I suppose I was hoping that Humax would fix the EPG speed issue, widely reported on here and hopefully spotted by their beta testers with the unofficial 1.03.06 & 11, with the final OTA release. Unfortunately not which got me thinking about the problem as I tried to get to sleep last night.
What we know is:
The problem occurs both with and without custom firmware.
Good : ) I'm always happy to eliminate CFW early on.
The performance appears to be related to the number of entries in the schedule (I'm as yet unconvinced that expired or non-matching entries contribute to the problem more than any other type of entry in the schedule);
So, this could mean that every time the EPG is scrolled and the screen refreshes, the software is doing a scan of the schedule database to work out where to put the scheduled icons... I can't see why this would be worse than previous versions though.
A factory reset followed by schedule restore or re-input helps some people. It didn't help me. Some people find the problem returns;
As prpr has pointed out in another thread, a factory reset wipes the schedule database. If restoring the same schedule following a factory reset results in better performance then perhaps automatically cleaning up the database on each boot would be beneficial? (SQLite3 Vacuum). Last time I looked at it the Humax software copied the database into memory at startup (I suspect using the SQLite Online Backup API) and only periodically flushed it back to disk. I haven't looked at 1.03.xx in detail.
The EPG is now updated in the software update window
I can't see how this could have an effect but if they have broken something in the EPG update code then the EPG database could be getting larger each day. 'diag epgrange' will show you. Mine's actually going back to Jan 31st at present so I will do an EPG reset from maintenance mode later and see if it helps.
Humax have begun to implement remote scheduling and EPG over IP in 1.03.xx
This could be one of the new portal apps that they're crowing about. This could be a factor - more investigation required.. a network packet trace while scrolling the EPG could be interesting.
So, let's see if we can bottom this out and come up with a fix ourselves or at least a set of things for people to try over and above the existing good advice to remove expired entries and factory reset.
Can anyone think of a way of quantifying the lag? How about bringing up the EPG on channel 72 and holding down the right button until the end of data is reached and timing that? That would let us broadly compare boxes and see if specific things fix the issue.
What we know is:
The problem occurs both with and without custom firmware.
Good : ) I'm always happy to eliminate CFW early on.
The performance appears to be related to the number of entries in the schedule (I'm as yet unconvinced that expired or non-matching entries contribute to the problem more than any other type of entry in the schedule);
So, this could mean that every time the EPG is scrolled and the screen refreshes, the software is doing a scan of the schedule database to work out where to put the scheduled icons... I can't see why this would be worse than previous versions though.
A factory reset followed by schedule restore or re-input helps some people. It didn't help me. Some people find the problem returns;
As prpr has pointed out in another thread, a factory reset wipes the schedule database. If restoring the same schedule following a factory reset results in better performance then perhaps automatically cleaning up the database on each boot would be beneficial? (SQLite3 Vacuum). Last time I looked at it the Humax software copied the database into memory at startup (I suspect using the SQLite Online Backup API) and only periodically flushed it back to disk. I haven't looked at 1.03.xx in detail.
The EPG is now updated in the software update window
I can't see how this could have an effect but if they have broken something in the EPG update code then the EPG database could be getting larger each day. 'diag epgrange' will show you. Mine's actually going back to Jan 31st at present so I will do an EPG reset from maintenance mode later and see if it helps.
Humax have begun to implement remote scheduling and EPG over IP in 1.03.xx
This could be one of the new portal apps that they're crowing about. This could be a factor - more investigation required.. a network packet trace while scrolling the EPG could be interesting.
So, let's see if we can bottom this out and come up with a fix ourselves or at least a set of things for people to try over and above the existing good advice to remove expired entries and factory reset.
Can anyone think of a way of quantifying the lag? How about bringing up the EPG on channel 72 and holding down the right button until the end of data is reached and timing that? That would let us broadly compare boxes and see if specific things fix the issue.