[Scilab-users] test_run failed: Slave Scilab exited with error code 127
Samuel Gougeon
sgougeon at free.fr
Tue Aug 6 15:38:57 CEST 2019
Le 06/08/2019 à 15:03, philippe a écrit :
> .../...
> I've just done this an get :
>
>
> -> test_run("ast",[],"no_check_ref")
> TMPDIR = /tmp/SCI_TMP_32597_VRKa5r
>
> Module 001/001 - [ast] :
>
> 001/154 - [ast] Sand.........................................
> ( /usr/-6.0.2//bin/scilab -nwni -nb -quit --timeout 15m -f
> /tmp/SCI_TMP_32597_VRKa5r/Sand.tst > /tmp/SCI_TMP_32
> 597_VRKa5r/Sand.res ) 2> /tmp/SCI_TMP_32597_VRKa5r/Sand.er
>
>
> the command line "/usr/-6.0.2//bin/scilab" should be instead :
> /usr/share/scilab-6.0.2/bin/scilab
>
> the problem seems to come from line 811
>
> if (getos() <> "Windows") & ~isfile(SCI+"/bin/scilab") then
> SCI_BIN = strsubst(SCI,"share/scilab","");//<---line 811
> else
> SCI_BIN = SCI;
> end
>
> there is two occurrences of "share/scilab" in the path so the
> substitution is not valid :
>
>
> --> SCI
> SCI =
>
> /usr/share/scilab-6.0.2/share/scilab
If this path specific to your installation? I mean, some other Linux or
oMacOS users run tests without this issue.
Anyway, this has to be fixed.
>
> the line 811 should be replaced with something like :
>
> SCI_path=strsplit(SCI,'/');SCI_BIN = strcat(SCI_path(1:$-2),'/');
>
> but this modification freeze scilab when launching test_run although
> running "/usr/share/scilab-6.0.2/bin/scilab -nwni -nb -quit --timeout
> 15m -f /tmp/SCI_TMP_32597_VRKa5r/Sand.tst" from a terminal is ok ...?
Have you checked that this is *actually* the run command, with the same
disp spy instruction?
If yes, then there may be 2 another bug after the first one.
>
> Thanks a lot Samuel for the help, do you want me to add a bug report ?
Yes please. It will need your contribution to test and valid the solution.
It will be hard (and heavy) to get a specific scilab installation to
test the patch.
Samuel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20190806/bd2dd7c1/attachment.htm>
More information about the users
mailing list