Hi all, sorry for double-posting, but I found the fix. For me the code does not look that 'sexy' now but it works. I do not know if I can make it better but for now I will stick to this :) I pushed it to the 'eip' branch of my fork. Have a nice weekend,
Etienne On 2020/03/13 14:48:35, Etienne Robinet <[email protected]> wrote: > Hi, > Thanks for the advice. I am trying to pass the length down the subclasses, > but I'm stuck. This does not work as it seems: > https://i.imgur.com/77bbdBN.png CipRRData does not know what 'len' is nor its > value as it seems. > > I wanted to inspire mysefl from the CotpPayload, but unfortunately, the first > byte of the whole packet is a discriminator (command). > > Etienne > On 2020/03/13 13:52:16, Christofer Dutz <[email protected]> wrote: > > Hi Etienne, > > > > I think you can solve your problem in two ways. > > In all you need to pass down the length of the packet in total from the > > root type (which knows the length). > > 1) You keep on just reading bytes and parse in the protocol logic class > > (Like in the S7 driver or KNX) > > 2) You directly parse PlcValues (using "dataIo" types) > > > > I would prefer option 2 but 1 is simpler. The reason I'm doing 1) in S7 and > > KNX is that I need to know the type from the request to decode in the S7 > > case and in the KNX case I need to know the type from an external source in > > order to decode it. > > > > > > Chris > > > > Am 13.03.20, 14:45 schrieb "Etienne Robinet" <[email protected]>: > > > > Yes this is exactly what I need. If I get the remaining bytes, I can > > know the element numbers as I have their type! > > > > Etienne > > > > On 2020/03/13 13:40:09, Julian Feinauer <[email protected]> > > wrote: > > > Hey, > > > > > > there ist he possibility to get the remaining size oft he bytes of > > the message. Does that help? > > > Otherwise I misread your question. > > > > > > Julian > > > > > > Am 13.03.20, 14:37 schrieb "Etienne Robinet" <[email protected]>: > > > > > > Hi Julian, > > > so how should I declare this field in the mspec if I can not get > > its size? > > > Thank you, > > > > > > Etienne > > > > > > On 2020/03/13 13:35:52, Julian Feinauer > > <[email protected]> wrote: > > > > Hi Etienne, > > > > > > > > first, Congratulations on your Progress! > > > > To you question: > > > > This is a common issue with many protocols. > > > > We solve that in the protocol layer by keeping the request > > until we get the response (see for Example how we did it for S7). > > > > So you cannot solve that in mpsec at compile time but have to > > decode the byte[] you get with the Information from the Request. > > > > > > > > Hope that helps! > > > > Julian > > > > > > > > Am 13.03.20, 14:02 schrieb "Etienne Robinet" > > <[email protected]>: > > > > > > > > Hi Chris, > > > > I have yet another question. When sending a request for > > multiple elements (let's say 10 elements of an array), you get a response > > from the PLC with all the data at the end of the packet. The problem is, in > > the request there is the number of elements we want, but not in the > > response. So so the protocol knows how many elements there are in the > > response packet, but not the packet itself. This is quite a problem as for > > the parse, we need to know the length of the 'data' array containing the > > response data (for now it was only depending on the type, but I would need > > type*elements). > > > > > > > > I tested a bit by modifying the generated IO, and 1 way to > > do it is to get the remaining bytes and divide this by the datatype size. I > > just wanted to ask if someone would know how to declare this in the mspec, > > as I don't want to touch at the template. I also thought about having an > > attribute on the response, but I don't know how to put an attribute that > > won't get parsed/serialized. Hope I am clear enough, but this is a code > > sample that worked (modifying generated sources): > > > > https://i.imgur.com/Xm6DxEZ.png (notice the error but this > > code works if I write it myself) > > > > And here is how I put it in the mspec: > > https://i.imgur.com/Ye1Kl9q.png (you can see the number of element is not > > present) > > > > > > > > Any help is welcome of course! :) > > > > Etienne > > > > > > > > > > > > > > > > > > > > > > > > On 2020/03/12 21:34:12, Christofer Dutz > > <[email protected]> wrote: > > > > > Hi Etienne, > > > > > > > > > > that's amazing news :-) ... congratulations :-) > > > > > > > > > > Also had a look at your mspec and I think you have done a > > great job with that. I wasn't quite sure about the relation between > > CipRRData and CipExchange, but your solution looks rock-solid. > > > > > > > > > > And now reading that you even managed to get something > > working, that makes me very happy. > > > > > > > > > > Tomorrow I'll be a little consumed with signing the > > contract with the KNX foundation but I'll try to have a look at your fork. > > > > > > > > > > Thanks for your great work :) > > > > > > > > > > Chris > > > > > > > > > > > > > > > > > > > > Am 12.03.20, 17:50 schrieb "Etienne Robinet" > > <[email protected]>: > > > > > > > > > > Hi all, > > > > > again a productive day, I pushed to my branch and the > > driver support reading, multiple reading and the camel component works (in > > karaf) and takes a List of strings. Tested it on a different PLC and it > > also worked! Next I'm going to implement the array readings and then maybe > > writing (tests needs to be done too). > > > > > > > > > > Etienne > > > > > > > > > > On 2020/03/12 07:18:32, Etienne Robinet > > <[email protected]> wrote: > > > > > > Hi Chris, > > > > > > yes that's what I thought so I managed to work > > around like this: > > > > > > > > > > > > > > https://github.com/etiennerobinet/plc4x/blob/eip/protocols/eip/src/main/resources/protocols/eip/eip.mspec > > > > > > > > > > > > And this works for reading! I managed to make a > > quick test and read via the tag name. Now my question is: can I create > > sub-types that are also discriminated with sub-subtypes? That would allow > > to create the ReadRequest only, as the fields before are mainly > > constant/implicit. > > > > > > Etienne > > > > > > > > > > > > On 2020/03/11 20:24:14, Christofer Dutz > > <[email protected]> wrote: > > > > > > > Hi Etienne, > > > > > > > > > > > > > > you are defining the type CipRRData twice ... > > once as one of the sub-types of EipPacket and once as a dedicated > > discriminated type. > > > > > > > > > > > > > > Chris > > > > > > > > > > > > > > > > > > > > > PS: I have no idea why I didn't finish writing > > this email and I just noticed when closing everything down ... sorry for > > the late reply. > > > > > > > > > > > > > > > > > > > > > > > > > > > > Am 11.03.20, 09:30 schrieb "Etienne Robinet" > > <[email protected]>: > > > > > > > > > > > > > > Hi all, > > > > > > > > > > > > > > I have a quick question. I've been working on > > the CIP encapsulation but hitting a problem with the mspec design. Here is > > the error I am facing: https://i.imgur.com/iCfh59n.png > > > > > > > > > > > > > > The problem here is that this CipRRData > > should also be of type EipPacket. When the command of an EipPacket is > > '0x66' this means SendRRData (for Read/Write and Request/Response so I have > > to discriminate it on the sub level). The problem is that after generation, > > CipRRData implements Message but does not extend EipPacket. How should I do > > it? Here is the mspec: > > > > > > > > > > > > > > [discriminatedType 'EipPacket' > > > > > > > [discriminator uint 16 'command'] > > > > > > > [implicit uint 16 'len' > > 'lengthInBytes - 24'] > > > > > > > [simple uint 32 'sessionHandle'] > > > > > > > [simple uint 32 'status'] > > > > > > > [array uint 8 'senderContext' > > count '8'] > > > > > > > [simple uint 32 'options'] > > > > > > > [typeSwitch 'command' > > > > > > > ['0x0065' EipConnectionRequest > > > > > > > [const uint 16 > > 'protocolVersion' '0x01'] > > > > > > > [const uint 16 'flags' > > '0x00'] > > > > > > > ] > > > > > > > ['0x0066' EipDisconnectRequest > > > > > > > ] > > > > > > > ['0x006F' CipRRData > > > > > > > [const uint 32 > > 'EipInterface' '0x00000000'] > > > > > > > [const uint 8 'timeout' > > '0x0000'] > > > > > > > ] > > > > > > > ] > > > > > > > ] > > > > > > > [discriminatedType 'CipRRData' > > > > > > > [simple uint 8 > > 'itemNb'] > > > > > > > [array CipItem 'items' > > length 'itemNb'] > > > > > > > [discriminator uint 8 > > 'CipService'] > > > > > > > [typeSwitch 'CipService' > > > > > > > ['0x52' CipUnconnectedRequest > > > > > > > [simple CommandData 'data'] > > > > > > > ] > > > > > > > ['0xCC' CipReadResponse > > > > > > > [reserved uint 8 > > '0x0000'] > > > > > > > [simple uint 8 > > 'status'] > > > > > > > [simple uint 8 > > 'extStatus'] > > > > > > > [simple uint 8 > > 'dataType'] > > > > > > > [array uint 8 > > 'data' length 'dataType'] > > > > > > > > > > > > > > ] > > > > > > > ] > > > > > > > ] > > > > > > > > > > > > > > [type 'CipItem' > > > > > > > [simple uint 16 'typeID'] > > > > > > > [simple uint 16 'size'] > > > > > > > ] > > > > > > > > > > > > > > [discriminatedType 'CommandData' > > > > > > > [reserved uint 8 '0x02'] > > > > > > > [reserved uint 8 '0x20'] > > // setRequestPathLogicalClassSegment > > > > > > > [reserved uint 8 '0x06'] > > // set request class path > > > > > > > [reserved uint 8 '0x24'] > > // setRequestPathLogicalInstanceSegment > > > > > > > [reserved uint 8 '0x01'] > > // setRequestPathInstance > > > > > > > [discriminator uint 8 'CipService'] > > > > > > > [typeSwitch 'CipService' > > > > > > > ['0x4C' CipReadRequest > > > > > > > [simple uint 8 > > 'RequestPathSize'] > > > > > > > [reserved uint 8 '0x91'] > > > > > > > [array uint 8 'tag' > > length '(RequestPathSize*2) - 1'] > > > > > > > [reserved uint 16 '0x0001'] > > //itemCount set to 1 for now > > > > > > > ] > > > > > > > ] > > > > > > > > > > > > > > ] > > > > > > > > > > > > > > > > > > > > > [enum int 16 'CIPDataTypeCode' [uint 8 > > 'size'] > > > > > > > ['0X00C1' BOOL ['1']] > > > > > > > ['0X00CA' REAL ['4']] > > > > > > > ['0X00C4' DINT ['4']] > > > > > > > ['0X00C3' INT ['2']] > > > > > > > ['0X00C2' SINT ['1']] > > > > > > > ['0X02A0' STRUCTURED ['88']] > > > > > > > ['0X02A0' STRING ['88']] > > > > > > > ['0X02A0' STRING36 ['40']] > > > > > > > ['-1' UNKNOWN ['-1']] > > > > > > > ] > > > > > > > > > > > > > > > > > > > > > Thanks for any help! > > > > > > > > > > > > > > Etienne > > > > > > > On 2020/03/10 15:18:47, Etienne Robinet > > <[email protected]> wrote: > > > > > > > > Hi all, > > > > > > > > I've managed to implement the EIP Session > > Register/Unregister (used for connected message which is best for high > > frequency fetching). I will push it to my branch and create a document > > explaining my steps. > > > > > > > > Next I want to do was to to the CIP > > encapsulation part for accessing tag by their name. The thing is, CIP is > > (from what I heard) proper to Allen Bradley. Should I leave it in the EIP > > driver or modify and adapt the current AB driver later on? For now I will > > continue on eip. > > > > > > > > > > > > > > > > Etienne > > > > > > > > > > > > > > > > On 2020/03/10 14:41:42, Cesar Garcia > > <[email protected]> wrote: > > > > > > > > > Hi, > > > > > > > > > > > > > > > > > > You can document the process step by > > step, you will surely find observation > > > > > > > > > points. > > > > > > > > > > > > > > > > > > I am working with the handwritten S7 > > driver, but in the future I would > > > > > > > > > support the team in migrate to mspec, so > > the experience you will gain with > > > > > > > > > the Etheret / IP is very important. > > > > > > > > > > > > > > > > > > Best regards, > > > > > > > > > > > > > > > > > > El mar., 10 mar. 2020 a las 9:17, > > Christofer Dutz (< > > > > > > > > > [email protected]>) escribió: > > > > > > > > > > > > > > > > > > > Hi Etienne, > > > > > > > > > > > > > > > > > > > > I would strongly suggest you install > > the Antlr plugn for the IDE you use. > > > > > > > > > > The problem is that ANTLR seems to > > gobble up a lot of errors and tries to > > > > > > > > > > be smart in a lot of cases. > > > > > > > > > > Whenever I have problems like this I > > use the ANTLR visual parser to parse > > > > > > > > > > a block of mspec (one type at a time) > > and try to narrow down the cause. > > > > > > > > > > > > > > > > > > > > Chris > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Am 10.03.20, 13:31 schrieb "Etienne > > Robinet" <[email protected]>: > > > > > > > > > > > > > > > > > > > > Hi all, > > > > > > > > > > I've been struggling with > > implementing the EIP driver. I started > > > > > > > > > > writing the mspec after creating both a > > module for the protocol and the one > > > > > > > > > > from the driver. I copied and adapted > > the pom(s) from the AB-ETH but only > > > > > > > > > > the enum is generated. > > > > > > > > > > Here is a link to the forked branch: > > > > > > > > > > > > https://github.com/etiennerobinet/plc4x/tree/eip > > > > > > > > > > > > > > > > > > > > I've been studying the EIP/CIP > > protocol and now I am confident what > > > > > > > > > > the packages should contain but I can > > not figure out how to generate this > > > > > > > > > > with the templates. Am I missing > > something? > > > > > > > > > > > > > > > > > > > > Etienne > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > *CEOS Automatización, C.A.* > > > > > > > > > *GALPON SERVICIO INDUSTRIALES Y NAVALES > > FA, C.A.,* > > > > > > > > > *PISO 1, OFICINA 2, AV. RAUL LEONI, > > SECTOR GUAMACHITO,* > > > > > > > > > > > > > > > > > > *FRENTE A LA ASOCIACION DE > > GANADEROS,BARCELONA,EDO. ANZOATEGUI* > > > > > > > > > *Ing. César García* > > > > > > > > > *Cel: 0416-681.03.99* > > > > > > > > > > > > > > > > > > *Cel: 0414-760.98.95* > > > > > > > > > > > > > > > > > > *Hotline Técnica SIEMENS: 0800 1005080* > > > > > > > > > > > > > > > > > > *Email: [email protected] > > > > > > > > > <[email protected]>* > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
