@asemke, Could you confirm this R backend bug after applying and testing this patch with images? I have found a solution, that fix this for me, but the solution changes mime-types and I begin to doubt, that this is bug of Cantor and not mimetype errors in my system.

In D12987#265161, @sirgienko wrote:

@asemke, Could you confirm this R backend bug after applying and testing this patch with images? I have found a solution, that fix this for me, but the solution changes mime-types and I begin to doubt, that this is bug of Cantor and not mimetype errors in my system.

I can confirm that BUG 394018 and also the issues with the plots is fixed with your change. However, I'm seeing yet another inconsistency in Cantor's code with respect to the names of the temporary folders and files. Check for example maximaexpression.cpp:72 and similar for juliaexpressions and maybe for other backends. Should we remove the subfolder structure in rserver and go for a simple

m_tmpDir = QDir::tempPath() + QString::fromLatin1("/cantor_rserver-%1").arg(getpid());
QDir dir;
dir.mkdir(m_tmpDir);

?

In D12987#265244, @asemke wrote:

I can confirm that BUG 394018 and also the issues with the plots is fixed with your change.

If I understand you correctly, you have non empty output of the commands like d <- c(1,2,3), didn't it?

In D12987#265259, @sirgienko wrote:

In D12987#265244, @asemke wrote:

I can confirm that BUG 394018 and also the issues with the plots is fixed with your change.

If I understand you correctly, you have non empty output of the commands like d <- c(1,2,3), didn't it?

Yes.