[webif] Version 1.0.0 Released

af123 said:
You have a pending uncorrectable sector. What happens if you try a hard disk check using the on-screen TV menus? I suspect it will fail and suggest that you reformat.

You need to run a full disk check from maintenance mode to repair this before it causes problems with recordings.


Posted on the move; please excuse any brevity.
Click to expand...
Sorry for the delay in getting back to you. I attempted the on-screen Hard Disk Check but no problems were reported. I have upgraded to the latest version of the custom firmware (2.16) but the error message persists on the Web Interface. Is there an option to disable this screen? I have seen no problem with recordings for the past few weeks (since this message appeared).

I will probably need to switch to Maintenance Mode and run a disk check and report back my findings. Thanks as usual for your great help!
 
af123 said:
;
  • Fix issue with decrypting single recordings .
Click to expand...

I have two files recorded which were named 691.TS (7.43 GiB)and 822.TS (3.59GiB). There is no decrypt option on webif. There are also a few other unusual features.

The first file is 2 hours 36 minutes and I presume it is Hi Def but when I press the ‘i’ button twice it doesn’t give details. (Interestingly the duration is 01:00 to 03:36). Playback features are limited.

I have an identical file that was scheduled and recorded (HiDEf) 7.68 GiB which webif says is 2 33 minutes duration. With full playback functionality.

The second file 2 hours 02 minutes and I presume it is Standard Def but when I pres the ‘i’ button twice it doesn’t give details. (Interestingly the duration is 01:00 to 03:02)

Strange again I have a recording identical (time 2.02 duration) as a normal stored file. With full play back functionality.


It seems that the box recorded both films twice.

On the 691.TS and 882.TS files on web if when I click on the name I get “ Media Details” then File :…. Size :….. and Info. (no synopsis etc).

The info contains

Code: 
Size    7.43 GiB
Info    libpostproc 52. 0.100 / 52. 0.100
[aac_latm @ 0x441300] audio config changed
[h264 @ 0x4341e0] non-existing SPS 0 referenced in buffering period
[h264 @ 0x4341e0] non-existing PPS referenced
[h264 @ 0x4341e0] non-existing SPS 0 referenced in buffering period
[h264 @ 0x4341e0] non-existing PPS 0 referenced
[h264 @ 0x4341e0] decode_slice_header error
[h264 @ 0x4341e0] mmco: unref short failure
Last message repeated 1 times
[h264 @ 0x4341e0] Increasing reorder buffer to 1
[h264 @ 0x4341e0] Increasing reorder buffer to 2
[h264 @ 0x4341e0] Increasing reorder buffer to 3
 
 
 
[h264 @ 0x4341e0] mmco: unref short failure
[mpegts @ 0x41bb30] max_analyze_duration 5000000 reached at 5013333 [mpegts @ 0x41bb30] Could not find codec parameters (Unknown: none ([5][0][0][0] / 0x0005))
[mpegts @ 0x41bb30] Could not find codec parameters (Unknown: none ([11][0][0][0] / 0x000B))
Last message repeated 1 times
[NULL @ 0x453a50] start time is not set in estimate_timings_from_pts
[NULL @ 0x459cc0] start time is not set in estimate_timings_from_pts
[NULL @ 0x45ff30] start time is not set in estimate_timings_from_pts
[mpegts @ 0x41bb30] PES packet size mismatch Last message repeated 1 times
.........

Code: 
3.59 GiB
Info    libpostproc 52. 0.100 / 52. 0.100
[mpeg2video @ 0x4341e0] mpeg_decode_postinit() failure
Last message repeated 18 times
[mpegts @ 0x41bb30] max_analyze_duration 5000000 reached at 5016000
[mpegts @ 0x41bb30] Could not find codec parameters (Unknown: none ([5][0][0][0] / 0x0005))
[mpegts @ 0x41bb30] Could not find codec parameters (Unknown: none ([11][0][0][0] / 0x000B))
Last message repeated 6 times
[NULL @ 0x439d40] start time is not set in estimate_timings_from_pts
 
........

I had tried to manually record a program, whilst one was being scheduled then stopped that recording. The next morning the .ts files appeared.

This seems unusual that the box can record the program twice.
 
These files (691.TS and 822.TS) are the intermediate products of incomplete decrypt operations. Normally they would be cleaned up at the end, but if (for example) the Humax got turned off before the operation completed, the process would start again next time the box was turned on. If you have the correct files then you can delete the <number>.TS files. If you do not have the correct files, you might find the non-decrypted originals in the undelete recycle bin.

"auto.log" or "unencrypt.log" (accessible from the Diagnostics page on the WebIF) might contain a record of what happened.
 
Webif automatic decryption doesn't use files with names like that and it doesn't create any temporary files in the same directory as the recording. This must be something else (perhaps unencrypt ?)
 
Alright, unencrypt.log then (previous post updated).

For anyone still using unencrypt, note that the auto-decrypt options in the WebIF are much cleaner and unencrypt is effectively unsupported. It is however still useful if you have an HDR-FOX with no network access and rely on loading custom package bundles by USB.

PS: I'm still running unencrypt on HDR1!
 
AndrewB said:
Still a problem on the current Chrome, I'm afraid. Version 26.0.1410.64 m.

Some patterns in the behaviour that I've noticed:

Observed effect: The file info pop-ups lose their background and formatting; the EPG list has no grid formatting at all and is just a long list of programme names; buttons are un-styled. The data on the page is correct; just that the styling is missing or incomplete.

Affected pages: I've observed the problem on the EPG (/cgi-bin/xepg.jim) and the file list (/browse/index.jim), but no other pages so far.

Patterns in repeatability: The problem never occurs when a page has been refreshed (Ctrl-F5). It only ever happens when navigating from one page to another. A full page refresh always fixes the problem for me, but then it re-occurs when navigating back to the pages.

Patterns in HTTP headers: There is a difference between HTTP requests which result in a broken page versus a correctly rendered page. The difference is that the correctly rendered page always has the following HTTP parameters in the request header:

Cache-Control:no-cache
Pragma:no-cache

Conversely, these HTTP parameters are always missing in the instances where the page ends up broken. Of course, this isn't necessarily causal, and it means little to me... But it might help to point at where the problem lies? It does feel like part of the page isn't being loaded into the browser correctly - like some of the CSS definitions or something?

I've made some screenshots of the captured HTTP responses. The first shows when the page worked, and the second when it failed:

View attachment 539 View attachment 540

HTTP violations in general: I don't think this is necessarily related to the problem; but my Fiddler proxy reports an HTTP protocol violation on every page that the Hummy serves. This violation is that the response has no Content-Length or "connection close" - but, that affects every page regardless of success.

My web-if: This is important..! I'm still on 0.13.3-4. So even though this is being discussed in the 1.0.0 thread, it's important to note that this Chrome problem affects the previous version of web-if.

Hope that helps.

Cheers
Click to expand...

I am still getting that behaviour, too. Using the latest version of Chrome on Win 7 and Win 8 and everything up to date on the T2.

I can't remember this always being the case, just recently.
 
My spare browser (Firefox) now has the Hummy set as its home page. I tried a number of so-called 'lightweight' and 'portable' browsers, but they all glitched on something in the Hummy displays - not the same somethings, unfortunately. Firefox has no problems with anything on the pages in question. It's not particularly unusual to find that a particular browser doesn't like a particular site, but to have a variety of issues with a selection of browsers suggests that all is not quite as it should be in the page code.
 
fenlander said:
My spare browser (Firefox) now has the Hummy set as its home page. I tried a number of so-called 'lightweight' and 'portable' browsers, but they all glitched on something in the Hummy displays - not the same somethings, unfortunately. Firefox has no problems with anything on the pages in question. It's not particularly unusual to find that a particular browser doesn't like a particular site, but to have a variety of issues with a selection of browsers suggests that all is not quite as it should be in the page code.
Click to expand...


My feeling is that one of those <link/>ed files is not being loaded correctly.
 
Black Hole said:
No problems with Safari on the iPad :)
Click to expand...

I haven't tried Chrome on a tablet. It might be a PC thing.

You used to be able to tweak settings in browsers so that pages could be forced to reload, never using a cached copy. These cut-down browsers seem to have hidden those settings away. Internet Explorer still has it, though.

I would use Firefox, but Chrome has this nice feature of syncing settings (bookmarks, etc) between all my devices, and it doesn't work correctly in Firefox, unfortunately.
 
Use Portable FireFox and have the WebIf as my home page. As fenlander states, it works just fine. Actually it is proving quite convenient.
 
fenlander said:
My spare browser (Firefox) now has the Hummy set as its home page. I tried a number of so-called 'lightweight' and 'portable' browsers, but they all glitched on something in the Hummy displays - not the same somethings, unfortunately. Firefox has no problems with anything on the pages in question. It's not particularly unusual to find that a particular browser doesn't like a particular site, but to have a variety of issues with a selection of browsers suggests that all is not quite as it should be in the page code.
Click to expand...
If you use Firefox, get the Html Validator add-on. You will be amazed at how invalid/crap the markup is on almost every site out there (and this forum is no exception).
When I was having problems with my own pages, all the funny behaviour between browsers went away when I cleaned up the HTML.
I have tried persuading other people of the virtues of doing this over the years... with varying levels of success.

Sadly, the WebIf is no different - there is loads of deprecated stuff, technically invalid stuff, and occasionally other more serious errors.
Whether this is the cause of the above problems I do not know, but it is not a good point to start from.
 
I'm more than happy to do some cleaning up of the back end web interface code where it makes sense, but not to the nth degree. For example, unquoted plain attributes don't pass strict validation but there is nothing fundamentally wrong with them.
I think the Chrome problem is somehow to do with caching of the dynamic HTML pages - nobody seems to be having problems with dynamically generated CGI ones which send cache control headers. Over on the Foxsat they have the same problem with recent Chrome builds and their web interface pages do pass W3C validation but they have modified the web server to send cache control headers on more content; maybe I should do the same.
 
af123 said:
I'm more than happy to do some cleaning up of the back end web interface code where it makes sense, but not to the nth degree. For example, unquoted plain attributes don't pass strict validation but there is nothing fundamentally wrong with them.
I think the Chrome problem is somehow to do with caching of the dynamic HTML pages - nobody seems to be having problems with dynamically generated CGI ones which send cache control headers. Over on the Foxsat they have the same problem with recent Chrome builds and their web interface pages do pass W3C validation but they have modified the web server to send cache control headers on more content; maybe I should do the same.
Click to expand...

I think you are right there. Dynamic stuff needs (meta) headers that tell a browser not to cache files.
 
I've been getting this on chrome on my PC, but I'd fine on chrome on my nexus. The ie tab in chrome fixes it. Most interesting is that I've also discovered my bt home hub admin page is also affected with images not downloading.
 
annonimouse said:
Most interesting is that I've also discovered my bt home hub admin page is also affected with images not downloading.
Click to expand...


Gosh, well done! I had forgotten about that! I don't use the BT hub any more, but there were also problems in Firefox, to do with background images not loading, eventually traced to antivirus and anti-adware software. Something to do with the string "adv" appearing in the file names as I recall.

So, could this be a clue to what is happening here?
 
Possible problem report:

This might (possibly) just be finger trouble, but I don't think so. I am raising the observation in case it is verified as a pattern of misbehaviour.

I decided to record Book at Bedtime this next fortnight, because it is a reading of Le Carre's new novel A Delicate Truth. Because I run flatten, I used the WebIF to modify the schedule entry to record to folder "A Delicate Truth", and confirmed the modification was in the pending list (but did not action it - there were plenty of reboots to come). I created a folder of that name, and used the WebIF OPT+ to mark it no-flatten and auto-audio (extract MP3). I thought everything was set up fine.

Late last night, after the recording and quite a while to run the auto processes, I was flicking through idly (HD-FOX by network share) and came across the A Delicate Truth folder with nothing in it. Not thinking too much I tried switching the media type to music, still nothing. Then I started getting cross thinking it had not recorded at all.

The recording was actually in My Video, unconverted, having been recorded to "Book at Bedtime" and flattened. When I checked the schedule the entry was there unmodified, so I changed folder again and committed it straight away to verify the update. When I moved the recording to the desired folder it converted, so the folder now contains the .ts and a .mp3.

I am wondering why the folder change did not commit the first time around. As far as I can see it should have done.
 
I've uploaded a new version of the mongoose package which sets no-cache headers on the plain HTML and server-side include files. With any luck, this should improve things for chrome users.
 
Sorry, but even with the updated mongoose package and a box reboot, the Webif is still broken when using Chrome. Well, it is for me. It's really weird. I can be clicking backwards and forwards in the WebIf and all is fine and dandy, then randomly, a page is broken.

FYI: Chrome v26.0.1410.64 m
 
I was hopeful for a while there, but then the EPG (grid view) broke. It worked correctly several times before it finally failed. So far I haven't been able to break any of the other pages.
 
You must log in or register to reply here.
Back
Top