@James,

Since yesterday after last git pull I cannot compile anymore a working Indi.

Whatever I start ends up with "undefined symbol: _ZN4INDI9Telescope13ISSnoopDeviceEP8xml_ele_" error

3rdparty driver even do not compile and show "CMakeFiles/indi_eqmod_telescope.dir/eqmod.cpp.o:(.data.rel.ro._ZTV5EQMod[_ZTV5EQMod]+0x78): undefined reference to `INDI::Telescope::ISSnoopDevice(xml_ele_*)'" at compilation

Do you have same issue?

I wonder what I am doing wrong

============ After deeper verification ... I had some old files from previous builds not de-installed
I allways clean up previous builds with xargs rm < install_manifest.txt but something went wrong ... I'm really get old

You're not doing anything wrong. The code was updated to fix a C-language compliance issue

So update libindi and sudo make install and it should be good.

Thank you Jasem,

I finaly got to compile libindi.
all the 3rdparty drivers also except "indi-rtlsdr" (I suppose dependency for libdspau issue)
I will continue to search ...
================
A the end I don't know if I made something wrong or not but fact is that I had an old version of "libdspau.so" an "libdspau.so.1" in my /usr/lib/ directory.
After removal full compilation worked again.

Not that I need all the 3rdparty drivers, but I hate when something does not work ...

I upgraded today to the latest stable KStars and INDI.

The good news is that the driver no longer crashes, after the fixes for the tty_nread changes. So slewing is normal, ...etc.

The bad news: One problem: when starting an Align (1 star or 3 star), the 'Align Process' line does not change to the value returned from :A?#. It still says 'Manual Alignment Process Idle', instead of the return value from said USB command.

OnStep itself is not the problem. When I disconnect INDI and connect from the command line using picocom, then use the Android app to start an Align, and issue a :A?#, I get back 613#. So the OnStep side of things are normal. It is the INDI driver that does not read and/or report the result of :A?# correctly.

@ Khalid,

I think something is broken in the driver code.
I am not in position to test it but for sure it has someting to do with refresh of Align variables
I believe the OSAlignOn variable is reset somewhere (could not see where).
I did this conditional to minimize polling but think it is not necessary.
Have same issue in version 1.4 and fixed it by discarding test

===================== lx200_OnStep.cpp ======================
// Update OnStep Status TAB
IDSetText(&OnstepStatTP, "==> Update OnsTep Status");
if (OSAlignOn) //don't Poll if no Aligning
{
if(!GetAlignStatus()) LOG_WARN("Fail Align Command");
}
==================================================

@ Khalid,

I checked the last version of stable and can reproduce what you notice about align.
Sine I discovered the same issue in the development version I did suspect my code but the fact that Stable has the same issue make me think that something in Indi changed and affected the behavior.

Here my notes after debugging:
To maintain some status I use global variables defined as part of "class LX200_OnStep" in lx200_OnStep.h
The strange thing is that these variables are reinitialized each time a function from Indi is called as if there would be declared again and again ????
I checked older versions with the same code and this did work.
Moving the variables declaration out of the class did solve this issue but I still do not understand the problem so nothing is solved.

I suspect my coding approach is wrong and I missunderstand something in Indi.
Any idea?

these variables are
char OSStat[20];
char OldOSStat[20];

char OSAlignStat[10];
char oldOSAlignStat[10];
bool OSAlignProcess=false;
bool OSAlignFlag=false;
bool OSAlignOn=false;

char OSPier[2];
char OldOSPier[2];

---

Finally I found that the variables were overwriten via overflow by Indi, so all of them should be sie of RB_MAX_LEN.

Just move definition of RB_MAX_LEN into header instead of cpp

char OSStat[RB_MAX_LEN];
char OldOSStat[RB_MAX_LEN];

char OSAlignStat[RB_MAX_LEN];
char oldOSAlignStat[RB_MAX_LEN];
bool OSAlignProcess=false;
bool OSAlignFlag=false;
bool OSAlignOn=false;

char OSPier[RB_MAX_LEN];
char OldOSPier[RB_MAX_LEN];

---

Thanks for finding the root cause.

So, will you create a pull request for the fix against master, or should I do that?

Khalid,

I tried to create branches on git to have multiple versions together but I discovered it is not as easy as that.
I tried to create the pull request starting from github.com/azwing/indi/tree/ONStep1.3 (the branch where I put the fix for version 1.3) but it ended up in a disater .
So yes if you could do the pull request would help, I am sure you manage git much better than me.

Thanks

Done.

Tested it, and I see the status of the alignment again.

Here is the PR for Jasem.

github.com/indilib/indi/pull/687

Alain,

Jasem merged the fix into master. So that part is taken care of now. The next stable should have the fix.

Did you get a chance to test the plate solve align with OnStep (the patch that I created a while ago)?

Unfortunately no, only test was indoor but after the fix of yesterday and looks like it works as expected.

Hope I can put everithing togetehr before winter so I can do intensive testing in the observatory even without motorization of roof but this is another story.
I really need to finish the works before snow is comming ....