Closed dornhege closed 10 years ago
Thanks, looks good to me! Should be rebased onto fuerte/groovy/hydro.
@jspricke : Do you have time to test & merge this this week? Otherwise I'll do it next week, when I'm back in Osnabrück.
I have only tested on groovy and LAN so far. I'll also be gone this week, but the 1 packet/scan assumption might fail on unreliable connections.
Sure, but that's still better than no TCP support at all, so I suggest to merge it first and fix it later if there turn out to be problems. The USB driver also makes the 1 packet/scan assumption, which always works if polling at a reasonable rate. Also, as long as we can't provoke the scanner into deviating from the 1 packet/scan assumption, we can't test if any code that deals with that actually works.
@jspricke already agreed on having a look at this. I'm pretty confident that if this works for groovy, it'll also work for fuerte and hydro without further testing (as long as it compiles).
The problem with moving the stop_scanner()
logic to the destructor is destruction order. The specialized classes would close the device before stop_scanner()
is called.
Ah, you are perfectly right with stop_scanner()
. Thanks for the update, will merge it after lunch.
Because of the merge in here, I've rebased the commits on top of all branches, thanks for the patches!
Added basic TCP support in branch tcp_support.