[etherlab-users] Limitations on Matlab Version - R2018b vs R2019a

[Moritz Schappler]

Hello Etherlab community,

I am trying to set up a new realtime PC and would like to use the
current version of Matlab/Simulink (and possibly future versions in the
next years).
I installed the etherlab toolchain as of March 24th 2020 from the source-forge hg
repos:
etherlab-code (1198 (4d38cecf1184); 2020-01-09), etherlabmaster-code
(2724 (fff0749b52f5); 2020-01-24) and pdserv (version 3.2.0, rev 651,
2020-01-10).

If I install everything on Matlab R2018b, the model compiles and can be
run. On R2019a (and R2020a), I receive errors and can not build (see log
at the end of the mail).
The log was created with a new Simulink model from scratch.
Is there a known compatibility list of versions of Matlab and Etherlab?
So does this work at for someone else? Then I would continue looking
into this.

Independent from the configuration in the question above, I also have a
patched version of EtherLab that removes pdserv and inserts the external
mode (see post in etherlab-dev mailing list).
This worked well until R2018b and did not work any more on R2019a (with
similar error messages). The patch is based on etherlab-code rev 1099
(2016-01-28).
I successfully rebased the patch until rev. 1142 (2018-01-18). Going
further would take a high additional effort due to the major changes in
hrt_main.c.
Since the whole purpose of the rebase is to get it running on the latest
version of Matlab (and potential future versions),
I would like to know, if this is possible in principle or if there are
major changes in Matlab R2019a declining this.

Thank you and best regards,
Moritz Schappler
(Research Assistant at the Institute of Mechatronic Systems at Leibniz
Universität Hannover)

P.S. Please forgive a potential double-post, as I am not sure if the same email from March 24th got delivered. 

> ### Creating
> /home/schappler/IMES/REPO/ethercat/etherlab-slaves/ixxat/new_2019a_etl_hrt/new_2019a.mk
> from /mnt/data/usr/local/MATLAB/R2019a/rtw/c/etherlab/etherlab_hrt.tmf
> ### Building new_2019a:
> /mnt/data/usr/local/MATLAB/R2019a/bin/glnxa64/gmake -f new_2019a.mk 
> INCLUDE_MDL_TERMINATE_FCN=1 ISPROTECTINGMODEL=NOTPROTECTING
>     CC  new_2019a_capi.o
>     CC  new_2019a_data.o
>     CC  new_2019a_header.o
>     CC  rtGetInf.o
>     CC  rtGetNaN.o
>     CC  rt_nonfinite.o
>     CC  ecrt_support.o
>     CC  new_2019a.o
>     CC  hrt_main.o
> g++      -o ../new_2019a new_2019a_capi.o new_2019a_data.o
> new_2019a_header.o rtGetInf.o rtGetNaN.o rt_nonfinite.o ecrt_support.o
> new_2019a.o hrt_main.o          -L/usr/local/lib  -lm -lpdserv -ldl
> -lrt -pthread
> new_2019a.o: In Funktion `new_2019a_step':
> new_2019a.c:(.text+0x15): Nicht definierter Verweis auf `ecs_receive'
> new_2019a.o: In Funktion `new_2019a_initialize':
> new_2019a.c:(.text+0x2c2): Nicht definierter Verweis auf `ecs_init'
> new_2019a.c:(.text+0x316): Nicht definierter Verweis auf
> `ecs_start_slaves'
> new_2019a.o: In Funktion `new_2019a_step':
> new_2019a.c:(.text+0x17c): Nicht definierter Verweis auf `ecs_send'
> new_2019a.o: In Funktion `new_2019a_terminate':
> new_2019a.c:(.text+0x356): Nicht definierter Verweis auf `ecs_end'
> collect2: error: ld returned 1 exit status
> gmake: *** [../new_2019a] Error 1Ž-

-- 
__________________________________________________________
M. Sc. Moritz Schappler
Wissenschaftlicher Mitarbeiter und Gruppenleiter
Robotik & autonome Systeme

Gottfried Wilhelm Leibniz Universität Hannover
Institut für Mechatronische Systeme
An der Universität 1, Gebäude 8142 (Etage 1, Raum 108)
30823 Garbsen

Fon: +49 (0)511 762-4116
Fax: +49 (0)511 762-19976

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20200406/d4c56cd1/attachment-0002.htm>