Bug#664966: Add an option to disable T3CCITT encoding

[Jonathan Nieder <jrnieder@gmail.com>]

tags 664966 + upstream
quit

Hi Vincent,

Vincent Bernat wrote:

> cupsfilters 1.0.7 include a fix for Brother printers to workaround a
> bug in the Postscript interpreter:
>
>      - pdftops: Additional workaround for a bug in the PostsSript interpreters
>        of Brother printers, calling Ghostscript with special command line
>        options (LP: #955553).
>
> Unfortunately, this fix relies on a patch in Ghostscript which is not
> applied in Debian. Please, apply it.
[...]
> --- a/base/gdevpdfb.c
> +++ b/base/gdevpdfb.c
> @@ -293,16 +293,18 @@
>       * that would need to be passed.
>       */
>      if (pres) {
> -        /*
> -         * Always use CCITTFax 2-D for character bitmaps.  It takes less
> -         * space to invert the data with Decode than to set BlackIs1.
> -         */
> -        float d0 = image.Decode[0];
> +        if (!pdev->NoT3CCITT) {
> +            /*
> +             * Always use CCITTFax 2-D for character bitmaps.  It takes less
> +             * space to invert the data with Decode than to set BlackIs1.
> +             */
> +            float d0 = image.Decode[0];
[...]
> --- a/base/gdevpdfx.h
> +++ b/base/gdevpdfx.h
> @@ -704,6 +704,11 @@
>      bool IsDistiller;
>      bool PreserveSMask;
>      bool PreserveTrMode;
> +    bool NoT3CCITT;                 /* A bug in Brother printers causes CCITTFaxDecode
> +                                     * to fail, especially with small amounts of data.
> +                                     * This parameter is present only to allow
> +                                     * ps2write output to work on those pritners.
> +                                     */

This is internal to the PDF-writing driver, so not an ABI change (phew).

Makes sense.  Please report this upstream at
http://bugs.ghostscript.com/ and let us know the bug number so we can
track it.

Thanks,
Jonathan