Hi everyone,
I hope you're doing well.
the 3.2 release is now almost ready. Sorry for the long hiatus that
elapsed before I could actually finalize the last features I planned.
mod_rivet test suite has been extended to allow testing with any MPMs
and in combination with different MPM bridges. The only non-default
meaningful test would be with the prefork MPM and the worker bridge.
Selecting a MPM causes mod_rivet to load the default bridge. This new
feature still need documentation, but I think I can write something
about it rather soon
-- Massimo
On 5/3/20 12:27 AM, Massimo Manghi wrote:
Hi guys, I hope you're doing well and if the pandemic reached your
whereabouts I hope you got around it uneventfully. In Italy the pandemic
hit very hard, especially here in the densely populated north where a
large portion of the economic and industrial capacity of the country is
located. I live not far from Lombardy where the pandemic hit very hard.
Given the intense communication with this region we too had a relatively
large number of casualties. Things are now steadily improving, but the
world ahead of us is going to be very different....
Anyway, being stuck at home most of the time gave me the chance to work
on the 3.2 version (which is currently the master branch). I think I'll
work on it next week trying to improve the test system, but I think the
module is ready.
New features:
- Threaded bridges (such as the worker bridge) cried loud for the
possibility of having single threads terminate by calling ::rivet::exit
without affecting other threads. This feature can be enabled by setting
the SingleThreadExit configuration parameter. The prefork bridge is
unaffected
- New command ::rivet::thread_id that returns the thread
identification in a format that emulates command ::thread::id (saving
the need to load the Thread package in order to get this information).
The command returns the thread id also as a decimal number, which is the
way it is printed in the Apache error log
bug fixes
- The lazy bridge has been simplified by removing some counters and
variables used during its development for testing. Also the access to
some variables has been protected by requiring lock on a mutex.
- Rivet_Upload modified. The upload command crashed under some
circumstances (invalid argument list) due to wrong argument checking.
Functions in src/mod_rivet_ng/Tclwebapache.c called by Rivet_Upload are
now responsible to return a value to the interpreter. Even though this
required the task of handling the result to be written within each
function, my opinion is that this approach allows a more precise
handling of Tcl objects. All of these functions already use Tcl status
codes in order to communicate with the caller, Rivet_Upload now simply
returns the code they return without further processing
-- Massimo
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscr...@tcl.apache.org
For additional commands, e-mail: rivet-dev-h...@tcl.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: rivet-dev-unsubscr...@tcl.apache.org
For additional commands, e-mail: rivet-dev-h...@tcl.apache.org