On Jun 12, 2011, at 7:21 PM, Josh Andler wrote:
2011/6/12 Jon Cruz <jon@...18...>
First and foremost I was pointing out that existing code already allows one to achieve the needed goal without adding any changes to the library. Pushing out changes to app code that require end users to wait for their distros to pick up new versions of libs (or, more often, to wait and then upgrade to a newer distro to get newer libs) is often a poor choice.
Often a poor choice? Yes. Sometimes for the betterment of the project? Yes. 0.49 or whatever version it will end up as is truly a game changer for us. New versioning scheme, it appears the cmake cross-platform building will take, cairo renderer, purging our internal copies of libcroco (of which Abhishek is now a maintainer) & libgdl... to me, this release will really be one of those "you're best off on the most up-to-date distros" releases. Will we require it? No. Will there possibly be known issues with older libraries. Yes. I'm making this as a general statement, not just in regards to any patches we may be waiting for gdl to accept.
That is true... however the main point in this specific case is that the functionality under discussion (icons) can be achieved without changing any API's and pushing things out to external libs. And combined with the fact that the one change being looked at introduces many problems, avoiding a lib API change can be good here.