Challenges -- Who's havesting fixes.

stéphane ducasse ducasse at
Wed May 24 08:22:55 UTC 2006

Hi andreas

> Hi Stef -
> First, I'm not sure what you mean by "bashing". Really, if I bash  
> something or someone it looks usually quite different from what I  
> wrote.

Indeed, I know you! and you know me so I was overreacting as usual.

> I would even go as far and claim that what I wrote is something  
> that you could probably relate to if you were trying to do some  
> package maintenance independent from the image maintenance.
> But since you asked let's stay calm and look at a few things: For  
> one thing, I have not seen any message asking about new versions  
> recently; the last time I saw a message where you asked for it  
> ( 
> January/099843.html) I responded quite explicitly pointing to the  
> latest versions ( 
> dev/2006-January/099871.html) If there has been anything in the  
> meantime please point me to it.

I could not find the email I (dreamed) sent

I could only find that:
From: 	  stephane.ducasse at
	Subject: 	[ANN] 3.9 beta is out
	Date: 	4 mai 2006 23:13:47 HAEC
	To: 	  squeak-dev at
	Reply-To: 	  squeak-dev at

Here is the beta version of 3.9
What should we do now:

	- harvest fixes for MC for some little bugs with traits
	- harvest fixes of traits
	- put back backwards compatible canUnderstand: and respondTo:
	- test and fix all the tests :) (this is a call)
	- harvest all fixes that will come
 >>>	- integrate all the package team work
	- remove the ToolsPlus
	- and everything I forgot	

	From: 	  stephane.ducasse at
	Subject: 	Mail for the network/IO team
	Date: 	13 mai 2006 14:13:17 HAEC
	To: 	  squeak-dev at
	Reply-To: 	  squeak-dev at

Hi guys

are you harvesting changes for the network? Or is this adventure dead?
Should I harvest the network changes?


> Second, let's look at Mantis. For the issues reported in the  
> categories of packages that I maintain there is no indication that  
> they have any relations to the changes that happened downstream,  
> which indeed leads me to believe that, yes, I am the only person  
> harvesting these packages, dealing with the reported bugs and fixes  
> (the other reason is that I would expect anyone interested in  
> helping with it to contact me to get access to the upstream  
> repositories which I'd be happy to grant but which hasn't happened  
> yet).


> And really, there is nothing wrong with that - I said I'd deal with  
> the issues and I am as you can see if you look at Mantis and the  
> issues that were resolved since I said I'd do it.

Ok could you push some of your packages to me so that I try to do a  
pass of all the pending items
(shit the list start to be really long).

> [BTW, for those not following that process almost all of the fixes  
> have a discussion thread attached to it which goes to show that I  
> typically don't just take the fixes but rather go over them  
> carefully discuss the necessary details and integrate them  
> afterwards (or reject them; that happens as well) - a process which  
> takes quite a bit of time and care and which makes it particularly  
> frustrating to deal with downstream changes that just "show up" in  
> an image. I also try to keep the packages in sync between the  
> various versions that I use (Squeak, Tweak, Tinlizzie, Croquet) and  
> that's even harder].

I really want to help the people and would like to avoid as much as  
possible to get fix that should not
be in. Less work for me less for you. Do you have some examples that  
I understand what crept in?

> Third, when I was looking at the latest 3.9, I found that all but a  
> single package that I am involved in with maintenance one way or  
> another (Balloon, Compression, FFI, Flash, Graphics, ToolBuilder,  
> TrueType; the single exception being the GraphicsTests package)  
> have downstream modifications that are not in the upstream  
> repositories.

Sorry you confused me with downstream and upstream. You mean that we  
pushed stuff in the image packages. Normally since we agree on that  
we payed attention not to do it. This is strange because we skipped  
all the packages you maintained.

> Now, I can understand why this is *sometimes* necessary, but when I  
> start looking at the changes in detail then I found that many of  
> them aren't required by other changes (which is the main reason why  
> I could see that they had to be included downstream) and that most  
> of these changes really should have been done upstream (if at all -  
> they really should be subject to the same review process that I am  
> running for changes reported at Mantis because otherwise they have  
> a really unfair advantage).


> And I can't for the heck of it find even a single Mantis report at  
> all for any of these changes (and again, if there are, please point  
> me to it) or even an attempt to communicate about these changes  
> (e.g., no emails that I could find talking about these changes  
> either) - they just "appeared" in the image at some point (and  
> again if there is a place where these things are discussed please  
> point me to it).

Strange I would like to understand which ones they are because I  
payed attention to not touch the packages you maintain.

> All in all, I believe this entitles me to say that for me it's  
> frustrating to see these downstream changes that are done outside  
> of the currently defined processes, which I believe include that  
> the responsibilities for harvesting done in a package lie with the  
> package maintainer not with the downstream image maintainer and  
> that changes should generally be documented at mantis.

Yes Ok I agree on that.

> And if I'm wrong about this, I'd really like a clarification to  
> understand what exactly the role of a maintainer is and what is  
> expect of this role. Because the way it is right now it *is*  
> frustrating to do package maintenance and all your yelling and  
> screaming won't really change that.

Normally the maintainer of the package is in charge of his package  
and most of the time
we should not have fix/changes in by other path. Now I would like to  
know the ones that creeps in
and we should have a way to communicate when something changes so  
that we have a control.

I will look at that next week (arghhhhhhhhh)
But I would appreciate that you let me know some examples so that I  
can understand
the why and how this happens.

Now if you can push some packages I would really try to release a new  
But been alone and with no time so far is difficult. I hope that the  
nex two weeks will be better.
But without internet (Yes this can happen, unbelievable).


> Cheers,
>   - Andreas
> stéphane ducasse wrote:
>> andreas
>> you mean you are harvesting your packages but nobody else!
>> Now I asked several time to get some new version of the maintained  
>> packages and I got
>> no answer so far.
>> So please do not bash us if we harvested a change that was for  
>> your package.
>> I do not like this kind of game!
>> Stef
>> On 24 mai 06, at 07:54, Andreas Raab wrote:
>>> Peace Jerome wrote:
>>>>>> All I wanted was an official answer to the question
>>>>>> "Who's havesting fixes."
>>>>> so far me and (marcus but he stopped).
>>>> Ok. Thank you for the direct answer. (And for taking
>>>> the responsibily.)
>>> Actually, this is only partially correct. Fixes that belong to  
>>> packages are harvested by the package maintainers. It is quite  
>>> frustrating for a package maintainer to see downstream  
>>> modifications magically appear in some image without even knowing  
>>> about them.
>>> Cheers,
>>>   - Andreas

More information about the Squeak-dev mailing list