[PATCH] gnu: r: Add pango to inputs.

[PATCH] gnu: r: Add pango to inputs.

[Ricardo Wurmus]

Hi Guix,

our R package is currently built with Cairo, which allows R to use the Cairo
rendering backend.  When R starts up and loads grDevices, however, it checks
for the availability of Pango, not Cairo.  So even though the Cairo backend
works just fine, R falls back to using the primitive Xlib backend.

You can confirm this by running this in R:

    getOption('bitmapType')
    options(bitmapType='cairo')

The first command should show you "Xlib" with R from Guix.  The second sets
the backend to Cairo, which works without problems.

I decided against patching the detection mechanism in R because I don't know
if R might use Pango features elsewhere.  Instead I opted to add Pango to the
inputs.  Pango pulls in Cairo, so the package closure is a little bigger than
before.  Pango makes up 0.7% of the total size, so I think that's okay.

~~ Ricardo

Ricardo Wurmus (1):
  gnu: r: Add pango to inputs.

 gnu/packages/statistics.scm | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

-- 
2.8.4

[PATCH] gnu: r: Add pango to inputs.

[Ricardo Wurmus]

* gnu/packages/statistics.scm (r)[inputs]: Remove cairo, add pango.
---
 gnu/packages/statistics.scm | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gnu/packages/statistics.scm b/gnu/packages/statistics.scm
index 1288d9d..abc0c7e 100644
--- a/gnu/packages/statistics.scm
+++ b/gnu/packages/statistics.scm
@@ -165,7 +165,7 @@ be output in text, PostScript, PDF or HTML.")
        ("which" ,which) ; for tests/Examples/base-Ex.R
        ("xz" ,xz)))
     (inputs
-     `(("cairo" ,cairo)
+     `(("pango" ,pango)
        ("curl" ,curl)
        ("tzdata" ,tzdata)
        ("gfortran" ,gfortran)
-- 
2.8.4

Re: [PATCH] gnu: r: Add pango to inputs.

[Ludovic Courtès]

Ricardo Wurmus <ricardo.wurmus@mdc-berlin.de> skribis:

> * gnu/packages/statistics.scm (r)[inputs]: Remove cairo, add pango.
> ---
>  gnu/packages/statistics.scm | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/gnu/packages/statistics.scm b/gnu/packages/statistics.scm
> index 1288d9d..abc0c7e 100644
> --- a/gnu/packages/statistics.scm
> +++ b/gnu/packages/statistics.scm
> @@ -165,7 +165,7 @@ be output in text, PostScript, PDF or HTML.")
>         ("which" ,which) ; for tests/Examples/base-Ex.R
>         ("xz" ,xz)))
>      (inputs
> -     `(("cairo" ,cairo)
> +     `(("pango" ,pango)

Maybe mention “for the cairo BitmapType backend” in a comment so that
nobody views it as a thinko.

Thanks,
Ludo’.

Re: [PATCH] gnu: r: Add pango to inputs.

[Ricardo Wurmus]

Ludovic Courtès <ludo@gnu.org> writes:

> Ricardo Wurmus <ricardo.wurmus@mdc-berlin.de> skribis:
>
>> * gnu/packages/statistics.scm (r)[inputs]: Remove cairo, add pango.
>> ---
>>  gnu/packages/statistics.scm | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/gnu/packages/statistics.scm b/gnu/packages/statistics.scm
>> index 1288d9d..abc0c7e 100644
>> --- a/gnu/packages/statistics.scm
>> +++ b/gnu/packages/statistics.scm
>> @@ -165,7 +165,7 @@ be output in text, PostScript, PDF or HTML.")
>>         ("which" ,which) ; for tests/Examples/base-Ex.R
>>         ("xz" ,xz)))
>>      (inputs
>> -     `(("cairo" ,cairo)
>> +     `(("pango" ,pango)
>
> Maybe mention “for the cairo BitmapType backend” in a comment so that
> nobody views it as a thinko.

Okay, I added a comment and pushed.  Thanks!

~~ Ricardo

Re: [PATCH] gnu: r: Add pango to inputs.

[Ricardo Wurmus]

Ricardo Wurmus <ricardo.wurmus@mdc-berlin.de> writes:

> Hi Guix,
>
> our R package is currently built with Cairo, which allows R to use the Cairo
> rendering backend.  When R starts up and loads grDevices, however, it checks
> for the availability of Pango, not Cairo.  So even though the Cairo backend
> works just fine, R falls back to using the primitive Xlib backend.
>
> You can confirm this by running this in R:
>
>     getOption('bitmapType')
>     options(bitmapType='cairo')
>
> The first command should show you "Xlib" with R from Guix.  The second sets
> the backend to Cairo, which works without problems.
>
> I decided against patching the detection mechanism in R because I don't know
> if R might use Pango features elsewhere.  Instead I opted to add Pango to the
> inputs.  Pango pulls in Cairo, so the package closure is a little bigger than
> before.  Pango makes up 0.7% of the total size, so I think that's okay.
>
> ~~ Ricardo

I also just stumbled upon an old message to the mailing list, which I
had completely forgotten about:

   http://lists.gnu.org/archive/html/guix-devel/2015-09/msg00279.html

This is the very same problem and it is solved by giving R what it
wants: Pango.

~~ Ricardo