[UPDATES] 18 for Squeak3.7beta

dway at mailcan.com dway at mailcan.com
Sat May 22 05:58:36 UTC 2004

Latest approved items for 3.7beta.  This should be the second-to-last
batch of updates for 3.7beta... the next batch will be in about a week,
and we will move to 3.7gamma then and also open 3.8alpha.

If there are any really important outstanding fixes out there, now is
the time to reply to them to bring attention to them.  (Note that only
fixes for bugs introduced during 3.7 will probably be considered at this
late point, unless they're trivial fixes.)  And of course reviewing
fixes during the harvesting parties helps a lot.

- Doug


5906PLMOfManyModelAccess-hpt -- Hernán Tylim -- 5 April 2004
I found on PluggableListMorphOfMany an access to the model that wasn't
using the selectors that were given by the user but a hardcoded
Because this force the user to implement #listSelectionAt: on the model,
and because the class is already using a pluggable way to access the
selection, this .cs removes the hardcoded call (the only one in the
class in #itemSelectedAmongMultiple:) for a self listSelectionAt: which
access the model in the right way.
Also it fixes the class comment where was stated that the model must
implement #listSelectionAt: and #listSelectionAt:put:. The latter were
not used"

5907InspectElementFix -- Yoshiki Ohshima -- 9 March 2004
When printString contains return characters, the menus that assume the
resulted string fits one line fails.  This fixes one problem I
encounterd. Also, the short-cut in the WriteStream>>nextPutAll:
short-cuts the limit check in LimitedWriteStream.  This change fix it.



5910PseudoClassAllCallsOn-nk -- Ned Konz -- 29 April 2004
Fixes browsing of class references from a code-file browser.

5911renameHTTPRequest-avi -- Avi Bryant -- 30 April 2004
v2: rename it to HTTPDownloadRequest instead, and change its subclasses
to match, per Michael's wishes.
Rename HTTPRequest to BasicHTTPRequest to avoid name conflicts with


5913DeprecatedScrollMsgs-nk -- Ned Konz -- 28 April 2004
A number of scrollbar-related messages were deprecated recently, but
their senders weren't edited.
This change set changes most of the senders of these messages to call
the appropriate new methods.
Perhaps it should be required that when you deprecate a method you
should run with showDeprecationWarnings turned on for a while before


5915WeekMorphFix-nk -- Ned Konz -- 28 April 2004
This fixes the walkback upon constructing a PDAMorph in post-5892

5916RunArrayTestFix-fbs -- Frank Shearar -- 28 April 2004
When I first wrote RunArrayTest>>testScanFromANSICompatibility I used
malformed strings. A TextDoIt must end with a ';;'. This changeset adds
those ';;' strings to the test data."

5917BalloonCanvasImage-nk -- Ned Konz -- 1 May 2004
I caught this by accident; if you have a BalloonCanvas on a different
form than the Display, the old code would write to the Display instead
of your form.
Also fixes a problem seen with framed rectangles and nil fillColors.


5919KCPInitializeNewFix -- stéphane ducasse -- 20 May 2004
make sure that new: also invoke initialize.
Note that for speed gain we could in the future shortcut this call on
String, Array, ByteArray...."

5920BrowserCommentCodePane -- Robert Hirschfeld -- 12 May 2004
Fixes the following issues:
#1 Hierarchy browser opens without comment pane.
#2 (Hierarchy) Browser does not display class definition after message
category deselection.
#3 Opening a Browser from another one automatically selects the -- all
-- category if there was no method selected."

5921UtilitiesAuthorInitials-gm -- German Morales -- 15 May 2004
I can't believe that nobody noticed this before: the word 'initials' was
wrongly typed.

5922MVCCursorDownFix -- Boris Gaertner -- 29 April 2004
This change set is a proposal to fix a problem that was described in
BFAV archive ID 14350.
The error in method ParagraphEditor>>sameColumn:newLine:forward: is that
it uses 'lines size' to obtain the number of lines in a paragraph. This
works for instances of NewParagraph, but not for instances of Paragraph.
To obtain the number of lines in a paragraph, one has to write
'paragraph numberOfLines'."

5923PSPointSizeFix-nk -- Ned Konz -- 13 April 2004
Postscript output from Squeak has been in pixel units. However, we're
outputting requests for fonts in point units, making the fonts appear
3/4 as big as they should on output.
This change set fixes that problem by also requesting fonts in pixel
v3 (13 April): fixed discrepancy between (ascender+descender) and
(unitsPerEm) in TTCFont. This will rename existing point sizes but keep
glyphs the same size.
v2 (25 March): improved decoding of font names and attributes.
Factored out knowledge of font naming to TextStyle class.
"Repair bad sizes *before* installing new code (see TTCFont
	| description computedScale cached desiredScale newPointSize repaired |
	repaired _ OrderedCollection new.
	TTCFont allInstancesDo: [ :font |
		cached := (font cache copyFrom: $A asciiValue + 1 to: $z asciiValue +
			detect: [ :f | f notNil ] ifNone: [].
		cached := cached ifNil: [  font formOf: $A ] ifNotNil: [ cached value
		description _ font ttcDescription.
		desiredScale _ cached height asFloat / (description ascender -
description descender).
		computedScale _ font pixelSize asFloat / font ttcDescription
		(((computedScale / desiredScale) - 1.0 * cached height) abs < 1.0)
ifFalse: [
			newPointSize _ (font pointSize * desiredScale / computedScale)
			font pointSize: newPointSize; flushCache.
			repaired add: font.
			font derivativeFonts do: [ :df | df ifNotNil: [
				df pointSize: newPointSize; flushCache.
				repaired add: df. ]].

<This automated message brought to you by the UpdateIncorporationTool.>
<See the UpdateIncorporationTool package on SqueakMap for details.>

More information about the Squeak-dev mailing list