessages.errNonExistentOrInvalidInstallLocation,
					null);
		}

On the other hand, when using the
org.eclipse.dltk.launching.interpreterInstalls extension point the
ScriptRuntime.addInterpreterExtensions() uses the same validation
function to validate the extension point's "home" attribute, which is
supposed to be, according to the description: "Path to the home
installation directory"


This makes the logic of the validateInstallLocation of the install
types fuzzy in my case and buggy in the Tcl or Ruby cases if I wans to
provide my own installs.


Regards,
Gabriel

PS: Hey, I know you got pissed off because of my many emails, but I
think they are helpful. And I am happy to help. Also, all my problems
form a good database in the archive for others to see if they
encounter similar problems.

-- 
MSc Gabriel Petrovay
MCSA, MCDBA, MCAD
Mobile: +41(0)787978034