[webif] Web Interface version 1.0.17 released

[DelftBlue]

.... presumably the total excludes the TSR buffer.

Yes, about 20 GB is reserved for the Time Shift Recording buffer. Maybe that could be included in the Disk Utilisation list and then the figures would add-up and match the banner numbers.

 

[prpr]

Minor fix to stop units list overrun (when I screwed up a calculation elsewhere and got no units!)

humax# diff -u /mod/webif/lib/pretty_size~ /mod/webif/lib/pretty_size
--- /mod/webif/lib/pretty_size~
+++ /mod/webif/lib//pretty_size
@@ -3,7 +3,7 @@
  proc pretty_size {size} {
  set units {bytes KiB MiB GiB TiB}
-  for {set i 0} {$size > 1023} {incr i} {
+  for {set i 0} {$size > 1023 && $i < [llength $units] - 1} {incr i} {
  set size $($size / 1024.0)
  }

This fix is still outstanding...

 

[Black Hole]

Why are the matching entries for Reggie Perrin not expanded in the schedule listing, when the matching entries for other recordings are?

 

[cdmackay]

In the last few days, I've started seeing problems with the single page per channel EPG view, i.e. something like:

https://humax/epg/xservice.jim?service=8448

but only for some channels.

The problem is that I get nothing other than the banner header with the pie chart. No tables, no nothing.

The few channels include Ch4 & E4. Most channels are fine, however.

The EPG itself has the data: I can see it in grid view, and via the on-telly UI. It's solely the single page per channel view in WebIF that is broken, for these few channels.

This started a day or so before the recent retune (which I did yesterday, and which made no difference).

The webif-error.log has this to say when it happens:

Runtime Error: /mod/webif/lib/epg.class:363: child killed by signal SIGSEGV
in procedure 'epg' called at file "/mod/webif/html/epg/xservice.jim", line 20
in procedure 'epg' called at file "/mod/webif/lib/epg.class", line 463
at file "/mod/webif/lib/epg.class", line 363

which would appear to indicate that it is /mod/bin/epg that is crashing?

Anyone have any ideas, please?


edit: apols, my details:

Web interface version: 1.0.17-8
Custom firmware version: 2.23 (build 2035)
Humax Version: 1.03.12 (kernel HDR_1.02.32)

 

[4ndy]

There seems to be a more general issue showing EPG in both webif and RS. I can't get either to display anything after Friday.


On screen the EPG is populated for most channels, though Film4+1, The Chart Show TV and True Entertainment show nothing beyond Friday.

Edit: Forgot to mention that it is the same on all three HDRs

Sent from my iPad using Tapatalk

 

[4ndy]

I have cleared the epg cache data in maintenance mode on all three machines and all now show a full week epg for most channel. Can't beleive that it was coincedence that all three HDRs had the same problem

Individual channel view is not working though. Get message about no epg data cached. Get similar message in status banner too.

 

[prpr]

I'm affected too, so it's obviously a general problem. A small amount of investigation reveals the "epg" process which creates the sqlite database used by the WebIf isn't running. Running it manually reveals this:

humax# epg -f /mnt/hd1/dvbepg/epg.dat sqlitedumpd /mnt/hd1/epg.db
epgd: DB time 1416769194, last 0
epgd: Regenerating.
Segmentation fault

I have taken a copy of epg.dat in case it is useful to af123, but the thing shouldn't be crashing in the first place, so certainly needs a fix or making more robust against whatever bad data it's chewing on.

 

[prpr]

Another fix for a table heading bug:

humax /mnt/hd2/mod/webif/html/epg # diff -u xservice.jim~ xservice.jim
--- xservice.jim~
+++ xservice.jim
@@ -55,10 +55,10 @@
  <th>Time</th>
  <th>Programme</th>
  <th>Synopsis</th>
+  <th></th>
  </tr>
 }

-set i 0
 foreach record $records {
  altrow
  $record get_channel_info

Maybe the header should say "Guidance" or "Warning", or maybe it's better left blank...

 

[prpr]

I have cleared the epg cache data in maintenance mode on all three machines and all now show a full week epg for most channel. Can't beleive that it was coincedence that all three HDRs had the same problem

Individual channel view is not working though. Get message about no epg data cached. Get similar message in status banner too.

Clearing the epg.dat file and rebooting makes it work for a few channels for a few seconds here, then the "epg" utility crashes. There is definitely a problem with the data upsetting this, not just a corrupt disk file.

 

[prpr]

Fix for "nown" on the Channel Information page:

humax /mnt/hd2/mod/webif/cgi-bin # diff -u channel.jim~ channel.jim
--- channel.jim~
+++ channel.jim
@@ -64,7 +64,7 @@
  set ehs 0
  foreach chan $channels {
  lassign $chan x lcn x name
-  set name [string range $name 1 end]
+  set name [system strip $name]
  if {$lcn >= 800} { incr ehs }
  switch $name {
  "BBC THREE" { set mux "PSB1/BBC A" }
@@ -104,8 +104,8 @@
  altrow reset
  foreach chan $channels {
  lassign $chan x lcn x name x prv x auth
-  set name [string range $name 1 end]
-  set prv [string range $prv 3 end]
+  set name [system strip $name]
+  set prv [system strip $prv]
  altrow
  puts "
  <td class=va>[epg channelicon $name 50]</td>

 

[Black Hole]

I realise this is something different, but I have never been able to get a WebIF EPG on my HD-FOX. There used to be an error message something like "no EPG data available", but now nothing.

 

[af123]

I'm affected too, so it's obviously a general problem. A small amount of investigation reveals the "epg" process which creates the sqlite database used by the WebIf isn't running. Running it manually reveals this:

humax# epg -f /mnt/hd1/dvbepg/epg.dat sqlitedumpd /mnt/hd1/epg.db
epgd: DB time 1416769194, last 0
epgd: Regenerating.
Segmentation fault

I have taken a copy of epg.dat in case it is useful to af123, but the thing shouldn't be crashing in the first place, so certainly needs a fix or making more robust against whatever bad data it's chewing on.

Same here, I'll see what's going on.

 

[af123]

New EPG package uploaded.. the culprit is the data for the Hollyoaks Omnibus which has something wrong with its guidance text descriptor. Shouldn't have made the EPG utility crash though so I've fixed that.

 

[af123]

For those who are interested, EPG entries are made up of a set of descriptors defined in ETSI EN 300 468. In Freeview, descriptor 137 is used to carry the guidance text.

That Hollyoaks EPG entry has a guidance descriptor which just specifies the language - no actual content. Must be an error on the broadcaster's side. Usually a programme without guidance just doesn't contain the descriptor.

Checking descriptor: 137/0x89 (4)
00000000: 89 04 00 65 6e 67  ...eng

Versus

Checking descriptor: 137/0x89 (51)
00000000: 89 33 00 65 6e 67 4d 61 79 20 63 6f 6e 74 61 69  .3.engMay contai
00000010: 6e 20 63 6f 6e 74 65 6e 74 20 75 6e 73 75 69 74  n content unsuit
00000020: 61 62 6c 65 20 66 6f 72 20 73 6f 6d 65 20 76 69  able for some vi
00000030: 65 77 65 72 73  ewers

 

[prpr]

It's probably worth saying that after this package upgrade, you either need to reboot your box or telnet in and run this:

humax# /mod/etc/init.d/S60parseepg start

I'm a bit surprised there isn't a script in the package to start this...

 

[prpr]

I realise this is something different, but I have never been able to get a WebIF EPG on my HD-FOX. There used to be an error message something like "no EPG data available", but now nothing.

I get this and it was no different today. I guess the Humax software doesn't save the EPG data to disk unless it sees a certain size of disk. 8G is not enough, which is quite ridiculous, but I guess there is no point (for its own operation) unless there is space for the Tsr buffer and to record.
I have just 'solved' it on mine by using rsync to copy the relevant file from an HDR. Works nicely now.

humax# rsync -av -e='ssh -i /mod/.ssh/db_rsa' 192.168.xxx.yyy:/mnt/hd1/dvbepg/epg.dat /media/drive1/epgsavedata

Of course you need dropbear-ssh and rsync installed and setup etc.
Then you can put the above into a Cron job entry and forget about it.

 

[Black Hole]

I don't find I need to look at the WebIF EPG for my HDs - but thanks for explaining the problem.

 

[raydon]

For those who are interested, EPG entries are made up of a set of descriptors defined in ETSI EN 300 468. In Freeview, descriptor 137 is used to carry the guidance text.]

Yes, unfortunately that definition of descriptor tag 137 (0x89) merely states "user defined" with no hint of exactly what content Freeview assign to it. There is the three character ISO 639 language descriptor immediately followed by the guidance text, but the language code is also preceded by another field. No idea what this is. If the byte is zero the next field byte is not present. If it's a one it's always followed by another byte containing either one or zero. Any ideas ? Some sort of PG code perhaps ? BTW This is as it appears in the EIT stream in the .ts, not sure about EPG.dat

 

[af123]

No idea. It's <descriptor tag> <descriptor length> <unknown length> [<unknown>] <language> <warning - to end>

Could it indicate whether the warning string is compressed?

Edit: nope...

 

[af123]

It's probably worth saying that after this package upgrade, you either need to reboot your box or telnet in and run this:

humax# /mod/etc/init.d/S60parseepg start

I'm a bit surprised there isn't a script in the package to start this...

The epg package just provides the utility. The startup script is part of the webif package as that needs the generated database.