2013.11.05 15:41 "Re: [Tiff] memory management function calls", by Chris Liddell
On 05/11/13 13:28, Bruce A. Mallett wrote:
Very opportune: I'm just finding this thread as the result of a search (after having received an assignment to update from 3.8.x to a later libtiff). A project I am on uses libtiff in an embedded system containing multiple components. These components have different allocators so over the years similar modifications to libtiff have been made locally to provide a way for each component can pass in its own allocator (and deallocator, etc). The fact that your mod. provides for passing a memory context aligns well will our main use case: one of our allocators requires context info so that it can be told about the expected use & duration of the object to be allocated as well as a suggested pool to use.
Having our own set of local changes complicates updating to a newer release, it would be very beneficial to us if this kind of capability were baked into the distribution.
Perhaps with two of us pushing the use case, it might give added weight to possibly including the functionality.
If you haven't already done so, it would be helpful if you could have a look at the changes I chucked together and see whether it covers your requirements - obviously, my focus was how Ghostscript uses libtiff, so I'm aware that I may have glossed over things that others might need.
On 08/02/2013 04:04 AM, Chris Liddell wrote: > On 25/06/13 12:12, Chris Liddell wrote:
>> So, as I said, I took a punt at these changes (sorry for the delay - >> other projects etc....).
>> I created a browsable git repo here:
>> You can take a look at the diff there.
>> I don't *think* I've changed the existing ABI, where such would be >> required, I've added a new API call.
>> Comments, questions, criticisms, general rants all (somewhat) welcome... > So, not much love (or otherwise) on this?
> Not even a complaint about putting up the patch in an unfavoured form?
> I'd expected to be berated for something by now!