feat: support start/stop fragment marker

[wyfo]

See eclipse-zenoh/zenoh#1597

[github-actions]

PR missing one of the required labels: {'bug', 'new feature', 'dependencies', 'documentation', 'breaking-change', 'internal', 'enhancement'}

[github-actions]

PR missing one of the required labels: {'dependencies', 'internal', 'documentation', 'enhancement', 'breaking-change', 'new feature', 'bug'}

[jean-roland]

Just a nit and I see there are some that precedes that PR, but all conditions should be booleans so the == true / false would be redundant and can be removed.

[wyfo]

See #813 (comment)
I asked before writing the code, but didn't understand Sasha's answer ^_^' (I would never have the idea to write == true if I had not seen it everywhere in the code)

[Mallets]

@sashacmc @jean-roland @wyfo this is about MISRA-C rules: https://stackoverflow.com/questions/44131637/misra-c-2012-rule-10-1-boolean-operand-to-be-used-in-case-where-expression-is-of

[sashacmc]

AFAIK, MISRA-C rules about using non-boolean values as boolean expression:
e. g.

int i = 1;
if (i) {}

but this kind of expression is absolutely legal:

bool b = true;
if (b) {}

[Mallets]

Well, yes but I also believe it depends on the static analyser... I'm not saying we should use == true, I was just providing a bit more context :)

[Mallets]

All CI is green but CI status checks which is stuck. I'm going to manual merge the PR.

[sashacmc]

All CI is green but CI status checks which is stuck. I'm going to manual merge the PR.

This is a known issue, @diogomatsubara is working on.