[lsb-discuss] Questions on standardizing "IJS"

Bastian, Waldo waldo.bastian at intel.com
Thu Aug 3 15:20:56 PDT 2006


>Instead of requiring ghostscript in LSB specification, does it make
>sense to just require it (or any other implementation providing IJS)
for
>testing? LSB will basically expect IJS and as it stands today, most
>distro basically will provide it through GS and hence satisfy LSB. Does
>this make sense?

You have to look at this from a printing workflow point of view. IJS is
only useful if you can install a PPD file that causes your IJS interface
to be called. And currently that means that your PPD file must make
reference to ghostscript.

If you don't want to pull in ghostscript I think you need to find a way
to tell the printing system via the PPD file "use some magic to print
this file to this IJS interface" without explicitly mentioning
ghostscript.

Let's call that magic "lsb-print-via-ijs-foo" for now. The question then
becomes if it is possible to replace things like

*FoomaticRIPCommandLine: "gs %B -q -dBATCH -dPARANOIDSAFER -dQUIET -dN&&
OPAUSE %A%Z -sOutputFile=- -"

with

*FoomaticRIPCommandLine: "lsb-print-via-ijs-foo"

and what kind of command line options lsb-print-via-ijs-foo should be
able to support?

Note that requiring magic LSB commands might prove to be just as popular
as requiring lsb-ld.so, so some caution might be in place here.

Cheers,
Waldo




More information about the lsb-discuss mailing list