128 replies to this topic

[christophecvr]

@All

 

Many already have seen the  previous topic :

 

http://forums.openpl...olock-function/

 

You can see the fuss I needed to create to have any attention unfortunately still not enough.

 

At first to all You should thank Theparasol, he's the first to draw the attention to a possible issue cause eDebug uses lock function (unfortunately required for every debug printing) . And that e2 is using the eDebug a bit at wild.

 

With this message I just correct the main tittle off the subject (edebugnolockfunction). That was wrong and adding a nolock to eDebug was one off my really not to do things, I already said why in the topic self.

 

I here just add an updated e2 patch against last enigma2 for some severe wrong buggy lines in servicemp3.cpp. Off course the lines are only commented. Since they could be perhaps even should be reset with a new dbug option the day e2 has been updated with a correct debug function which only runs on users demand .

 

This last is .... Adding a brandnew piece off code which makes developers and testers able to enable or disable debug utility's based on module used and level. That's not a easy task and needs the full support of every pli member.

 

If some off the devs (actually a lot off them ), do continue to hide themselves behind some expressions like :

 

- Wrong attitude (sorry I'm the messenger and do have a lack off diplomacy but I'm 100 % honnest)

- I do have no problems ?? (well that's just you're test in a to limited area)

- It's the driver (sometimes it is trough but so what in not more then 10 % of the cases)

- and ..... so what a lot off complete things which make no sense .

 

What to do : .....

 

Well in this mail a patch to e2 which only does comment some lines who do cause problems AND THEY DO in to servicemp3.cpp player.

There are other modules who do cause general problems just due to the use off eDebug.

 

If this can't be seen by the pli developers , core members and or moderators. Then it's useless to even think about new advanced development.

 

Its nice to have exemplary functions into c/c++ modules used by e2 but useless if they do not work or worse if they have have critical boolean bugs into them , they are caused by for example the use off eDebug.

 

Well I hope that some off the "devs core members and ..." Do an effort in order to see the REAL problem and also UNDERSTAND it if not  then ........

 

At this time we are at the if not , ... (lucky that I build my own image I solve problems for my self but other users ...)

 

 

Attached Files

•  e2-servicemp3-some-eDebug-commenting-update.patch   4.38KB   29 downloads

[MastaG]

Thank you chris.
Will add it in for sure :-)

[christophecvr]

Thank you chris.
Will add it in for sure :-)

I'll hope that the current pli "important members do see the problems and unstable e2 due to it , the arrogant kings of them unfortunately do not " 

[littlesat]

Does it solve real mean big issues! That is why we only commented out the most important one.... This because the real solution is working with debug levels.... But that takes some time to develop...

I think we should not push this tempory work-a-round that is candidate for revert later when it has no real concrete added value... The buffer percentages did had indeed hadbig proven impact..! That was why it was accepted

A good complete patch with eDebug levels is what we really want...

To accept this interim solution we need to be convinced with some evidence...

We really do understand your effort and time you spend with this topic...!!!! And other openpli topics....

Edited by littlesat, 2 February 2016 - 22:01.

[Taapat]

@littlesat but what is wrong if it's commented some debug messages? I see that in code at some already is notes that they are "extra debug info for testing purposes CVR should_be_removed later on"?
They can be removed, but maybe for developers is better that they remain commented.

Of course, if introduced debug levels they may be able to stay, but then they should be changed to specify the debug level.
If they have commented, then later it makes easier to understand that they needed level only for developers.

 

I also do not like the christophecvr attitude, but it seems that about this question you are asking too much from him.

[christophecvr]

Does it solve real mean big issues! That is why we only commented out the most important one.... This because the real solution is working with debug levels.... But that takes some time to develop...

I think we should not push this tempory work-a-round that is candidate for revert later when it has no real concrete added value... The buffer percentages did had indeed hadbig proven impact..! That was why it was accepted

A good complete patch with eDebug levels is what we really want...

To accept this interim solution we need to be convinced with some evidence...

We really do understand your effort and time you spend with this topic...!!!! And other openpli topics....

Well actually you all saw that with the most direct issue is solved , that by the buffer .

But that from the subtitles has already be confirmed by MastaG (with a lot off testers behind that the general subtitle behaviour is much better and actually even works ok while before it just worked sometimes ok sometimes not and ...)

Then the start off mkv media which has a toc chapter entry but especially eighter by resuming a movie had sometimes strange issues which leaded to a partially missing off audio and especially subtitles. Those are really all gone after the patch I posted . This by duo2 and dm8000 and others ... leave it to fill up other boxes by MastaG.

 

But waht You do not understand :

- I do not wan't A work-around

-The patch IS NOT A WORK-AROUND it just stops unruly code from acting and screwing normal working. Due to the using off eDebug.

 

BUT I JUST USE COMMENTING BECAUSE : The day when a serious debugging feature is implemented into e2 we can add those line as a debug (which only runs when users asks to run in a certain debug level).

 

The point is.

 

Commenting any possible cause off to heavy debug for now. (and commenting now since and a later we perhaps will use the lines in a certain debug level xxx).

 

As for how to solve and add a good debug into e2 which for now is NOT there , we need to to it easy, and it is a very very difficult complicated job.

[christophecvr]

@Taapat

 

 

I also do not like the christophecvr attitude, but it seems that about this question you are asking too much from him

 

Very ?? questionable  remark.

 

Ok I do have a very straight forward and honest way to say what I discovered and propose (propose not force).

My remarks are not made to break ,bring down or even insult some one.

I'm just straight forward and really honest . That all with a tremendous lack off diplomacy , that's me only honest point.

 

If being straight to the point and honest is a bad attitude then ........, ???

 

 

but it seems that about this question you are asking too much from him

Explain this ???

[mirakels]

I'll hope that the current pli "important members do see the problems and unstable e2 due to it , the arrogant kings of them unfortunately do not " 

 

 

christophecvr please stop this attitude or I think people will introduce special code that will un-stabalise your enigma2 experience forever. You are in now way honest and 'straight' forward. You are rude and impolite and overreact.

 

There is no reason to talk like this. You will only make it worse. It is that someone pointed out this new thread of yours, otherwise I would not even have read it.

 

We are all just people having wifes/husbands, kids, work and finally hobbies (maybe the last 2 should be switched, but that depends). If you want to make life hard for yourself please go do that somewhere else. But respect other people hanging round here.

[christophecvr]

 

I'll hope that the current pli "important members do see the problems and unstable e2 due to it , the arrogant kings of them unfortunately do not " 

 

 

christophecvr please stop this attitude or I think people will introduce special code that will un-stabalise your enigma2 experience forever. You are in now way honest and 'straight' forward. You are rude and impolite and overreact.

 

There is no reason to talk like this. You will only make it worse. It is that someone pointed out this new thread of yours, otherwise I would not even have read it.

 

We are all just people having wifes/husbands, kids, work and finally hobbies (maybe the last 2 should be switched, but that depends). If you want to make life hard for yourself please go do that somewhere else. But respect other people hanging round here.

 

Looks like You are one off the "Arrogant Kings"

 

You are just trying to deny real facts discovered and proved 100 % . So NO  I will never stop if no corrective actions are taken to a very very server bug.

 

Perhaps You should go off for a while , and let it handle of by real experienced and or other Young volunteers to solve the current (very heavy bugs) .

 

I'm not one off the real experienced nor a very Young guy but .....

 

Anyway You're really strange I never never asked to introduce at wild what any code actually just did the opposed .

 

So sorry but ????? or perhaps better

[athoik]

See: https://github.com/O...6a5578789643408

[Taapat]

Explain this ???

 

Sorry for my bad English.
By "this question" I think your patch.

[christophecvr]

See: https://github.com/O...6a5578789643408

Well based on that here a e2 patch which makes use off new debug.

 

To solve the urgent issues caused by over debugging.

problems are :

- causing out off sync (at random) situations by use off srt subtitles.

- causing (again it's a random issue sometimes well other times not) auto off sync start audio/video segments.

- The worsted issues are the (and also at random unexplained and not logic full sandkeeper events sometimes followed by full stb freeze).

 

With this patch the standard users (using init 3). Will not be bottered by issues caused by over use off edebug.

 

testers who start from console :

 

When just using enigma2.sh , You will see the log like it is by the regular user after this patch.

When using enigma2debug.sh You will see the full log like it was before.

 

That's off course just a start but the urgent real buggy issues caused by the over use off eDebug and so on ... will be gone.

The testers yes You know that some freaky issues may appear when using enigma2debug.sh , so no problem.

Attached Files

•  e2-start-Debug-level.patch   3.18KB   22 downloads

[MastaG]

 

Now we start without debugging by default

 

Now we start without debu

[Taapat]

I do not know whether it is a good idea to create two almost equal scripts.
Maybe better is in enigma2debug.sh set some variable for ENIGMA_DEBUG_LVL and from enigma2debug.sh run enigma2.sh.

In enigma2.sh check this variable and if it exist add it or add ENIGMA_DEBUG_LVL=3 to LD_PRELOAD.

 

But maybe it would be better by default in enigma set ENIGMA_DEBUG_LVL to 3, and add different level from enigma2debug.sh if necessary.

[littlesat]

enigma2debug.sh is not required at all...... For those who want to debug something else should be done... namely init 4 -> Set ENIGMA_DEBUG_LVL environment variable -> init 3.... later it can be restored to 4....

 

In addition in the E2 code this could be arrange somewhere in a menu... change debug level.... (probably also while it is running > work in progress < )

 

I'm now waiting for a servicemp3.cpp patch that is using this new feature...

 

It can be done by a level... set it that way that they are currently ignored.... or by a tag....

 

This instead of commenting them out....

Edited by littlesat, 4 February 2016 - 16:23.

[christophecvr]

I do not know whether it is a good idea to create two almost equal scripts.
Maybe better is in enigma2debug.sh set some variable for ENIGMA_DEBUG_LVL and from enigma2debug.sh run enigma2.sh.

In enigma2.sh check this variable and if it exist add it or add ENIGMA_DEBUG_LVL=3 to LD_PRELOAD.

 

But maybe it would be better by default in enigma set ENIGMA_DEBUG_LVL to 3, and add different level from enigma2debug.sh if necessary.

That's What it indeed should become in future:

 

But actually in first time the really to much used debug (by standard user) should be stopped.

In normal operation enigma2 is started trough initittab (see gui start)

 

In future also the standard debug level 0 should be set but the first errors or events who leads to a enigma2 crash are adapted to well be printed always to log.

 

But indeed setting the default debug level to three now, could also be a could solution.

 

Testers who need then only have to start with

ENIGMA_DEBUG_LVL=4 enigma2.sh

But the regular user will have the standard which does not print eDebug.

 

But it's well trough that some eDebug lines in now should be changed from eDebug ( ...   ) to for example eLog( lvl, ...)

[christophecvr]

enigma2debug.sh is not required at all...... For those who want to debug something else should be done... namely init 4 -> Set ENIGMA_DEBUG_LVL environment variable -> init 3.... later it can be restored to 4....

 

In addition in the E2 code this could be arrange somewhere in a menu... change debug level.... (probably also while it is running > work in progress < )

 

I'm now waiting for a servicemp3.cpp patch that is using this new feature...

 

It can be done by a level... set it that way that they are currently ignored.... or by a tag....

 

This instead of commenting them out....

So then it's best to set standard debug level to three for now. users who wan't more just need to start with

ENIGMA_DEBUG_LVL=4 enigma2.sh

Attached Files

•  e2-debuglevel-to-3.patch   1018bytes   9 downloads

[christophecvr]

Maybe even better.

 

Here a patch (only servicemp3.cpp) where debug level (6) and (5) haven used.

 

5 which in future should be a real debug not supposed to be runned by standard user.

6 Trace mode maybe that heavy that it anyway will disturb normal streaming

 

So for now if enigma2 is runned normally, You will not have those 5 and 6 ones.

 

But iff You run :

ENIGMA_DEBUG_LVL=5 enigma2.sh (the extra debug is included may and does cause some delays who sometimes triggers sandkeepers)
ENIGMA_DEBUG_LVL=6 enigma2.sh (trace mode so heavy that it will cause other issues , but it's to study a specific problem)

Attached Files

•  e2-newdebug-only-servicemp3-adapted.patch   4.04KB   8 downloads

[christophecvr]

Here I implemented base setup for enigma2 debug level trough user interface.

 

Now A user can change debug level trough menu,settings,systemsettings,Userinterface  last line Enigma2 Debug Level.

 

Testers developpers can also start in console enigma2 with :

 

ENIGMA_DEBUG_LVL=<x> enigma2.sh

 

Which will supersede the enigma2 UserInterface setting.

 

 

Attached Files

•  e2-DebugLevel-Configurable-trough-user-interface.patch   3.89KB   7 downloads

[athoik]

Hi christophecvr,

Is there a reason that eConfigManager::getConfigIntValue not working and you implement getConfigString?

Also why not a simpler version to get debugLvl?
 

-	debugLvl = getenv("ENIGMA_DEBUG_LVL") ? atoi(getenv("ENIGMA_DEBUG_LVL")) : 4;
+	debugLvl = getenv("ENIGMA_DEBUG_LVL") ? atoi(getenv("ENIGMA_DEBUG_LVL")) : eConfigManager::getConfigIntValue("config.usage.e2_debug_level", 4);

Edited by athoik, 6 February 2016 - 18:09.