On 14.12.2012 16:27, Malte Forkel wrote:
Am 14.12.2012 10:23, schrieb Klaus Schmidinger:
On 13.12.2012 23:19, Malte Forkel wrote:
Hi,
I noticed the following when using SVDRP: With the command NEWT, you can't create a timer if there already exists a timer for the same channel, day, start time and stop time. But you can create some dummy timer and then use the MODT command to make it identical to the existing timer. Here is a sample session:
...
Looking at svdrp.c and timers.c, cSVDRP::CmdNEWT will use cTimers::GetTimer to look for an existing timer, cSVDRP::CmdMODT does not seem to perform a check.
I'm not sure about the rationale. What would I break if I dropped the test in cSVDRP::CmdNEWT? Are there better workarounds than creating a dummy timer?
NEWT performs a simple test to prevent the user from inadvertently creating duplicate timers. With MODT things are somewhat more complex, because the timer is already in the list, so it would have to check whether there is an *other* timer with the same settings. Sure, this could also be done, but then again just let's keep things simple...
Oh, I would like to argue for dropping the test in NEWT! I think that might make things even more simple (at least for me :-)) and more consistent. After all, there is nothing that keeps me from creating duplicate timers with any of the other interfaces I use. And if an application wanted to make sure it does not create duplicate timers, comparing the new definition to the output of LSTT would be rather easy to implement. Plus, VDR itself does not seem to have any problems with duplicate timers.
On the other hand, if there is a test in NEWT, I think it should check other attributes as well, e.g. active vs inactive and recording path.
My use case is this: I'm writing a script that changes the names ofsome timers created by EPGSearch. Because EPGSearch will revert any changes to these timers, the script disables the original timer and then creates a new one with a different name. The latter part requires some trickery at the moment, because NEWT rejects the new timer even though it is different from the original onein the two attributes mentioned.
Well, if nobody objects I have no problem with removing that check from NEWT.
Klaus