Experimental Physics and
Industrial Control System

Tim Mooney <[email protected]> · Fri, 23 Sep 2011 12:47:56 -0500 (CDT)

Pierre,
Probably you should look at the aSub record code, in base/src/rec.
The aSub record doesn't make a 2D array of strings as an array of
pointers to arrays of characters.  It makes a 1D array of (for
example) NOVA*MAX_STRING_SIZE characters, and sets VALA to the
address of that array.
Tim

----- Original Message -----
> From: "Perrier Pierre" <[email protected]>
> To: "Mathias Steiner" <[email protected]>, [email protected]
> Sent: Friday, September 23, 2011 11:45:49 AM
> Subject: RE: Asub record problems to update waveforms of string-type
> Thanks for your help. Indeed, now my waveform is updated at each call
> of my "asub_notify" function. Unfortunately, I've another problem,
> strings in the waveform are completely wrong (ex : "&@ñ««««") like a
> pointer error. So, I think the problem comes from the cast of
> precord->vala , what do you think about it ? I don't know what is the
> type of vala. Does anyone know ?
> Maybe someone has an example of "asub" record process function which
> updates waveforms of string-type.
> 
> Cheers,
> 
> Pierre Perrier
> 
> 
> 
> -----Message d'origine-----
> De : Steiner, Mathias [mailto:[email protected]]
> Envoyé : vendredi 23 septembre 2011 16:47
> À : Perrier Pierre
> Objet : RE: Asub record problems to update waveforms of string-type
> 
> Pierre,
> 
> This may only be tangential to your problem, but I have found that
> when my aSub records update an 'analog out' variable, the value
> immediately updates everywhere.
> 
> OTOH, when updating strings or waveform strings, they only "sort of"
> update.
> 
> A caget will provide the updated string value, but EDM will show the
> old value. This is even true when the EDM window is killed and
> restarted -- which I don't understand at all.
> 
> Adding FLNK, "$(P):CRExeSeq.PROC" to your aSub, though not pretty, may
> fix the immediate problem.
> 
> Cheers -Mathias
> 
> 
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Perrier Pierre
> Sent: Friday, September 23, 2011 10:01 AM
> To: [email protected]
> Subject: Asub record problems to update waveforms of string-type
> 
> Hello,
> 
> I have some problems with the use of the “asub” record. I’m trying to
> update a waveform record with two strings. So, my EPICS database looks
> like this:
> 
> record(waveform, "$(P):CRExeSeq") {
> field(PINI, "NO")
> field(DTYP, "Soft Channel")
> field(UDF, "0")
> field(NELM, "2")
> field(FTVL, "STRING")
> }
> 
> record(aSub, "$(P):_notifyCRExeSeq") {
> field(SCAN, "Event")
> field(EVNT, "4")
> field(OUTA, "$(P):CRExeSeq")
> field(SNAM, "asub_notify")
> field(PINI, "NO")
> field(FTVA, "STRING")
> field(NOVA, "2")
> field(NEVA, "2")
> }
> 
> The “$(P):CRExeSeq” record is updated by the “$(P):_notifyCRExeSeq”
> record when an observer class executes the “post_event(4)” function.
> The problem is that my “asub_notify” function doesn’t really update
> the string waveform. In this function, I process the update in the
> following way:
> 
> // I get the new informations from an other class
> QStringList l_infoAnotifier =
> g_drvSequenceurMetier->getCRExecutionSequence();
> 
> //Here, I process the update
> if(precord->nova <= l_infoAnotifier.size()){
> // I create a char**
> char** l_info =
> (char**) malloc(l_infoAnotifier.size()*sizeof(char*));
> 
> for(int i=0;i<l_infoAnotifier.size();i++){
> l_info[i] = (char*) malloc(40*sizeof(char));
> if(l_infoAnotifier.at(i).length()<=40){
> strcpy_s(l_info[i],40,l
> l_infoAnotifier.at(i).toStdString().c_str());
> } else {
> strcpy_s(l_info[i],40,"Error too much char");
> }
> }
> //Copy
> try {
> char** l_stringout = (char**) precord->vala;
> memcpy_s(l_stringout,(40*precord->nova)*sizeof(char),
> l_info,(40*l_infoAnotifier.size())*sizeof(char));
> } catch (...) {
> printf("Error in the asub_notify function\n");
> }
> 
> //Delete
> delete[] l_info;
> }
> 
> In the same way, I created a process function which seems to work for
> updating waveform of double-type data. Here, I think the problem comes
> from the string type and the “memcpy” functions.
> Does anyone know how to realize asub functions for string waveforms?
> Or know why this code doesn’t work?
> 
> 
> 
> Cheers,
> 
> 
> Pierre PERRIER
> Stagiaire - MCO LIL
> Sopra group.
> Parc Innolin
> 9 rue du Golf 33700 Merignac
> Phone : +33 (0)5 56 12 79 59
> [email protected] - www.sopragroup.com
> 
> Ce message peut contenir des informations confidentielles dont la
> divulgation est à ce titre rigoureusement interdite en l'absence
> d'autorisation explicite de l'émetteur. Dans l'hypothèse où vous
> auriez reçu par erreur ce message, merci de le renvoyer à l’émetteur
> et de détruire toute copie.
> 
>  Pensez à l’environnement avant d’imprimer.

-- 
Tim Mooney ([email protected]) (630)252-5417
Software Services Group (www.aps.anl.gov)
Advanced Photon Source, Argonne National Lab

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System