Hi Massimo,
It is a pity that you cannot reproduce it. May be it is a problem in my
system setup. Or is it related to TCL 8.4.19? Which TCL version are you
using?
I'll have a look to 'ab'. I did tests using a shell and wget, e.g.:
for i in {1..10} ; do wget -O- -q -S
"http://localhost:8082/sventemp/AAA/handleAAA.rvt"; ; done
for i in {1..10} ; do wget -O- -q -S
"http://localhost:8082/sventemp/BBB/handleBBB.rvt"; ; done
Please note I simplified the script a little bit more (one procedure
with regsub and source) to:
<?
proc procB { } {
source include.tcl
set args {}
regsub "astring" "anotherstring" "globalvar_local" args
}
procB
?>
May be it is possible to make it more simple. But for now it seems that
I've reached a local(?) minimum.
I hope I can test it today on a different system. I will let you know.
Best regards
Sven
Am 19.01.2012 00:23, schrieb Massimo Manghi:
Tested it and everything seems to be fine here.... (both when
ServerLimit set to 1 or left to its default)
An alternate way to test your scripts is by using the apache utility
'ab' (Apache Benchmark). It can push the server to its limit and
therefore possible bugs that show up under some extreme circumstances
might be reproduced.
Anyway I don't understand why the nested procedures ProcA and ProcB in
handleAAA.rvt should be decisive to cause the bug to occur. Is it the
symptom of some problem in your set up? Subtle problems with some
library?
On mer, 2012-01-18 at 22:32 +0100, Sven wrote:
Hi Massimo,
I attached a "project" where I can reproduce it. Maybe there are some
other things that are important I did not mention yet (hopefully not).
There are some important things:
1) that quirky code for at least one rvt file: see AAA/handleAAA.rvt in
the attachment
2) the [pwd] in handleAAA.rvt seems to prevent the problem to occur!!!
So I comment it out.
3) the count of serving Apache instances/children shall be low enough:
'ServerLimit 1' was perfect for this purpose
Point 2) indicates that there is some issue with the TCL interpreter.
Isn't it?
Everything is possible, but I don't think this is the most plausible
hypothesis. The interpreter could have (and certainly has) bugs o
unwanted side effects but I don't expect them to occur in a basic
command like 'pwd'. I recommend you take a look at 'ab' and do some more
tests. Let me know.
-- 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
