It took me a very long time to write all of the Quicktime code for pix_film/record/video and have it run well. I think pix_record took a few months to write working on it nearly full time. After that it was a lot more work to have them all run for months at a time without crashing. I was being paid to do it - along with the DirectShow code and a bunch of other things.
A lot of the issues stemmed from how far apart Apple's intended use for the code and what GEM needs to do were. I would expect the newer frameworks that are much more limited than the original ones to be a decent struggle for performance and stability.
Chris
PS - I have no interest in writing code for anything Apple or any other desktop OS again.
On Mon, Oct 30, 2017 at 12:38 PM, cyrille henry ch@chnry.net wrote:
Hello,
I've check Iohannes github: There is not a lot's of release critical issue.
Image, film and video under osX and having a way to compile for window and the main issues.
There is also lot's of stuff regarding FTGL, but they look to be obsolete. At least, we will know if they are obsolete after being able to compile Gem on the concerned platform.
Since Iohannes is currently too buzzy to help, I think we should organize development in 2 parts :
1st : doing the minimum to have Gem usable on all platform. We could then release a v0.94 and use it to test on lot's of computer and reporting all bugs. Then, a 2nd development round could create a 0.94.1 bugfix release with a longer lifetime.
if there is no objection, I will have a look at the platform we could use to centralize financing and stuff. cheers c
Le 12/10/2017 à 10:34, cyrille henry a écrit :
hello Max
Le 12/10/2017 à 10:04, Max a écrit :
Bonjour Cyrille,
This sounds good, there was the idea of some bounty system years ago too. I think it's crucial for this process to
- break up the job in small manageable parts
yes, that's the idea of using the github bug tracker. Spiting jobs in bugs. (and use flags to create priorities)
- well define their scope and define what "job completed" entails
I think it imply that the push request is accepted, and the code is merged upstream.
- set up a system that doesn't come with a huge administration overhead
yes, this is a problem. The reduce amount of developers interested in this job will make things simpler.
It almost seem there should be a platform for this out there...
yes, I don't think administration will be a show stopper.
https://www.bountysource.com/teams/pure-data/issues
this one looks like it parses github issues.. kind of shady.
On 2017년 10월 12일 09:46, cyrille henry wrote:
Hello,
Since the "get money, paid a developer" idea arise on the mailing list, I spend last weeks trying to think how to organize the last bit of development that need to be done.
I think we need a clear roadmap : a list of all object / platform that need to be fixed. My proposition is to use/update the bugtracker and to organize everything that need to be done under the "release critical" tag. There is a bit of work there. For example "native film reading in osX/64 bit" is not mark as release critical. Also, there is 2 critical bug regarding W32 build : do we need to be able to compile Gem on mingw AND mscv?
With this list of "release critical" bugs, it will be easier to estimate the work that need to be done. It will allow to estimate the development time, and the bounty we could offer for this work.
With the financial help of New Blankets, other institutions, associations or users, I confident that we will be able to finance this work (or part of this work).
does this sound good?
cheers Cyrille
GEM-dev mailing list GEM-dev@lists.iem.at https://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@lists.iem.at https://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@lists.iem.at https://lists.puredata.info/listinfo/gem-dev
GEM-dev mailing list GEM-dev@lists.iem.at https://lists.puredata.info/listinfo/gem-dev