On Thu, Jul 24, 2014 at 6:07 PM, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx> wrote: > On 24.07.2014 17:14, jscipione@xxxxxxxxx wrote: >> >> fef862e: TextView.h rearrage virtual methods for FBC > > As has already being said (and I told you in IRC yesterday) this is > superfluous. > >> 8e268f7: TextView: Fix FBC size breakage >> >> In hrev46796 I added two new private methods: _PreviousLineStart(), >> and _NextLineEnd() which increased the size of the class breaking >> binary compatability because I forgot to decrement the reserved array. >> >> This commit decrements the reserved array restoring the class size >> to the original size fixing the binary compat issue. > > Maybe you should ask if you don't understand something. I told you exactly > what you have to do, and you managed to do it wrong: > - you added boolean members, and that changed the class size. > - I even told you yesterday that methods don't change the class size > - The boolean members only account for a single int (ie., as I told you 356 > against 360 bytes size), so the array size must be 5, not 4 Axel, calm down and take a deep breath. I didn't figure out the problem at first but after some more research I learned what was going on and corrected the array size to be 5 rather than 4 in hrev47568. [1] > Is it really that hard to even read? Maybe rereading the IRC log before > breaking everything just a bit differently would have made sense? Unfortunately you mentioned the problem in the development channel which isn't logged AFAIK so I couldn't go back and re-read the discussion. Maybe you should read the entire commits log before sending off an angry email and then you'd know that the problem was already corrected. >> dc4ae0e: TextView: Fix regression for Home and End >> ... while Shift is held down in a selection. > > Ouch, you didn't even manage to get this one right. > THE BEHAVIOR IS BROKEN WITHOUT HOLDING DOWN SHIFT, TOO. And that was what I > mentioned to you. Again, calm down, I realize that there was a problem also when the shift key was not held down and I solved the problem in both cases. > At least nice that you fixed another regression, even though you seem to > have borked it, too. The BeOS solution definitely makes sense. Please > implement that one instead. I actually implemented the BeOS solution as well but I was and am reluctant to use it. There are pro's and con's on both sides and I am not convinced that the BeOS solution is superior. > Is it possible to forbid you to do any changes that change more than two > characters? This is really more than just a little annoying. I'm sorry that you are annoyed, all I can do is apologize and to try and fix any issues that come up as soon as possible. Also, changing the array size only involved changing 1 character and I managed to still screw that up so I don't think that would completely solve the problem. [1] https://cgit.haiku-os.org/haiku/commit/?id=d11a323120b7aae53f178877ca4c4cc59ab6a1d1