Casper Hornstrup wrote:
-----Original Message----- From: ros-dev-bounces@reactos.com [mailto:ros-dev-bounces@reactos.com] On Behalf Of Alex Ionescu Sent: 9. marts 2005 22:17 To: ReactOS Development List Subject: Re: [ros-dev] Registry and ObjectManager (was FreeLdr Part II)
<snip>Talk about all the work that went into the branch</snip>
I'll try to be more clear since you seem to be confused about my points. I'm NOT arguing that the product is bad. I haven't looked at the branch yet, but I'm sure a lot of great work went into it. ReactOS is still a volunteer effort and everyone can (and should) be able to decide for him/herself what they want to contribute. This is a great motivation factor, but it's also a disadvantage so we should make sure that our processes minimize the damages that can happen from that.
What I AM arguing for is that THE PROCESS could be better. Traditionally there are two types of branches.
- Release branch. Usually created some time before a release to have a
more stable product at the time of a release. Only bugfixes go to this branch.
- Feature branch. Usually created to achieve a well-defined purpose
(e.g. the implementation of a new feature). Only changes required to achieve the well-defined purpose are committed to this branch.
Alex_devel_branch is a combination of the two and it has some disadvantages.
I agree, but I just want to clear up something.... this was never supposed to happen. I never planned for my branch to be ready for 0.2.6 release. However, both Steven and James wanted 0.2.6 delayed until the merge, so that some stuff they considered blockers could be fixed. This is why it became an ugly combination, and I know it's been frustrating.
- Bugs that that could have been fixed on trunk are annoying to trunk
developers. The bugfixes that go only to the branch are not in trunk (until merged there from the branch) where most developers work and is thus annoying to these developers.
This comment continues to seriously piss me off, but I will try to be as diplomatic as possible. The bugfixes that went in my branch are my own work, my own code, and wholy owned and copyrighted by me (just like any other code anyone writes). It was my decision to write that code, and it is my decision if I want to share it. There is nothing stopping me from deleting the whole branch tomorrow morning (except for Thomas's patch), if I feel like it. Of course, I will not do that, because I believe in the effort that everyone brings to this project, and I want to help out.
Those bugs have been in trunk *forever*. This, the developpers have been annoyed *forever*. If this were truly the case, then either:
1) All the developers would have nervous breakdowns by now, having been "annoyed" for the last 7 years. 2) They would've found and fixed the bugs, being higly "annoyed" by them.
This did not happen. Why? Because those bugs were 1) unknown to the developers, and also 2) not affecting them at this stage. Those who knew about it simply told themselves "it doesn't break anything for now, so someone will fix it later". The exception to this is Steven's bug.
So please don't try to tell me that developers are being "annoyed". Developers should be *thankful* that I took the time to review the whole Executive/Kernel (ex/ke) and fixed dozens of bugs and implemented new APIs, and that Thomas added more SEH to the functions. I have not once heard any developer speak of this "annoyance", so maybe it would be wise not to use "the developers" in your sentence.
I am really not trying to be an asshole here, but I consider the work that ANY developer does a gift. I consider your work on rbuild a gift to ReactOS and to me. I am not annoyed by the fact rbuild is not in trunk now, but that you've been using it for months. Instead, I thank you for your effort and dedication, and I anxiously await the result. I would give and expect the same attitude towards and from any other developer. You are not annoying me for having worked on your branch for months now, why am I annoying you for having worked on mine for two weeks? One could argue that the deficencies of "make" are much more annoying then some kernel bugs which won't come into play until later when we get serious about stability and drivers. I appreciate the fact you've decided to attack them, no matter how long it takes (btw, this also goes to Royce and everyone else working on it, as well as to Hartmut for Cc).
- Risk of duplicating work. We'll have more branches to track bugfixes on
so it's harder to know which bugs has been fixed and which hasn't.
I have tracked and merged all bug fixes, because I'm doing a good job at it. Instead, you should be happy that the bugs which plagued my branch did not affect trunk until they were fixed, which saved a lot more grief.
Casper
Best regards, Alex Ionescu