[cups.development] Port from Debian [Squeeze] to Ubuntu 12.10

[Peter M. Groen]

On Thu, 18 Apr 2013 06:19:49 -0700, Michael Sweet wrote:

> Peter,
> 
> On 2013-04-18, at 8:46 AM, "Peter M. Groen" <pgroen at osdev.nl> wrote:
>> 
>> This question is asked many times on the different forums, but no
>> answer found...
>> 
>> I'm switching from Debian (cups 1.4.4) to Ubuntu (cups 1.6.1) and
>> suddenly my compiler is complaining about "forward declarations" in
>> ipp.h and I'm a bit stuck at the moment.
>> ...
>> This produces the following errors:
>> 
>> -------------[Compiler output]-------------
>> ../lib/printer/printerutil.cpp:xxx:x error: invalid use of incomplete
>> type 'ipp_t {aka struct _ipp_s}'
>> -------------[/Compiler output]-------------
>> 
>> There is no reason (as far as I can see) why this happens. Is the
>> interface changed that much? Do i miss something?
> 
> 
> This change is listed in the "what's new in CUPS 1.6" help document;
> basically we added accessor functions to libcups and made the ipp_t and
> related structures private in CUPS 1.6.  Aside from making the API a
> little more future-proof, the change also helps to avoid common coding
> errors that occurred with developers manipulating the structures
> directly.
> 
> For legacy code you can add "#define _IPP_PRIVATE_STRUCTURES 1" before
> including <cups/ipp.h> or add "-D_IPP_PRIVATE_STRUCTURES" to your
> compiler flags, however that should only be used as a last resort -
> future versions of CUPS will remove this workaround and require you to
> use the accessor functions.
> 
> For this code I would recommend using the ippNewRequest function (added
> in CUPS 1.2, so it has been around for a very long time) to create your
> IPP request and set the header fields appropriately:
> 
>     ipp_t *request = ippNewRequest(IPP_GET_PRINTER_ATTRIBUTES);
> 
> Among other things, ippNewRequest initializes the IPP version and
> assigns a valid unique request ID for each new request.  Your code can
> assign an out-of-range number (<= 0 or > 2^31-1) depending on the
> implementation of rand() and doesn't properly account for newer versions
> of IPP - two of the common coding errors I was talking about...
> 
> _________________________________________________________
> Michael Sweet, Senior Printing System Engineer, PWG Chair

Hi Michael,

Thank for an extended answer. This is lacking in all the other forums 
"out there". Seems to me I'll have my work cut out for me then. 

Thank you and I'll f'up if things are building again.

Kind regards

-- 
Peter M. Groen