Line numbers now physical in setbpt ?
François Vogel
fvogelnew1 at free.fr
Sun Jan 18 22:18:01 CET 2009
Hi all,
I have seen no reply to my message below, could Serge please tell me
what's the situation on this subject?
Thanks.
Francois
François Vogel said on 11/01/2009 17:50:
> Hi,
>
> I'm slowly realizing that apparently in Scilab 5 all line numbers are
> now physical line numbers while they were logical line numbers up to
> Scilab 4.1.2.
>
> I think it's now clear from bug 3407 that line numbers reported in error
> messages and in lasterror() are now physical (were logical in Scilab 4.x).
>
> What I'm now realizing is that also what is managed by setbpt/delbpt
> must now be physical and no more logical.
>
> Could Serge please confirm this statement?
>
>
> Example (run it in Scilab 4.1.2 and in Scilab recent master to compare):
>
> function pllines
> disp("line 2 2"); // logical physical
> disp("line 3 3")
> disp("line 4 4")
> disp("line 5 5")
> disp..
> ("line 6 7")
> disp("line 7 8")
> disp("line 8 9")
> disp("line 9 10")
> endfunction
>
> setbpt("pllines",6)
> pllines
>
>
> Scilab 4.1.2 ends with:
>
> line 6 7
> Stop after row 6 in function pllines :
>
>
> -1->
>
> I.e. Scilab 4.1.2 executed the full continued logical line 6 before
> stopping.
>
> Compare with the output in Scilab 5:
>
> line 5 5
> Stop after row 6 in function pllines.
>
>
> Type 'resume' or 'abort' to return to standard level prompt.
>
> -1->
>
> Here we can see that it does NOT executed logical line 6 completely,
> thus stopping in the middle of it. However, it nevertheless states that
> it stopped after execution of line 6.
>
>
> Further observation: add setbpt("pllines",10) and you'll see it stops in
> Scilab 5, meaning line 10 is still in function pllines. In Scilab 4.1.2
> it doesn't stop because line 10 is after the end of the function pllines
> definition.
>
>
> Thanks for clarifying this. If setbpt/delbpt are now really supposed to
> deal with physical line numbers, then I'll have to make changes in the
> Scipad debugger for Scilab 5.
>
> Francois
More information about the dev
mailing list