Subject: UPDATE: 8.3alpha development report - DN [1]
Jeffrey Hobbs <jeffrey.hobbs@scriptics.com> - 20 Oct 1999 - comp.lang.tcl
For those who have an interest in the current status of 8.3,
here's a little report. Several changes have occured in the
past couple of weeks that might make it interesting for some
to test the NetCVS dev source tree.
* There are about 330 lines of changes since 8.2.0 (some were
back-ported to 8.2.1, most not).
* There is an entirely new regular expression engine, courtesy
of Henry Spencer. This one should use much less stack space.
* I have gone through and Purified all of Tcl on Unix. I can
now run the full range of tests, threaded or non-threaded,
without a single leak. There are some UMRs and such, but these
seem to be innocuous. Of note, I plugged a leak for each created
process on Unix, as well as leaks for each thread were cleaned up.
* new 'file channels' method
* a lot of clean up in the test suites
* new 'array unset' method
* new 'clock clicks -milliseconds' switch
* new Tcl_AppendResultVA (should be happy for LinuxPPC now)
* fix for insidious crash when releasing literal table
* fix for event/io problems with threads
* a few other things (see ChangeLog)
Tk didn't see much in this last haul, as I was focused on
Purify'ing Tcl (bring on the Spanish Inquisition!).
In the coming month, I plan on running Purify over Tcl on NT,
and focusing more on getting submitted patches into Tcl/Tk for
8.3 (one I've already mentioned, like Img and Dash).
I wanted to let you know about this "milestone", because I
consider the current CVS sources (for 10/20/99) to be very
suitable for testing purposes. 8.3 won't have any special alpha
releases as I add new features. I want to get what I can in and
make a 8.3b1 as soon as I can. Skipping full alpha releases adds
a lot more time to my work schedule for adding new things.
--
Jeffrey Hobbs The Tcl Guy
jeffrey.hobbs at scriptics.com Scriptics Corp.
Last modified
1999-11-04
1999-11-04
(195.108.246.50)
Note: you are looking at
the snapshot of an old wiki
- much of this information
is likely to be very outdated
