Re: Help in using gsapi_set_stdio() to capture stdout into a memory buffer

If you look in base/gserrors.h, you'll see that -12 is a limitcheck error,

But, firstly, gsapi_run_string() is a really bad idea for anything
other than a basic string, because Postscript can contain binary data
(and in fact, Postscript itself can be binary encoded), so it will
fool functions like strlen() which we use in that API call.

So, *maybe* gsapi_run_string_with_length() would work.

Personally, I would use gsapi_run_string_continue() in loop, passing
in buffers of data - probably keep to about 4k buffer size, that way
your not going to hit any other string length limits that might still
be in there.

Chris




On 24 March 2016 at 06:18, Joaquim Luis <[email protected]> wrote:
> Hi again. I'm afraid I have to get back to this because I got stuck in a
> mystery.
> The attached code, sorry for the #ifdefs mess used for debugging, works when
> compiled as a standalone program (use the -DI_AM_MAIN pre-processor flag).
> It works only with small files where we can just do
>
> gs_rasterize ps_file.ps
>
> However, if I built it as a dll and send as argument the postscript code as
> a string, which will be called either from Matlab or Julia, than the call
>
>         gsapi_run_string (GSinst, ps, 0, &GSexitCode)
>
> returns an error code GSexitCode = -12
>
> I confirmed that the problem is not in the PS in memory with the code chunk
> under the #if 0. That is, that piece of code runs well from the stand-alone
> but fails when run from a dll.
>
> I'm completely clueless with this. Can it be because gs tries to do
> something that it has not right to do when controlled by another process
> (the one that calls the dll)?
>
> Thanks again
>
> Joaquim
>
>
>> At 16:17 17/03/2016 +0000, Joaquim Luis wrote:
>>
>>
>>> NOTE: the code file is attached, as it was in my first message.
>>
>>
>> I'm afraid (for me) your email ended up in a Google spam folder and I had
>> to dig it out manually. Somewhere in there the code obviously fell off,
>> apologies for that.
>>
>>
>>              Ken
>
>
> _______________________________________________
> gs-devel mailing list
> [email protected]
> http://ghostscript.com/cgi-bin/mailman/listinfo/gs-devel