51 replies to this topic

[altobelli]

De blindscanoptie werkt niet meer na de laaste update.

De box opnieuw geflasht .d.d 03032015.

[Erik Slagter]

We hebben de blindscan plugin overgenomen van een aantal mensen die daar flink aan gewerkt hebben sinds de laatste versie die wij via de 3rd party plugins aanboden. Maar het lijkt er op dat hij niet helemaal geweldig werkt, ik zie meer klachten.

Edited by Erik Slagter, 4 March 2015 - 15:39.

[altobelli]

Ik hoop dat het snel wordt opgelost

[Huevos]

Erik, it has worked fine for several years in OE-A. Not sure why PLi are having so much trouble with this. Really is a shame as it is the most customizable blindscan plugin there is. Only limitation is the closed source binary.

[blzr]

actually, there's little openpli's fault in this (if any )

 

problem lays in xtrend bsp layer (which OE-A doesn't use),
because of 'strange' et-blindscan-dvbs-utils recipe, this package is not built as such, and dvbs blindscan binary is not provided, at the moment;
seems to be pretty easy to solve (IF one would have access to repo / contact with xtrend repo maintainer),

e.g. something like this:

diff --git a/recipes-bsp/drivers/et-blindscan-dvbs-utils.bb b/recipes-bsp/drivers/et-blindscan-dvbs-utils.bb
index 48b4cdf..1f63c2d 100644
--- a/recipes-bsp/drivers/et-blindscan-dvbs-utils.bb
+++ b/recipes-bsp/drivers/et-blindscan-dvbs-utils.bb
@@ -3,13 +3,15 @@ SECTION = "base"
 PRIORITY = "optional"
 LICENSE = "CLOSED"
 
-SRC_URI = "http://www.xtrendet.net/enigma2-plugin-systemplugins-blindscan_1.4_mipsel.zip"
+PACKAGE_ARCH = "${MACHINE_ARCH}"
+
+SRC_URI = "http://xtrendet.net/xtrend-dvbs-blindscan-${PV}.tar.gz"
 
 PROVIDES += "virtual/blindscan-dvbs"
 RPROVIDES_${PN} += "virtual/blindscan-dvbs"
 
 PV = "1.0"
-PR = "r7"
+PR = "r8"
 
 S = "${WORKDIR}"
 
@@ -17,11 +19,11 @@ do_compile() {
 }
 
 do_install() {
-    install -m 0755 -d ${DEPLOY_DIR_IPK}/${MACHINE}
-    install -m 0644 enigma2-plugin-systemplugins-blindscan_1.4_mipsel.ipk ${DEPLOY_DIR_IPK}/${MACHINE}
+       install -d "${D}/${bindir}"
+       install -m 0755 "${S}/avl_xtrend_blindscan" "${D}/${bindir}"
 }
 
 INHIBIT_PACKAGE_STRIP = "1"
 
-SRC_URI[md5sum] = "50e866cd0e33825260442ba39212f0ea"
-SRC_URI[sha256sum] = "f854a6924e932615dd76716b94fd80221ee35382cecc630a063d5f4f5db36a80"
+SRC_URI[md5sum] = "194e60464f9d65d68f625cda1bc19b4d"
+SRC_URI[sha256sum] = "f2d21945afc0a06a5ac1fc23617948380a4c117f940029251ae9373c69c5b730"

(just an example, of course in this case archive with avl_xtend_blindscan binary should be put on xtrend's server)
 

---
and one more related change, in meta-dream:

diff --git a/meta-dream/recipes-bsp/drivers/dreambox-blindscan-utils_1.7.bb b/meta-dream/recipes-bsp/drivers/dreambox-blindscan-utils_1.7.bb
index a42794c..0086113 100644
--- a/meta-dream/recipes-bsp/drivers/dreambox-blindscan-utils_1.7.bb
+++ b/meta-dream/recipes-bsp/drivers/dreambox-blindscan-utils_1.7.bb
@@ -1,6 +1,8 @@
 SUMMARY = "Utilities needed to do transponder blindscan with dreambox dvb receivers"
 LICENSE = "CLOSED"

+PACKAGE_ARCH = "${MACHINE_ARCH}"
+
 PROVIDES += "virtual/blindscan-dvbs virtual/blindscan-dvbc"
 RPROVIDES_${PN} += "virtual/blindscan-dvbs virtual/blindscan-dvbc"

Attached Files

•  et-blindscan-dvbs-utils.bb.txt   622bytes   2 downloads
•  xtrend-dvbs-blindscan-1.0.tar.gz   3.03KB   8 downloads
•  dreambox-blindscan-utils_1.7.bb.txt   911bytes   2 downloads

[blzr]

oh, and after further investigation, seems that this commit:
http://sourceforge.n...043a751829d871/
did 'adapt' a bit to much,
blindscan binaries are not located inside the plugin dir anymore...

diff --git a/blindscan/src/plugin.py b/blindscan/src/plugin.py
index a123053..250b98c 100644
--- a/blindscan/src/plugin.py
+++ b/blindscan/src/plugin.py
@@ -29,7 +29,6 @@ except:
 from time import strftime, time
 
 XML_BLINDSCAN_DIR = "/tmp"
-BIN_BLINDSCAN_DIR = "/usr/lib/enigma2/python/Plugins/SystemPlugins/Blindscan/bin/"
 XML_FILE = None
 
 _supportNimType = { 'AVL1208':'', 'AVL6222':'6222_', 'AVL6211':'6211_', 'BCM7356':'bcm7346_'}
@@ -602,18 +601,12 @@ class Blindscan(ConfigListScreen, Screen):
                        status_box_start_freq = temp_start_int_freq + uni_lnb_loc_osc[band]
                        status_box_end_freq = temp_end_int_freq + uni_lnb_loc_osc[band]
 
-               try:
-                       if os.path.exists(BIN_BLINDSCAN_DIR):
-                               os.system('chmod 755 %s/*' % BIN_BLINDSCAN_DIR)
-               except:
-                       pass
-
                if boxtype.startswith('vu'):
                        try:
-                               cmd = "%s%s %d %d %d %d %d %d %d %d" % (BIN_BLINDSCAN_DIR, self.binName, temp_start_int_freq, temp_end_int_freq, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid))
+                               cmd = "%s %d %d %d %d %d %d %d %d" % (self.binName, temp_start_int_freq, temp_end_int_freq, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid))
                        except: return
                elif boxtype.startswith('et'):
-                       cmd = "%savl_xtrend_blindscan %d %d %d %d %d %d %d %d" % (BIN_BLINDSCAN_DIR, temp_start_int_freq, temp_end_int_freq, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid)) # commented out by Huevos cmd = "avl_xtrend_blindscan %d %d %d %d %d %d %d %d" % (self.blindscan_start_frequency.value/1000000, self.blindscan_stop_frequency.value/1000000, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid))
+                       cmd = "avl_xtrend_blindscan %d %d %d %d %d %d %d %d" % (temp_start_int_freq, temp_end_int_freq, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid)) # commented out by Huevos cmd = "avl_xtrend_blindscan %d %d %d %d %d %d %d %d" % (self.blindscan_start_frequency.value/1000000, self.blindscan_stop_frequency.value/1000000, self.blindscan_start_symbol.value, self.blindscan_stop_symbol.value, tab_pol[pol], tab_hilow[band], self.feid, self.getNimSocket(self.feid))
                print "prepared command : [%s]" % (cmd)
 
                self.thisRun = [] # used to check result corresponds with values used above

on my et9500 this modified blindscan plugin.py works, but someone would need to check what's with vu+ boxes /?/

Attached Files

•  plugin.py   44.01KB   22 downloads

[Huevos]

And what do you think of the plugin?

[Erik Slagter]

blzr thanks for looking into this.

 

Please know that blindscan support for the first generation of xtrend receivers has been very flaky for some time now. From the moment the bsp layers is managed by xtrend itself, the necessary closed binary has been on-and-off in the blindscan package. I am not aware of the current situation because I don't have such a receiver. If it's still broken, xtrend will have to fix it. It looks like the recipe isn't completely correct to start with. But that would mean that the "old" blindscan plugin would be nonoperative as well.

 

The reason the dreambox-blindscan-utils are not "MACHINE_ARCH" is, IIRC, that it contains a generic cable tuner blindscanning utility, which is also used for other manufacturerers using this tuner as well.

 

So I guess the first step is to confirm if the xtrend blindscan package really installs the binary and check where it installs the binary. Maybe that's the same problem with the vu+ models.

[blzr]

From the moment the bsp layers is managed by xtrend itself, the necessary closed binary has been on-and-off in the blindscan package...If it's still broken, xtrend will have to fix it.

yes it is broken, and should be easy to fix (as I posted above)
 

It looks like the recipe isn't completely correct to start with. But that would mean that the "old" blindscan plugin would be nonoperative as well.

the 'old' blindscan 3rd-party plugin included all the blindscan vu/xtrend binaries, so whether they were supplied by the blindscan utils package was irrelevant, plugin worked...
 

The reason the dreambox-blindscan-utils are not "MACHINE_ARCH" is, IIRC, that it contains a generic cable tuner blindscanning utility, which is also used for other manufacturerers using this tuner as well.

but, because of that mips32el arch you have something like I described here:
http://forums.openpl...e-7#entry478567
(and, maybe for these 'other manufacturers' should be provided separate blindscan utils packages, and they shouldn't rely on dreambox-blindscan package /?/)
 

So I guess the first step is to confirm if the xtrend blindscan package really installs the binary and check where it installs the binary. Maybe that's the same problem with the vu+ models.

as it is now - xtrend blindscan binary is not installed at all;
with the changes to the et-blindscan-dvbs-utils recipe I suggested above - binary is installed in standard /usr/bin location,

as for the vu boxes, I don't know, probably you can tell me if/where the binaries are installed after installing virtual/blindscan-dvbs

[Erik Slagter]

Clear blzr, thx. I didn't realise all binaries were included in the old plugin, makes it even more ugly that I thought.

 

I will ask xtrend to fix the package.

 

I will try to find out who's using the dmm blindscan utils and will try to fix it.

 

And last but not least, the blindscan plugin needs to be changed so it looks for the binaries in the right place. I guess dimitrij could make a patch.

 

That's it?

[Erik Slagter]

blzr, just had a look at the xtrend the recipe, I don't think it works like you think

 

What they do is they download a complete ipk package and install the complete package into the package dispatch area. I am not sure it's that "correct" but I guess it works. I will check that.

Edited by Erik Slagter, 5 March 2015 - 10:54.

[blzr]

yeah, it 'works' exactly as you described i.e. doesn't work now (of this I can be pretty sure, basing on my own experience )
- 'our' new blindscan plugin depends on virtual/dvbs-blindscan
- virtual/dvbs-blindscan for xtrend is should be provided by et-blindscan-dvbs-utils package, which package doesn't exist as such
- instead it tries /?/ to install enigma2-plugin-systemplugins-blindscan (v.1.4, wih the binary inside plugin dir btw), while there's already enigma2-plugin-systemplugins-blindscan (git3963... version) installed...
 

[Erik Slagter]

It seems the problem is that xtrend provides a package with the same name as the generic package we build lately. And the plugin we build misses the dvb-s-blindscan dependency.

Edited by Erik Slagter, 5 March 2015 - 11:59.

[Erik Slagter]

Both packages are built, btw, but apparently the xtrend one is installed (1.4 vs git).

 

I will try to convice xtrend to supply a package that only has the binary and nothing else, ok?

[blzr]

well, I have no idea what was the purpose of such weird solution in xtrend's recipe, but I think that thanks to these problems now it's great oportunity to civilize this, finally...

[Erik Slagter]

Here with the vu+duo2 it works then the path to the blindscan binary is changed to /usr/bin/.

 

I'd rather see "/usr/bin" here instead of "/usr/bin/" and add the slash where it's used %s%s -> %s/%s

 

Also I am a bit disappointed found transponders are not shown during the process, that's something my simple version did so it can't be that difficult to implement.

 

The dependencies for the plugin are correct in the source:

Package: enigma2-plugin-systemplugins-blindscan
Version: 1.1-r0
Description: blindscan
Maintainer: Huevos (OE-A)
Depends: enigma2, virtual/blindscan-dvbs

But it seems they don't get included in the package.

[Erik Slagter]

well, I have no idea what was the purpose of such weird solution in xtrend's recipe, but I think that thanks to these problems now it's great oportunity to civilize this, finally...

I think I do have an idea but it's a bit strange though.

 

All we can do is ask xtrend, in the end.

[Erik Slagter]

Request to xtrend is on it's way.

[metoo]

can you also do a request for the DMM Remote controls etc for ET7x00 please?

[Erik Slagter]

IIRC the addition of non-et rc's has already been requested. They're not by any means required to implement it though.

Edited by Erik Slagter, 5 March 2015 - 13:04.