On Thu, 8 Nov 2012 22:08:32 -0200 Lucas De Marchi <[email protected]> wrote:
> On Thu, Nov 8, 2012 at 8:47 PM, Michael Blumenkrantz > <[email protected]> wrote: > > On Wed, 7 Nov 2012 13:46:36 -0800 > > "Enlightenment SVN" <[email protected]> wrote: > > > >> Log: > >> e/modules: Do not use recursive Makefiles > >> > >> > >> > >> Author: lucas > >> Date: 2012-11-07 13:46:35 -0800 (Wed, 07 Nov 2012) > >> New Revision: 78978 > >> Trac: http://trac.enlightenment.org/e/changeset/78978 > >> > > > > Wow, this is a huge patch which radically changes things. Nice idea. > > > > Prior to committing, however, I really wish you had: > > 1) Reviewed/tested this a bit more. The number of followup commits and bug > > reports is quite appalling considering that we are actually in a release > > cycle 2) Waited a few more days for the manager of this project (me) to be > > able to test/review/okay such a large change // asked for users and > > developers to test prior to committing > > 3) Spent a little more time fixing the known bugs that you mentioned in your > > previous email > > > > To reiterate, this is a good idea; however, I've already called out people > > for committing patches which did far less than this, so it would be wrong > > if I didn't do the same for you, who has much more egregiously abused svn, > > and during a release cycle at that. > > > > So, to clear things up for everyone who may have missed the implications > > which go along with going into alpha during our release cycles, consider > > this as the official message: > > > > We are in a feature freeze for E17. This should be treated the same as any > > previous feature freeze prior to a release. > > > > I will send another mail specifically on this topic to ensure that nobody > > misses it. > > > > Again, to re-reiterate, I agree with the changes you proposed and your > > method of doing it. And now I would greatly appreciate it if you could > > spend some time testing your work much, much more thoroughly. As soon as > > possible. > > I think you are not being fair. You're bashing me as if this was not > tested. It *was* tested much more than I wanted, but I was already > expecting some bugs for such a huge change. Even make distcheck > in-tree, out-of-tree, etc were tested. I think I'm being extremely fair. Reread my message; at no point did I bash you or your changes -- I was complimentary about both. > > Take a look in this patch and interdiff with the previous version I > sent to the mailing list. That was after hours of compiling and > testing. And it shows your dedication to making this feature work, something we all can appreciate. > > It's not the normal workflow of the project to send patches to mailing > list, but I even did that. And *you* acked it. If by "acked" you mean that I said "this is a good idea", then sure. But saying something is a good idea is a long way from saying "this patch is good". > > However I only use a limited amount of modules in E. I couldn't test > it more. Be fair, take a look in the follow up commits and you will > see that bugs were caused because of different versions of automake, > build out-of-tree in a directory not sharing a prefix, installing as > root so things like chmod work, external modules or modules that I > never used. As previously stated, I am being the most fair. The problems you cited would easily have been caught by asking for a bit more user testing. > > These could only be tested with a broad audience. I was confident > enough to commit and sit waiting on the mailing list and in IRC until > late in the night waiting for bug reports. People sent the bug > reports and they got fixed. Again, be fair and see the time frame > between the bug reports and the fixes. Sadly it was the only way to > get this tested. Much thanks to PrinceAMD, Seoz, Thanatermesis and > glima who pointed out most of the bugs. Again, I am being fair. I have complimented both your idea and the effort it took to fix all the issues that have arisen since your first commit. > > If it had not been done this way we would still have that first > version of the patch, with all the bugs it had. And the reason to get > this in *after* your talk and (possible) release of E was that I > wouldn't like to screw with it. No, if it had been done in the way that we do most very large changes -- keep patches/branches on the mailing list since we can't do such a thing in SVN -- the bugs would largely have been avoided in the first place. > > And yes, I'm surprised and really disappointed with your reaction. > Particularly because of the amount of bugs introduced in this very > same week and the past one in E and EFL makes me only think you are > taking me as the scapegoat. > > > > Lucas De Marchi Again, you seem to have ignored all the positive feedback that I have given you regarding this work from the first mention of it. You were not the only person that was called out on breaking things over the past week, just the only person on the mailing list; anyone who went to the dev day or LinuxCon can vouch for this. Since you seem to be skilled at reading only the negative points that I make, I will try to rewrite my points in a format that may be easier to understand: * THIS WAS A GREAT IDEA * I APPRECIATE YOUR WORK ON IT * IN THE FUTURE, PLEASE TEST A BIT MORE WITH THE COMMUNITY BEFORE COMMITTING * I AM GLAD THAT YOU ARE FIXING THE BUGS SO QUICKLY, AND PLEASE CONTINUE TO DO SO Thanks, and happy hunting ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_nov _______________________________________________ enlightenment-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
