I took care of it:

github.com/indilib/indi/pull/1860/files

So I tried a fresh install using astro-soft-build and ended up with version Build: 2023-03-07T16:40:59Z
With this version, saving to XISF rather than FITS results in a crash in the Nikon DSLR driver.

Is this unexpected with this recent build? How can I determine if something is out-of-date and needs updating? The previously mentioned thread entries are too cryptic for me to see where I've gone wrong.

Thanks

Does it works with simulator?

yes, it does.

It works with for me with my Canon 1100D. I don't know if there is something specific for Nikon. If you could run it through ekosdebugger to get better idea where it does crash?

So obviously I have configuration problems. So to be up-to-date I just did a git pull (nothing to update) and tried re-running ./build-soft-latest.sh
and I get...

[ 17%] Building CXX object libs/indibase/CMakeFiles/indidriver_OBJECT.dir/indiccd.cpp.o
In file included from /usr/include/libxisf.h:22,
from /home/blackastro/astro-soft/indi/libs/indibase/indiccd.cpp:38:
/usr/include/libXISF_global.h:22:10: fatal error: QtCore/qglobal.h: No such file or directory
22 | #include <QtCore/qglobal.h>
| ^~~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [libs/indibase/CMakeFiles/indidriver_OBJECT.dir/build.make:412: libs/indibase/CMakeFiles/indidriver_OBJECT.dir/indiccd.cpp.o] Error 1

Not sure what is out of sync?

Thanks

Yes today we merged new version of LibXISF that doesn't depend on Qt. That error is from old version. So update libxisf first.

So, in trying to follow in your footsteps, I added the following to a copy of ./build-soft-latest-xisf.sh in line 20

# add libxisf support
cd "$ROOTDIR"
[ ! -d "libxisf" ] && git clone --depth=1 git.launchpad.net/libxisf
cd libxisf
git pull origin
[ ! -d ../build-libxisf ] && { cmake -B ../build-libxisf ../libxisf -DCMAKE_INSTALL_PREFIX=/usr -DBUILD_LIBS=1 -DCMAKE_BUILD_TYPE=Release || { echo "XISF lib failed"; exit 1; } }
cd ../build-libxisf
make -j $JOBS || { echo "XISF lib failed"; exit 1; }
sudo make install || { echo "XISF lib failed"; exit 1; }

running this script eventually produced errors later in the script, in compiling fitsdata
/home/blackastro/astro-soft/kstars/kstars/fitsviewer/fitsdata.cpp: In member function ‘bool FITSData::loadXISFImage(const QByteArray&)’:
/home/blackastro/astro-soft/kstars/kstars/fitsviewer/fitsdata.cpp:464:28: error: no matching function for call to ‘LibXISF::XISFReader::open(QString&)’
464 | xisfReader.open(m_Filename);
| ~~~~~~~~~~~~~~~^~~~~~~~~~~~
In file included from /home/blackastro/astro-soft/kstars/kstars/fitsviewer/fitsdata.cpp:42:
/usr/include/libxisf.h:357:10: note: candidate: ‘void LibXISF::XISFReader::open(const String&)’
357 | void open(const String &name);
| ^~~~
/usr/include/libxisf.h:357:29: note: no known conversion for argument 1 from ‘QString’ to ‘const String&’ {aka ‘const std::__cxx11::basic_string<char>&’}
357 | void open(const String &name);
| ~~~~~~~~~~~~~~^~~~
/usr/include/libxisf.h:358:10: note: candidate: ‘void LibXISF::XISFReader::open(const LibXISF::ByteArray&)’
358 | void open(const ByteArray &data);
| ^~~~
/usr/include/libxisf.h:358:32: note: no known conversion for argument 1 from ‘QString’ to ‘const LibXISF::ByteArray&’
358 | void open(const ByteArray &data);
| ~~~~~~~~~~~~~~~~~^~~~
/usr/include/libxisf.h:360:10: note: candidate: ‘void LibXISF::XISFReader::open(std::istream*)’
360 | void open(std::istream *io);
| ^~~~
/usr/include/libxisf.h:360:29: note: no known conversion for argument 1 from ‘QString’ to ‘std::istream*’ {aka ‘std::basic_istream<char>*’}
360 | void open(std::istream *io);
| ~~~~~~~~~~~~~~^~
etc...

I think I'm closer, in that I generated a libXISF.so.0.2.0
but I'm still clueless as to what is generating this current issue.

Cheers

KStars is not merged yet so you need to fetch it from my fork invent.kde.org/nou/kstars/-/tree/xisf0.2

I can confirm now that with a compile from kde Build: 2023-03-13T19:37:51Z
that XISF file format storage also works with the gphoto indi_nikon_ccd aka. the Nikon DSLR driver, just as you would expect

Thanks!

Cheers

Glad that it works now.

I have come across a couple of random frames where it looks like it doesn't read from the camera correctly and you are left with a weird image artifact. See attached example top half of the image is ok but the bottom half is messed up. I have seen this happen at least 4 times over several nights. Usually, the next image will be fine. I don't remember seeing this issue while using the FITS format certainly not as frequently as I have seen of late. I'm not ruling out my own equipment but it seems to have cropped up since I switched to XISF. I'm running INDI server on a fanless pc on Ubuntu 22.04.02 and EKOS via my laptop. Images are stored remotely on the Ubuntu PC. camera is an Altair Astro 26C

Has anyone else come across this issue?

Logging for attached affected image.
[2023-04-17T22:47:09.610 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - "Capturing 600.000-second image..."
[2023-04-17T22:47:09.618 GMT Daylight Time INFO ][ org.kde.kstars.indi] - Altair AA26CTEC : "[INFO] FPS rate limit is set to unlimited "
[2023-04-17T22:47:09.633 GMT Daylight Time INFO ][ org.kde.kstars.indi] - Altair AA26CTEC : "[INFO] FPS rate limit is set to unlimited "
[2023-04-17T22:57:12.915 GMT Daylight Time INFO ][ org.kde.kstars.indi] - Altair AA26CTEC : "[INFO] Image saved to /Light/NGC 5195_Light_600_secs_002.xisf "
[2023-04-17T22:57:12.916 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - "Remote image saved to /Light/NGC 5195_Light_600_secs_002.xisf"
[2023-04-17T22:57:12.931 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - "Received image 2 out of 40."
[2023-04-17T22:57:12.954 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - Dithering...
[2023-04-17T22:57:12.956 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - "Dithering..."
[2023-04-17T22:57:12.959 GMT Daylight Time INFO ][ org.kde.kstars.ekos.capture] - "Dithering..."
[2023-04-17T22:57:12.964 GMT Daylight Time INFO ][ org.kde.kstars.ekos.guide] - "PHD2: Dithering started."