Re: a reminder about patches

Adam Megacz wrote:

>Two reminders:
>
>1. You MUST propose a patch if you want it to get okayed
>
>2. A patch can only be committed by the person who wrote it (this rule
>   applies to me too; I cannot/willnot commit patches I did not
>   write).
>  
>
Hold on...these were not gripes. These were a list of patches that I am 
currently working with. I posted this so that people other than me could 
get some work done.

>Out of your five gripes, three were already okayed and one
>  
>
By my count, one was already approved (charlie's initial size fix), one 
was never proposed (brian's backtrace patch), and three were not 
approved at the time I posted (530 fix, 538 fix, and the AWT performance 
UNDO).

>was never proposed (the fifth is a legitimate gripe, although I've had
>less than a day to reply since you posted it).
>
>Details follow:
>
>  
>
>>Sat Apr 10 23:00:38 PDT 2004  [email protected]
>>  * freetype.java cleanup (fixes 540)
>>Without this, the current nestedvm stuff barfs completely when trying
>>to render text.
>>    
>>
>
>I just okayed this
>  
>
I haven't seen any approval for this.

>  
>
>>Sun Apr  4 18:51:59 PDT 2004  [email protected]
>>  * brian's patch for fullpath backtraces
>>Old patch that is still applicable and helps debugging.
>>    
>>
>
>This was never proposed.
>  
>
That's fine. It wasn't mine to propose, and I wasn't paying close 
attention to it.

>  
>
>>Tue Apr  6 15:18:56 PDT 2004  [email protected]
>>  * fixes Bug 525 - correct initial frame size
>>I just added a test case for this. Charlie's fix works, though it does
>>make the window take up the entire screen in situations where it
>>shouldn't
>>    
>>
>
>http://lists.xwt.org/pipermail/patches/2004-April/000815.html
>  
>
Good...I'll ping charlie on it.

>  
>
>>Wed Apr  7 19:33:49 PDT 2004  [email protected]
>>  UNDO: improved Java2 AWT performance
>>This backs out the patch that makes font's virtually unreadable on
>>every platform except OSX.
>>    
>>
>
>Sorry, you're going to have to find a fix for it other than simply
>backing it out.  This is clearly a bug in the Win32 AWT implementation
>(if you believe otherwise, please explain).
>  
>
Well...it might well be a bug in AWT implementation, but if so, it's a 
bug in both Linux and Win32.  And you're welcome to reject the UNDO 
patch. Closing the bug and telling us not to reopeen it, on the other 
hand is utterly unacceptable.

>>Sun Apr 11 14:55:18 PDT 2004  [email protected]
>>  * respun fix for 538 to work with adam's latest patches
>>Updated fix for 538. Without this, boxes shrink to zero if
>>shrink=true, even if they have are filled with an image.
>>    
>>
>
>http://lists.xwt.org/pipermail/patches/2004-April/000842.html
>
>  - a
>  
>
I committted it two minutes after this approval.

This particular message of mine was a status, mixed with a request to 
help us get synched up. There are no patches of mine that you have 
approved that I haven't submitted, and I've never committed a patch that 
wasn't mine.

What has been extremely frustrating is the number of regressions. The 
last time I asked you about testing against anything in the widgets 
repository, you objected saying it was empty. Well, widgets.ibex.org is 
fully populated, and main.t, sampler.t, games.t, and alpha2.t all 
provide various ways to sanity check and prevent the many regressions 
I've seen in the last couple days. We're working on a more fomalized 
test suite, but in the meantime, please, please, please spot check 
against at least main.t and sampler.t.

-Tupshin