diff options
author | dgp <dgp@users.sourceforge.net> | 2009-01-12 16:50:03 (GMT) |
---|---|---|
committer | dgp <dgp@users.sourceforge.net> | 2009-01-12 16:50:03 (GMT) |
commit | 91fcbdbc1ce2836f8df968af33ce13bff991a90b (patch) | |
tree | 166fd653d036a82b765eb37770c9ae92a504164e /tests/error.test | |
parent | 8b9a703b9cb4b40327681b31f0b3c1c2ed01cea5 (diff) | |
download | tcl-91fcbdbc1ce2836f8df968af33ce13bff991a90b.zip tcl-91fcbdbc1ce2836f8df968af33ce13bff991a90b.tar.gz tcl-91fcbdbc1ce2836f8df968af33ce13bff991a90b.tar.bz2 |
* generic/tclBasic.c (Tcl_DeleteCommandFromToken): One consequence
of the NRE rewrite is that there are now situations where a NULL
objProc field in a Command struct is perfectly normal. Removed an
outdated comment in Tcl_DeleteCommandFromToken that claimed we
use (cmdPtr->objPtr == NULL) as a test of command validity. In fact
we use (cmdPtr->flags & CMD_IS_DELETED) to perform that test.
Also removed the setting to NULL, since any extension following the
advice of the old comment is going to be broken by NRE anyway, and
needs to shift to flag-based testing (or stop intruding into
such internal matters). Part of [Bug 2486550].
Diffstat (limited to 'tests/error.test')
0 files changed, 0 insertions, 0 deletions