Re: Ghostrscript/GhostPDL 9.18 Release Candidate 1

On 24 September 2015 at 11:23, Johannes Meixner <[email protected]> wrote:

>
> Hello,
>
> On Sep 24 11:32 Johannes Meixner wrote (excerpt):
>
>> I had a look at base/gximono.c but for me it is not obvious
>> what is actually meant with
>>  dev_value = color_cache[*psrc * spp_out];
>>
>> dev_value is a pointer to byte and I do not understand
>> the idea behind how that pointer value is calculated
>> so that currently I cannot provide a reasonable patch.
>>
>
> Instead of trying to understand that functionality
> in gximono.c I had a better idea and made a diff
> between gximono.c in Ghostscript 9.16 versus now
> in 9.18rc1 which shows the following:
>
> In Ghostscript 9.16 it was:
>  dev_value = color_cache + psrc[dda_ht.state.Q] * spp_out;
>
> In Ghostscript 9.18rc1 it is:
>   dev_value = color_cache[*psrc * spp_out];
>
> Because color_cache is also a pointer (to byte)
> "color_cache + <something>" results a pointer
> so that I assume the following is the right fix:
>
> In Ghostscript 9.18rc1 in gximono.c I replaced all
>   dev_value = color_cache[*psrc * spp_out];
> with
>   dev_value = &(color_cache[*psrc * spp_out]);
> so that the pointer and not the value gets assigned.
>

I looked at how the color cache is created, and came to the same conclusion
as you, I've put those changes in a commit, and will stick it up for review
by the relevant engineer (who is based on west coast USA).

Thanks!,

Chris

_______________________________________________
gs-devel mailing list
[email protected]
http://ghostscript.com/cgi-bin/mailman/listinfo/gs-devel