- Changed long to time_t for event times.
- Used PRIu64, PRIdMAX, and %lld for formatting and parsing time_t
and event IDs.
- Updated sscanf and printf calls to handle time_t and uint64_t correctly.
- Updated channelEPGformated to use time_t instead of long for _stoptime.
- Corrected sscanf format specifier in EpgCGI to handle time_t correctly.
- Changed handling of _stoptime in xmltvepgCGI to use time_t.
- Ensured gmtime and localtime functions receive time_t arguments.
These changes ensure compatibility with _TIME_BITS=64 and maintain
proper type usage for time functions, addressing different
definitions and maintaining robustness across kernel and library versions.
Ensure backward compatibility for rc_sync and channelEPGformated
- Added conditional compilation in rc_sync to check for input_event_sec macro.
- Used input_event_sec and input_event_usec macros if defined.
- Provided fallback to direct timeval members for older versions.
- Minor formatting adjustment in channelEPGformated for consistency.
This should ensure compatibility with both newer and older
kernel versions, different C library implementations,
and possible standard settings of build systems.
Addressing different definitions of the input_event structure maintains
robustness across kernel, library, and build system configurations.
Line 1: servicename
Line 2: epg now
Line 3: epg now duration X/Y
Line 4: epg next
Origin commit data
------------------
Commit: 866fc19dd0
Author: BPanther <bpanther_ts@hotmail.com>
Date: 2024-04-30 (Tue, 30 Apr 2024)
------------------
This commit was generated by Migit
Line 1: servicename
Line 2: epg info
Line 3: duration X/Y
Origin commit data
------------------
Commit: 3498c8c8a6
Author: BPanther <bpanther_ts@hotmail.com>
Date: 2024-04-30 (Tue, 30 Apr 2024)
------------------
This commit was generated by Migit
Origin commit data
------------------
Commit: bee4cb4959
Author: vanhofen <vanhofen@gmx.de>
Date: 2024-03-25 (Mon, 25 Mar 2024)
Origin message was:
------------------
- turn some comparsions for a better readability
------------------
No further description and justification available within origin commit message!
------------------
This commit was generated by Migit
Origin commit data
------------------
Commit: 0c0a1a1d5f
Author: TangoCash <eric@loxat.de>
Date: 2016-03-25 (Fri, 25 Mar 2016)
Signed-off-by: Thilo Graf <dbt@novatux.de>
Where is this documented?
Making changes to the API without any documentation is quite unpleasant.
Unfortunately, this has happened several times before.
It would be nice if someone could address this issue.
see: ./neutrino/doc/nhttpd/nhttpd_controlapi.html
We have a problem since the introduction of moviebrowser_mode.
Lastmode is no longer reset correctly, because the new mode is overlaid by another mode
and thus the original mode cannot be restored. Neutrino becomes unusable and has to be
restarted. (Example below)
Generally the opening of a window, like the MB or Channellist is also no Neutrino_mode in
the sense as it was invented once. It is just a window and not a mode that Neutrino is in.
Example: Neutrino is in TV_MODUS and you open the MB. You can see, Neutrino is still in TV_MODUS.
I reverted the code and now an open window can say for example to the lcd4linux 'hello i'm open, here is mb'.
When running a window we send some kind of AktionKey. After the actions in the window, the AktionKey is cleared.
Now to the current bug:
1. boot on a TV channel.
2. switch to a WebTV channel. (in my case PlutoTV Star Trek)
3. play a TS over the Moviebrowser and stop it. (MODE_WEBTV is detected and the WebTV stream is restarted)
4. call the MB a second time and start and stop the TS.
5. the WebTV stream does not restart anymore.
6. exit the MB now.
7. neutrino becomes not operable.
This was observed without physical frontends and in certain
circumstances while using the web interface, e.g. if users are
using bouquet functionalities.
