On 4/18/02 10:49 AM, "Jim Wilcoxson" <[EMAIL PROTECTED]> wrote:
> This test is a > 10-15 line TCL loop to read through a file containing around 500K lines, > do 3-4 string operations, set an ns_share array entry Can you try the same test with an nsv instead of an ns_share? > I'd gladly move if the benefit is there. I don't know that I'd agree that there's no benefit to using 8.x. I know there were some benchmarks early on in the history of 8.x support that showed a significant slowdown with 8.x, but I also saw some more recent ones, and I think a lot of the performance advantage of 7.6 has been reduced. Your own tests appear to show that either 7.6 or 8.x has a big advantage over the 7.4 in 2.3.3, although they do show 7.6 having more of an advantage. There are lots of benefits to using a mainstream Tcl. For one, current Tcl can handle binary data; while there are still issues with moving this into and out of AOLserver requests, you can manipulate images from within Tcl, which is sometimes handy. Also, I just got to rewrite something that used ns_sock and ns_select functions, because there was no way to do a non-blocking read using those functions, whereas using the Tcl 8.x native socket support worked a treat, plus I could test the functions outside of AOLserver, using just a tclsh. There are more 3rd-party tcl libraries for 8.x that you can take advantage of, which is getting very handy for doing XMLish stuff, since a lot of people are working on this for 8.x (and no one I'm aware of is working on it for 7.6). 8.x is also a little more than a few months old, so I don't know that it really fits in the "newfangled" category, as much as that 7.6 is approaching the paleolithic category. I recognize that it's not a simple drop-in-and-go to move to 8.x; I've had the migration problems before, where we had a bunch of customers stuck with 2.0 for the longest time and couldn't easily get them all moved to 2.3.3. I feel your pain, but I think there are benefits for starting to move toward 8.x now, where possible. And I think that the performance issues can be worked on. Jeff Hobbs is really helpful with performance things, so if it turns out to be a Tcl bottleneck, I'll bet he either has an alternative, a workaround, or will get a fix scheduled for an upcoming 8.x -- you won't be able to expect scheduled fixes for 7.6. Pete.
