On Fri, Aug 15, 2003 at 08:28:35PM -0700, Tim Rowledge wrote:
> 3.6 #5387 image.
> Run 'BaseImage Tests #12'
> 

> 
> FileDirectoryTest>testExists was checking completely the wrong directory
> in the final assert. Fixed. (Still fails on unix becase of vm caching
> the directory entry (?))
>
Hmm... test doesn't fail on macos X. (your changed test works, too). 
Need to look closer...
 
> TestsForTextAndTextStreams>testReplacement1 was failing because (I
> claim) Text>replaceFrom:to:with:startingAt: was simply messing up the
> runs array by inserting the entire replacement's runs into thereceiver's
> runs. Fixed - I think.
> 
Ok, I'l repost this as a normal [FIX] to the list.

> TestForTextAndTextStreams>testExampleText2 fails because Text cannot be
> reversed correctly - there is no implementation of #reversed for Text
> and the SequencableCollection version is unsuitable.  #testExampleText1
> fails for similar reasons; growing the collection of a TextStream is not
> done even vaguely properly.
>
Maybe this could help? 

[ENH] Text instance method 'reversed' [sm]
http://swiki.gsug.org:8080/sqfixes/3429.html

> BCCMTest>test06bClassDescriptionAllSubInstances fails because it is
> plain weird.
>
I'l remove this.
 
> ByteArrayBugz>testByteArrayLongAt fails because it attempts to put -1 in
> a ByteArray. It works ok for +ve numbers.
> 
> BitmapBugz>testBitmapByteAt fails because Bitmap>byteAt:put: fails to
> ensure the input value is actually a byte (and happily screws up the
> answer). Potential fix is to mask the value, another possibility is to
> raise an error. From the look of the SUnit it seems that raising an
> error is anticipated but somebody else should consider what was wanted.
> 
These tests were posted together with fixes, the fixes are allready
in the updatestream.

-- 
Marcus Denker marcus at ira.uka.de  -- Squeak! http://squeak.de