[squeak-dev] Strange Monticello behavior (at least for me)

Nicolas Cellier nicolas.cellier.aka.nice at gmail.com
Thu May 12 14:34:18 UTC 2011


I also saw this behavior when comparing changes of a merged package vs trunk.
I can't remember in what case exactly, but sometimes it looks like the
diff is performed against the elder merge ancestor and gives you a
wrong answer about changes.

Nicolas

2011/5/12 Alexander Lazarević <laza at blobworks.com>:
> Levente,
>
> that would be a perfect explanation, but unfortunately the quick fix I had
> in mind wasn't as innocent as I thought it would be, so shooting myself in
> the foot.
>
> Anyway thanks,
>  Alex
>
> 2011/5/12 Levente Uzonyi <leves at elte.hu>
>>
>> On Thu, 12 May 2011, Alexander Lazarević wrote:
>>
>>> I was just about to commit an innocent fix for MethodReference to the
>>> trunk,
>>> when Monticello presented me a list of >2000 changes it would actually
>>> commit. So I fetched the latest image from Squeak.org, updated the image,
>>> checked that everything was up to date and touched the classSymbol method
>>> again. And again I got that huge list of changes when I checked against
>>> trunk? As always I'm prepared to get a "you're holding it wrong" answer
>>> to
>>> this, but I wonder what's the deal.
>>
>> If the repository you're trying to save the package to (e.g.
>> package-cache, Inbox) doen't containt the ancestor of the given package,
>> then it will show more methods as changed. The worst case is when the
>> repository doesn't contain the package at all, so you'll see that all
>> methods have changed. In this case check the changes to the Trunk
>> repository, which (besides showing you the correct changes) will fetch the
>> ancestors of the package and save it to the package-cache. Then save your
>> package to the package-cache (it will show the correct changes in the save
>> dialog) and copy the saved version from the package-cache to the repository
>> you wanted to save it originally.
>>
>>
>> Levente
>>
>>>
>>> Alex
>>
>>
>>
>
>
>
>
>



More information about the Squeak-dev mailing list