On Thu, 2009-09-17 at 00:32 +0100, Steve Fryatt wrote: > I would imagine that there is a *vast* difference between bodging the > front-end to ensure that it keeps working with less dramatic (in API terms) > developments in the browser core, and completely restructuring the front-end > to make it work with what are (if my understanding is correct) very > significant changes to NetSurf's internal APIs.
There is, yes. I'd qualify the second half of the above with "parts of" -- most things can reasonably remain unchanged and still work (even if their current implementation leaves a lot to be desired). > If they didn't care about RISC OS, then I would imagine that dropping RISC OS > support from the browser core tomorrow would lose them a number of > developmental headaches. It would certainly simplify the buildsystem -- much of the existing one exists purely to support compilation on RISC OS (and working around the myriad quirks associated with that). The core code itself is entirely platform agnostic, so there's less to be gained there from dropping support for any given platform. > Personally, I'm grateful that a bunch of developers have worked out > how to develop a web browser on a platform that has proper > development tools (hence speeding up their work) and access to many > more developers that the RISC OS world can ever dream of, while still > enabling RISC OS to have its own native version of the finished > product with a native front-end. It turns out that such cross-platform development is no harder (and frequently easier) than developing for a single platform. In NetSurf's case, I can guarantee that it would be less developed, less robust, and less usable had development remained solely on RISC OS. Having access to such tools as valgrind and gdb is invaluable. The tooling on RISC OS doesn't even come close, and having to wait 20 minutes instead of 20 seconds for a full rebuild isn't likely to help development :) > All they're asking in return is that a RISC OS developer joins them to keep > the RISC OS bits up to date. Sorry, but that doesn't seem like a big ask to > me -- which is probably why I offered to help, although I don't seem to have > the time available right now either. I'm glad that at least one person "gets it". John.
