Pandoc.brew creates unnecessary png-s

[cocinerox]

Pandoc.brew(text="<%=0%>",output="x.md")

creates

/plots/x.md-1.png (what is a white square).

Cheers, L (Ps: Meet you in a week!:)

[daroczig]

Wow, that's strange indeed. I cannot reproduce this bug, but IMHO this was introduced by some changes in recordPlot in R 3.0.2. Could you please verify your sessionInfo()?

[cocinerox]

Sure:

R version 3.0.2 (2013-09-25)
Platform: i386-w64-mingw32/i386 (32-bit)

locale:
[1] LC_COLLATE=Hungarian_Hungary.1250 
[2] LC_CTYPE=Hungarian_Hungary.1250   
[3] LC_MONETARY=Hungarian_Hungary.1250
[4] LC_NUMERIC=C                      
[5] LC_TIME=Hungarian_Hungary.1250    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods  
[7] base     

other attached packages:
[1] pander_0.3.9

loaded via a namespace (and not attached):
[1] digest_0.6.3 tools_3.0.2 

[daroczig]

Cool, thanks! Compiling R 3.0.2 in the near future...

[cocinerox]

Managed to reproduce it with R 15.2 & pander 0.3.2:

R version 2.15.2 (2012-10-26)
Platform: i386-w64-mingw32/i386 (32-bit)

locale:
[1] LC_COLLATE=Hungarian_Hungary.1250 
[2] LC_CTYPE=Hungarian_Hungary.1250   
[3] LC_MONETARY=Hungarian_Hungary.1250
[4] LC_NUMERIC=C                      
[5] LC_TIME=Hungarian_Hungary.1250    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods  
[7] base     

other attached packages:
[1] pander_0.3.2

loaded via a namespace (and not attached):
[1] digest_0.5.2

[daroczig]

Hm. Did you ever see such issue before? I have not used pander on Windows for production, but getting my hands on a Windows VM to verify.

[cocinerox]

I suspect that this happened earlier as well, but did not investigate it...

[daroczig]

Yeah, indeed. And it also happened when you did not specify the output parameter, but the strange 1px images ended up in the tempdir() on Windows, so that you would not realize this issue :)
Will resolve this soon, thanks again.

[cocinerox]

Thanks for the fix, works ok! (However, I wonder why these imgs are created at all.)

[daroczig]

Short demo on the reason why those dummy images are created on Windows:

• Linux demo:

> t <- tempfile()
> png(t)
> recordPlot()
> dev.off()
null device 
          1 
> file.exists(t)
[1] FALSE

• Windows demo:

> t <- tempfile()
> png(t)
> recordPlot()
> dev.off()
null device 
          1 
> file.exists(t)
[1] TRUE

[cocinerox]

Further investigated in Windows:

> t <- tempfile()
> png(t,type = "windows") # default
> dev.off()
null device 
          1 
> file.exists(t)
[1] TRUE
> 
> t <- tempfile()
> png(t,type = "cairo")
> dev.off()
null device 
          1 
> file.exists(t)
[1] FALSE

Maybe a global option for custom type would be advantageous.

[daroczig]

After all, I decided to default to cairo instead of a custom global option, as there is no sense in using the windows type instead of cairo if the latter is available. As a matter of fact, since pander depends on R 2.15, this check is rather needless, as it seems all modern Windows builds do support cairo.

Hopefully this will not screw up something on Mac OS :)
Will verify tomorrow. Thanks again! And would love to hear your feedback if this workaround would not be okay, neat or something.

[cocinerox]

Nice, thanks!