Cannot open some FCS files created with FCSalyzer

[Anonymous]

Hi Sven,
I worked on data with your great FCSalyzer, created FCS files with my analyzes. Surprisingly, when trying later to open them again, some open fine, some don't, with this error message: "File could not be opened! The header does not identify it as a FCSalyzer document! Opening the file with Gedit does not indicate any visible difference in the header. This is on v 0.9.4.
By curiosity, I downloaded the last version (0.9.6), and could not open any FCS file. The program does not react (no error message, no nothing). I am on Ubuntu 15.10, and java version "1.8.0_51".
Thanks,
Philippe.

[Philippe P]

Just if it may help, here are 2 FCX flles (I meant FCX files in the initial post, not FCS files, sorry for the confusion). 20141107 FACS results opens fine with 0.9.4, while 20141021 gives the error. None open under 0.9.6.

[Sven Mostböck]

Hi Philippe,

thanks for your comments. I am also happy that you like the program.

I took a look at the files you provided. Thanks for uploading them, I would not have had a chance to find the bug without them. First I should mention that version 0.9.4 is not supported any longer. The newer versions (0.9.5 and up) were not planned to open old FCX files. However, I changed that now. The new version 0.9.7 has some bug fixes and should now be able to open your FCX files.

Please use the newest version 0.9.7 for analysing your data. Give it a try and let me know if you encounter more problems.

Best regards,
Sven

[Sven Mostböck]

Say Philippe,

are you the same Philippe who had problems with setting the document size a few months ago? I take it, that that problem has disappeared by now?

Best regards,
Sven

[Philippe P]

Hi Sven,
Yes, and yes.
And FCSalyzer 0.9.8-alpha opens all my files.
:-)
Thanks again for your great job!

[Philippe P]

Another bug I recently bumped into, that curiously was not there some months ago:
On histograms, lines cannot be colored any more. On p5 of attached FCX files is such a case.
Philippe.

[Sven Mostböck]

Hi,

this does not seem to be a bug. Instead, the user interface is not really clear. On the histogram overlay on p5, all three histograms are set to "Multigate color". With that option, events are colored in the color of the gat they fall into. That makes sense for dot-plots and is the behaviour we usually expect there. However, it is useless for histograms and I should remove it here.

Instead, the "user-defined color" should be pre-selected. This option is needed, if you want to set the color for the histograms yourself. In your "Histo color OK.FCX", all histograms are set to this option and hence the color selection works.

This is a fail of the user interface on multiple points. First, the pre-selected option "multigate" does not make sense for histograms and should be removed. Second, the color selection boxes should be disabled or empty for the "applied gate" option, as this option means that the histogram uses the color as defined in the gate that is used on the histogram.

I will change the dialog in the next version, However, as this is not a critical bug, I will wait with the release of the next version and will repair multiple things there.

Thanks!

[Philippe P]

And if it can help, here is an example of a parallel FCX file for which color can be set and changed without any problem.

[Sven Mostböck]

Hi Philippe,

thanks for the bug info and the files, I will take a look into them. A few months ago, FCSalyzer was updated to include FCS3 data files. That switch needed a lot of internal changes and I guess that's where things broke.

I am always amazed at how many different things can go wrong in a program. I test the program myself as well, of course, but don't manage to find all bugs.

Please keep the reports coming.

Reagrds,
Sven

[Philippe P]

Thanks Sven,
There are indeed a few bugs that I found, but always with a way around them. So I don't want tu bug (!!) you with them. If you nevertheless want to tackle an annoying one:
The "Ctr," and "Ctr." shortcuts to move to the previous and next files. On a French keyboard , "Ctr," works as expected, but not "Ctr." (nothing happens). My system, Ubuntu 15.10, is US, just the keyboard is set to French. May be using the left and right arrows instead would solve what looks like a keyboard mapping language problem?

[Sven Mostböck]

You are right, this is probably a mistake with the key mapping. This post made me realive that "period" and "comma" are not good choices, as their location might be different from keyboard to keyboard.

However, the arrow keys are already used for moving (ALT-arrows) and aligning (CTRL-arrows) selected plots. To add a third option for changing datafiles would probably invite disaster. Users would probably align plots when they want to change datafiles, etc.

I have now changed this to CTRL-"page-up" and "page-down" for the upcoming next version. What do you think?

[Philippe P]

I was thinking 1) select object; 2) click left or right arrow, without Ctr. But that may not be possible. CTRL-"page-up" and "page-down" sounds good to me!