Hi,
I'm encountering very strange slice+changes behavior and I am not sure wtf is going on... because it's showing that there are changes in untouched code... So imagine I've changed two packages "Spec-Core" and "Spec-Examples" If I create a slice for the package independently, then displaying changes looks fine Spec-Core Changes or Spec-Examples Changes so far so good, but if I in the slice maker select both packages... and then check the changes this shows up which is wtf... I didn't touch AbstractAdapter nor other classes. It's almost like it's comparing "Spec-Examples" against "Spec-Core" If I (after creating slice for both packages) select "Spec-Core" it's complaining about removed classes... which are from "Spec-Examples" However "Spec-Examples" still looks just fine What is going on? What am I missing? This is not the first time I am making slices, but I've never encountered this. Thanks, Peter |
Hi Peter,
Le 7 juil. 2015 à 14:19, Peter Uhnák a écrit : > Hi, > > I'm encountering very strange slice+changes behavior and I am not sure wtf is going on... > because it's showing that there are changes in untouched code... > > So imagine I've changed two packages "Spec-Core" and "Spec-Examples" > > If I create a slice for the package independently, then displaying changes looks fine > > Spec-Core Changes > <2015-07-07_14:09:56.png> > > or > > Spec-Examples Changes > <2015-07-07_14:10:26.png> > > so far so good, but if I in the slice maker select both packages... and then check the changes this shows up > <2015-07-07_14:11:02.png> > > which is wtf... I didn't touch AbstractAdapter nor other classes. It's almost like it's comparing "Spec-Examples" against "Spec-Core" It looks like the changes browser shows changes from the whole system. As workaround, I now select packages in the slice independently to do a diff. But, yes, it looks like something is broken with slices diff Christophe smime.p7s (5K) Download Attachment |
In reply to this post by Peter Uhnak
2015-07-07 14:19 GMT+02:00 Peter Uhnák <[hidden email]>:
I've already seen it when making slices, probably as far back as Pharo3. Per-package was fine, overall slice was marking plenty of changes wrong. Would you be able to reproduce it on purpose, or does it seems linked with that particular set of changes? Regards, Thierry
|
So is this just "visual" bug, or will it break things if I save it like that?
Well I've changed about ten packages and if I create the slice on top of any combination it looks broken. So at least to me it looks very consistent for any changes across two packages. Peter |
2015-07-07 14:53 GMT+02:00 Peter Uhnák <[hidden email]>:
I think it is just a visual bug (or something wrong in the way a slice collate the changes from its dependencies).
It would be great to have a test-case, that's all (and this is a lot). It seems the bug has been around for a while. Does anybody knows if other MC implementations (Squeak?) exhibit that bug? Thierry
|
In reply to this post by Peter Uhnak
Let us know. It would be good to see if we can reproduce it.
Stef Le 7/7/15 14:19, Peter Uhnák a écrit :
|
In reply to this post by Peter Uhnak
There is a bug on how it collects all definitions from all dependencies. @Peter can you fix this please, it is a missing yourself in MCWorkingCopy>>#loadRemoteDependenciesIn: ... ^ dependencies inject: OrderedCollection new into: [ :all :deps| all addAll: deps;yourself ] 2015-07-07 14:19 GMT+02:00 Peter Uhnák <[hidden email]>:
|
Hi Nicolai, I had to also do this (this is ugly and I shouldn't query class types, but maybe this should be fixed elsewhere anyway) MCDefinitionIndex>>addAll: aCollection aCollection do: [ :ea | (ea isKindOf: OrderedCollection) ifTrue: [ self addAll: ea ] ifFalse: [ self add: ea ] ] Because aCollection can now be also collection of collections. But after both those changes it looks like a correct changeset. Peter On Wed, Jul 8, 2015 at 11:59 AM, Nicolai Hess <[hidden email]> wrote:
|
You can write it like that:
MCDefinitionIndex>>addAll: aCollection aCollection flattened do: [ :each | self add: each ] Le 8 juil. 2015 à 12:45, Peter Uhnák a écrit :
smime.p7s (5K) Download Attachment |
In reply to this post by Peter Uhnak
2015-07-08 12:45 GMT+02:00 Peter Uhnák <[hidden email]>:
Yes?
|
Free forum by Nabble | Edit this page |