I have made a refinement in commit 972b0b5 to still warn if no explicit language was set. I believe the false positives
will be lower here, since lines are less likely to end with colons for completely separate reasons.
This is about balancing the known false positives against any use of this alternative engine specification facility.
This problem (false positives because of : in the first line) has come up a lot in the past as well as now. On the
other hand. are you (or anybody else on this list) aware of any current cases of people using this ability to specify a
different script engine? I'm not currently aware of any.
The only other current script engine implementations I'm aware of are apart from XEngine are DNE (long discarded from
core), ForthMinus (abandoned by Alondira over 4 years ago), some other DNE derived engine that was long abandoned and
your own proprietary engine. I don't count engines running on forked codebases since they have moved away from
OpenSimulator compatibility.
To be honest, I think this is an over-specified feature that was only brought in originally to avoid making a decision
about OpenSimulator's script engine. I find it hard to believe that anyone would want to deal with the complexity of
running some scripts in one engine and other scripts in another in the same simulator, just for some theoretical
performance advantage. I rather doubt you're running 2 engines for the same language since that would be a nightmare to
explain to users (happy to be corrected on this point).
Really, I think we should adopt a better way of specifying the script language/engine than looking for a : than be stuck
with this issue forever.
On 28/06/12 02:50, Melanie wrote:
I'm afraid that can't be done this way. Selecting an engine without
giving it an explicit language, e.g. // SCEngine: is legal.
Therefore, with this patch, the defaulting to XEngine is no longer
reported int these cases. Since the reporting was introduced to
ensure that scripts are not run on engines with lesser
functionality, the entire purpose of the message is completely
defeated by this. With this patch, we might as well not notify users
at all and just let them crash and burn.
Melanie
On 28/06/2012 02:49, Justin Clark-Casey wrote:
Okay Edmund, commit 25baa2d contains a change which should eliminate reporting
of these false positives. Please give it
a whirl.
On 26/06/12 01:38, Edmund Edgar wrote:
On 26 June 2012 08:58, Justin Clark-Casey <[email protected]> wrote:
I don't like the way script language/desired engine detection was done - a
clash with existing scripts could have been avoided by requiring some
special preceeding chracters rather than simple c#, lsl:DotNetEngine, etc
(like #! in shell scripts, for instance).
Agreed, but on the assumption that we're stuck with this convention
now, would it be possible to tighten up the pattern matching a bit? It
feels like the system should be able to tell the difference between
//c#
...and
// This is a comment consisting of multiple words. What it has to do
with script engines: Nothing.
Just checking for a colon seems a bit broad. One of the mind-numbing
busywork tasks on my TODO list this morning is cleaning up some
scripts somebody checked in since I last fixed our iar. Their editor -
I think it's the Eclipse plugin - has helpfully added:
// LSL script generated:
furniture-1.0.object_definitions.presentation_seating.lslp Sat Nov 19
14:36:29 Tokyo Standard Time 2011
_______________________________________________
Opensim-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/opensim-users
--
Justin Clark-Casey (justincc)
http://justincc.org/blog
http://twitter.com/justincc
_______________________________________________
Opensim-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/opensim-users
