Is there any example for how to use thIs?

[zeerd]

Means , need I modify my codes for the outputting logs with some special formale?

I did not found any documents to discuss this.

[svlad-90]

Hi @zeerd, let me provide you with some information regarding your question.

The DLT-Message-Analyzer is not a stand-alone application. It is only a plugin for the GENIVI ( COVESA ) dlt-viewer: https://github.com/COVESA/dlt-viewer

The installation guide of this repo explains how to install the DLT-Message-Analyzer plugin on top of the installed dlt-viewer: https://github.com/svlad-90/DLT-Message-Analyzer/blob/master/md/installation_guide/installation_guide.md

dlt-viewer's repo has INSTALL.md, which explains how to build and install the dlt-viewer application: https://github.com/COVESA/dlt-viewer/blob/master/INSTALL.md

---

But all above is only the "viewer" part. It will allow you to analyze *.dlt files.

In order to SEND the messages:

• you need dlt-daemon being installed on your target device: https://github.com/COVESA/dlt-daemon. It can be installed at least on Linux and Mac OS X operating systems. I haven't tested it on the other OS.
• you need to modify your application to use the dlt logging library: https://github.com/COVESA/dlt-daemon/blob/master/include/dlt/dlt.h. Analyze this header and headers, which it includes. Also, you can try to traverse the following document, to have a more extensive guide: https://github.com/COVESA/dlt-daemon/blob/master/doc/dlt_demo_setup.md.

In other words, you need to deploy the following environment on your device:

Hopefully, the above provides enough overview so that you could move on. Please, try it out.

If you'll have further questions - let me know. ))

[zeerd]

Seems, my question is not clear enough.

I means, is there a formale for the output strings?

For example:

In a.c , I could use DLT_LOG(C, DLT_LOG_INFO, DLT_STRING("Hello, B!"))

in b.c I could use DLT_LOG(C, DLT_LOG_INFO, DLT_STRING("Hello, A!"))

And then, how could zhe DMA know it is :

A -> B : Hello, B!
B -> A : Hello, A!

Need I do something like DLT_STRING("from A to B : Hello, B!") ? Is there a rule for this? I did not find a link to some doc of those things from the README.md in the root dir.

[svlad-90]

Hi @zeerd, Ok. Now your question is clear enough.

You are interested in the plantuml features. Particularly, in how to let the plugin know which part of the message is which token of the sequence diagram. Which part is the client? Which part is the service? Which part is the method name? Etc.

The documentation covers this. The following article explains it extensively: https://github.com/svlad-90/DLT-Message-Analyzer/blob/master/md/plant_uml/plant_uml.md. Read "General idea", "Simple example" and "Advanced example" sections carefully.

---

In short - the plugin does not have any hard-coded formatting rules for the strings. You are building those rules using the regex group names:

Group name	Meaning	Type
USID	Sequence id of the communication	optional
UCL	Name of the client	mandatory
URT	Request identifier	request_type - at least one of the request types should be filled in
URS	Response identifier	request_type - at least one of the request types should be filled in
UEV	Event identifier	request_type - at least one of the request types should be filled in
US	Service name	mandatory
UM	Method name	mandatory
UA	Call arguments	optional

In order to build a sequence diagram:

• you need to build a regex, which will contain all the mandatory group names
• have the strings being found, which will contain at least all the mandatory group names
• press the UML check-boxes on all rows, which you want to use inside the diagram
• switch to the "UML View" tab and press the "Create sequence diagram" button

Be aware, that if the output of your application is not unified, you can still work with this, creating UML-capable regex expressions for different types of output. Store them, as separate regex patterns and then use a combination of those to build a sequence diagram, which will contain rows with different formatting.

In other words, the less unified output you have, the more regex expressions you'll need to build. Still, it is possible to work even in that way. At least, I had such experience when was trying to build sequences of different domains, which had totally different formatting of the output messages.

If there will be any further questions - do not hesitate to ask me ))

[svlad-90]

Hi @zeerd, if your question is answered - please, let me know. In such a case I'll close it ))

[svlad-90]

As there is no further feedback, the issue is marked as closed.

@zeerd, please, reopen it in case if further help is needed regarding this topic. ))