by
Matt NewmanApologies to our readers for last weeks hiccup, our normal quality controls are back in place!The news group has been very busy this week (> 650 msg's), dominated by Tcl 8.1.x bugs and performance concerns.The Tcl 8.1.x thread on performance wins this week's award for the longest single thread in one week (over 70 messages). I encourage anyone who is seriously planning to migrate soon to check out this thread, see below for some highlights.- Main thread on Tcl 8.1.x performance. (see [1])
- Backgrounder on Tcl Performance. http://purl.org/thecliff/tcl/wiki/TclPerformance

- Nice explanation of the difference between the Tcl 8.0.x and Tcl 8.1.x byte compilers. (see [2])
- Real-world impact of Tcl 8.1.x changes - string index is no longer O(1) (see [3])
- Y2K False Alarm (2000 is a leap year). Should we be concerned about the Y2038 bug though? (see [4])
- Interesting outcomes of variable substitution. (see [5])
- Geometry managers (pack/grid) and reqwidth. http://216.71.55.6/cgi-bin/wikit/443.html

- A suggestion for system-wide and per-user options (for tcl and tk) that would map across various platforms. (see [6])
- Example of serial communications using Tcl. (see [7])
- Gethostbyname in Tcl. (see [8])
- Much needed but minor improvement for people who care about sub-second timing in Tcl. (Proposal) (see [9])
- New Socket Driver for windows in 8.1.1 - calling for willing testers. (see [10])