[Scilab-Dev] TCL : Milestone and further investigations.

[fvogelnew1 at free.fr]

Selon Bruno JOFRET <bruno.jofret at inria.fr>:
> It fixes BUG 2514 and 2640 (for me, do you all agree ?).

I think it does, I remember having checked some weeks ago (was however not
perfect I think), but I can't check on Linux right now since I'm far away from
home.


> - Recursive commands using ScilabEval & TCL_EvalStr are working.

I'n not so sure. Look at this (in debug(-1) mode so that you can see the output
of the queueing and unqueueing processes):

-->TCL_EvalStr("ScilabEval a=10")    <<<hit enter once here
TCL_EvalScilabCmd a=10              <<<this gets displayed: command is queued
 ans  =

               <<<no return value:normal for the ScilabEval instruction

-->            <<<At this point the command is not yet executed. Hit enter!
   Unqueuing a=10 - No option    <<<the queued command gets executed only now!

-->


Btw, in Scilab5 the width of the debug(-1) output has become much narrower than
it was in Scilab 4.x. Doesn't help readability and should be fixed too.


Moreover:
-->TCL_EvalStr("ScilabEval TCL_EvalStr(""pwd"")")
TCL_EvalScilabCmd TCL_EvalStr("pwd")
 ans  =



-->    <<< hit enter here, apart from this, this is OK below
   Unqueuing TCL_EvalStr("pwd") - No optio
     (end) n

BUT:

-->TCL_EvalStr("ScilabEval TCL_EvalStr(""set a 1"")")
TCL_EvalScilabCmd TCL_EvalStr("set a 1")
 ans  =



-->    <<< hit enter here
   Unqueuing TCL_EvalStr("set - No option
TCL_EvalStr("set
             !--error 31
Incorrect string.

while executing a callback


-->

Apparently a problem when the string passed contains spaces. Which means I think
you're wrong when you say that recursive commands using ScilabEval & TCL_EvalStr
are working, such as TCL_EvalStr("ScilabEval TCL_EvalStr(""ScilabEval a=7"")")


> - TCL_EvalFile :
> There is a simple solution,

A simple solution to what problem? Why doesn't it work any more? Why is the
current code not compatible with your changes? This looks strange to me since
it's a similar code for TCL_EvalStr and you state that the latter is working.


> I hope I did not forget any remarks/question.

Did I miss your answers to my email somewhere in January where I asked a number
of wuestions about what you're doing exactly (I don't have this email at hand
sorry, not even its send date)?


Francois