thank you Harald, I never exploited that bizarre name resolution and
luckily never bit me with some unexpected bug 🥶. I try to keep the
global variables to the smallest possible set, but whatever this
paradigm shift broke it definitely had to change
happy conference in Vienna...
-- Massimo
On 7/16/26 22:40, Harald Oehlmann wrote:
Great that you have found it !
Yes, the "bugfix" that v in
proc a {} {set v 1}
does not set v if it exists at global level breaks a lot but is really
necessary...
Thnks for all and greetings from the TCL conference in Vienna,
Harald
Am 16.07.2026 um 17:47 schrieb Massimo Manghi:
A new rivet RC is ready from
http://www.rivetweb.org/~manghi/rivet/rivet-3.0.0rc2.tar.gz
an incompatibility was spotted in the way the informational
array ::server is generated and due to the different variable names
resolution scheme adopted in Tcl9 it caused a regression
  -- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
