On Windows 'send' is disabled because it caused
interpreter hangs. However, this leads to a test failure.
Ideally win-send should be fixed, but failing that the
test should have a non-Windows constraint applied.
==== safe-1.2 Safe Tk loading into an interpreter FAILED
==== Contents of test case:
catch {safe::interpDelete a}
safe::interpCreate a
safe::loadTk a
set l [lsort [interp hidden a]]
safe::interpDelete a
set l
---- Result was:
bell cd clipboard encoding exec exit fconfigure file
glob grab load menu open pw
d selection socket source tk_chooseColor
tk_chooseDirectory tk_getOpenFile tk_ge
tSaveFile tk_messageBox toplevel wm
---- Result should have been (exact matching):
bell cd clipboard encoding exec exit fconfigure file
glob grab load menu open pw
d selection send socket source tk_chooseColor
tk_chooseDirectory tk_getOpenFile
tk_getSaveFile tk_messageBox toplevel wm
==== safe-1.2 FAILED
Logged In: YES
user_id=79902
A failing test isn't exactly a disaster.
Logged In: YES
user_id=32170
Certainly not a disaster -- but when I run Tk's tests and I
get some 20 different failures, how am I to tell which are
important and which aren't?
I happen to know that this one is pretty harmless, but
that's still no reason not to resolve it (esp. given I can't
see any action on fixing Windows 'send').
Logged In: YES
user_id=80530
Originator: NO
Tk 8.5b1 is out.
It's time to get [send]
on Windows re-enabled and
working properly, or it's
time to officially revert it
from the new feature list.
Logged In: YES
user_id=80530
Originator: NO
I think this has been resolved
in the sense that [send] on Windows
has been pulled from the feature
list for Tk 8.5.0.
Leaving open for maintainer resolution,
but dropping from release blocking
priority.