Hello folks, hello Carl-Daniel!
I am (again) trying to fix our development process. The development of 0.9.8 was an improvement already to what we did previously (since about the 0.9.4 release) although my own resources where quite limited in the second third of the cycle. But obviously it is not enough and we still have some huge problems (e.g. probing, 4B addressing) and I don't see how this is possible with the resources we currently have and use.
I have set out my plans already in the wiki: http://flashrom.org/Development_Guidelines
The summary is an additional branch named staging where patches are committed to allow for testing them by a broader audience without tainting trunk. Patches will be moved to trunk rather slowly, either after "enough" testing (and an additionally ack) or by the known reviewing process. This kind of mirrors what I do locally already but makes it more publicly.
The idea behind this is to still maintain the (at least) four-eyes principle but support it by provoking testing (and superficial reviewing) by non-regulars. My hope is that this helps to keep up our high quality standards with less reviewing strain on the regulars that can not invest as much time as they want to while allowing for faster integrating of changes.
I am happy to discuss changes to my proposal until the eventual release of 0.9.8, or the end of February, whatever comes first.
On Fri, 13 Feb 2015 09:07:51 +0100 Stefan Tauner stefan.tauner@alumni.tuwien.ac.at wrote:
I am happy to discuss changes to my proposal until the eventual release of 0.9.8, or the end of February, whatever comes first.
From the absence of comments on the topic I conclude there is a consensus in the community that the proposed approach should be implemented. I will do so starting on Sunday (2015-03-15).
On 10.03.2015 14:17, Stefan Tauner wrote:
On Fri, 13 Feb 2015 09:07:51 +0100 Stefan Tauner stefan.tauner@alumni.tuwien.ac.at wrote:
I am happy to discuss changes to my proposal until the eventual release of 0.9.8, or the end of February, whatever comes first.
From the absence of comments on the topic I conclude there is a consensus in the community that the proposed approach should be implemented. I will do so starting on Sunday (2015-03-15).
Excellent, thank you!
Having read the proposal again, I wholeheartedly agree with it. It is the best chance we have to get things in, and still keep the releases stable.
Regards, Carl-Daniel
-
Carl-Daniel Hailfinger -
Stefan Tauner