> One big enhancement here is Diego's new look, which will probably
> generate some feedback. :) But it's probably best to get it in (since
> it can become stale quickly), and make any further refinements
> afterward.

Did you harvest the new icons I designed? because if marcus does not  
approved them I cannot :)

> Speaking of going stale, here are the conflicts that I ran into. (With
> the NewLook enh, I think Ned's FITBFocusFix changes in these 3 methods
> may need to be reinstated, but the one method was changed too
> drastically for me to fix.

My impression is that the capture of the DNU message is not needed. Ned  
can confirm that.

>  With SARInstalled, it looks like
> SARInstaller class>>new needs to be removed from the SARInstaller
> package.)
>
>
> Checking NewLook-dgd.5.cs (#5278) for method conflicts with changesets
> after 5278FileListDnD-nk ...
> CONFLICT found for FillInTheBlankMorph class>>createAcceptButton with
> newer changeset 5351FITBFocusFix-nk
> CONFLICT found for FillInTheBlankMorph
> class>>setQuery:initialAnswer:answerExtent:acceptOnCR: with newer
> changeset 5351FITBFocusFix-nk
> CONFLICT found for FillInTheBlankMorph class>>createCancelButton with
> newer changeset 5351FITBFocusFix-nk
>
>
> Checking 5529SARInstallerFor36-21.cs (#5424) for method conflicts with
> changesets after 5430RemoveDeprecatedIn36 ...
> CONFLICT found for SARInstaller>>new with newer changeset
> 5503KCP102-5newInitPattern
>
> - Doug
>
>
> ------------------------------
>
> 5527squeakmapbase-1_08 -- Göran Krampe -- 8 Oct 2003
> Since version 1.07:
> - Small fix to SMSimpleInstaller>>download to set the FileStream to
> binary before writing. It worked fine with StandardFileStream but if  
> you
> set FileStream class>>concreteStream to CrLfFileStream then it didn't
> work... :-)
> - #forPackage:directory: now throws error if no installer is found.  
> This
> affects SMSqueakMap methods #installPackageWithId:,
> #installPackageNamed: and #upgradeOldPackagesDirectory: which now traps
> this and informs the user. (still ignoring headless operation, will  
> wait
> until SM2)
> Note that: SMLoader>>upgradeInstalledPackages doesn't catch this but  
> who
> uses that? :-). #installPackage and #downloadPackage don't either but
> they aren't available in the menu unless there is an installer found to
> deal with the selected package.
> - Uses withBlanksTrimmed in lastTransactionInLog (makes it work better
> with CrLfFileStreams I assume)
> - SMInstaller generally now uses Exceptions instead of true/false logic
> and callers now relies on Errors being signaled.
> - Only use name resolving in SMSqueakMap class>>pingServer: in a Squeak
> with version below network rewrite. The issue with proxies for older
> Squeaks is thus ignored - people using proxyserver will just have to  
> use
> a 3.6 image. Ok? :-)
> - Sightly smarter logic in synchWithLog (detecting missing directory  
> AND
> logfile).
> - Better error handling in SMSqueakMap>>installPackageNamed: and
> installPackageWithId: (typically used in scripts). They now also honors
> the SMLoader download dir Preference.
> - Better error handling generally. Patch pending for SM Package Loader
> so that install/download/upgrade/update map uses Exception handling
> gracefully. :-)
> - Improved DVS installer (that hopefully works now)
> - Various other small fixes not remembered right now...
>
> 5528SM-Loader-1_03
>
> 5529SARInstallerFor36-21 -- Ned Konz -- 10 October 2003
> 10 October (v21)
> - adds compatibility with new Monticello loader
> - adds ability to load .mcz Monticello files as well as .mcv and .mc
> "
>
> 5530ScriptableButtonFix -- Yoshiki Ohshima -- 28 September 2003
> If you bring up a ScriptableButton from the Supplies flap and press the
> cyan button to get the scriptor, it raises an error. This changeset
> trys to fix the problem."
>
> 5531TTTextStyle5-fix-dgd -- Diego Gomez Deck
> <DiegoGomezDeck at ConsultAr.com> -- 4 November 2003
> The previous version was serializing the font information using #name
> and des-serializing using #descriptionNamed: causing the lost of
> bold/italic information.
> This fix serialize using #fullName (and des-serialize using
> #descriptionFullNamed:) fixing the previous bug without breaking the
> serialization of objects serialized with the bug.
>
> 5532CollectionAtRandom -- Lukas Renggli -- 6 November 2003
> md: Just the latest version with a correct chanset preamble... ;-)
> -additional changes (and [et] [er]) from stepane.
> The message #atRandom is not thread-save when being used concurrently  
> on
> different collections or integers. As a test try to evaluate the
> following expressions in your image, the random generator provided by
> Collection gets corrupted:
> [ [ (1 to: 100) atRandom ] repeat ] forkAt: 30.
> [ [ 100 atRandom ] repeat ] forkAt: 30.
>
> This is a critical bug! As an example the superb web-framework Seaside
> uses random key-generation to identify its sessions. If other threads  
> in
> your image use #atRandom problems might arise.
> The attached change-set solves the problem by introducing a critical
> section to the Collection class providing the random generator."
>
> 5533TextMorphDropShadow-nk -- Ned Konz -- 17 October 2003
> This fixes a problem where TextMorphs with drop shadows would leave
> traces of their selections behind when they lost focus.
>
> 5534MorphRemovalNotify-nk -- Ned Konz -- 20 October 2003
> Version: 3.7a >5486
> This is a follow-on to the 3.7a update 5486MorphRemoval-ar-nk that
> updates 3.7 images to the equivalent of my v6 posting on 16 October.
> 16 October (v6) (Ned Konz):
> Added removedMorph: notifications to removeAllMorphs and
> removeAllMorphsIn:
>
> 5535SocketAndSocketStream-svp -- Stephen Pair -- 6 November 2003
> Various fixes and comment changes to Socket and SocketStream. Fixes a
> few places where there is a slight risk of infinite looping and lost
> data. Also changes FTPClient to work with these changes."
> AppRegistry
>
> 5536NewLook-dgd -- Diego Gomez Deck <DiegoGomezDeck at ConsultAr.com> --  
> 23
> June 2003
> SystemWindow Changes
> - a couple of small enhancements to system windows
> -- boxes sizes based on title font
> -- title area color more '3d'
> -- alternative boxes look
> Menus Changes
> - change the menu color to a gradient one
> -- see MenuMorph>>updateColor and senders
> -- new Preference's option called gradientMenu
> - change the menu item selected fillstyle
> -- see MenuItemMorph>>selectionFillStyle and senders
> - menui tems can have an icon
> -- see MenitItemMorph>>icon, MenitItemMorph>>icon:,
> MenitItemMorph>>hasIcon and MenitItemMorph>>drawOn:
> -- MenitItemMorph>>minWidth and MenitItemMorph>>minHeight consider the
> icon (if it's present)
> - add a set of icons to the world menu.
> -- new Preference's option named menuWithIcons (in menus category)
> -- see class MenuIcons
> -- a set of icons with an appropiate license is needed
> - new Preference's option named #menuAppearance3d. this new option
> uncouple the 3d appareance behaviour from menuColorFromWorld option.  
> now
> each option handle an aspect
> -- it's possible to have a flat menu with the color from the world, a  
> 3d
> menu with a color different from the world, etc.
> - new preference parameter named #menuLineColor.
> -- remove of unsed preferences (#menuLineLowerColor and
> #menuLineUpperColor)
> ScroolBar Changes
> - the size of the scrollbars is base on the default font height
> - new gradient option for the scrollbar's buttons
> - new set of icons for the scrollbar's buttons
> - the slider get 'pressed' when it's dragged
> - an cache is included to speed-up the time to get the images
> -- see LRUCache and references to it
> - two new Preferences options
> -- gradientScrollBars: use a gradient look
> -- alternativeButtonsInScrollBars: use an alternative set of arrows
> FillInTheBlank Changes
> - FillInTheBlankMorph uses the menu look preferences
> - defaultAnswerExtent proportional to menu font
> - red/green for [accept] and [cancel] buttons
> Preferences enable: #scrollBarsNarrow.
> Preferences
> addPreference: #alternativeWindowBoxesLook
> category: #windows
> default: true
> balloonHelp: 'If true, use an alternative look for Window''s boxes.'.
> Preferences
> addPreference: #menuWithIcons
> category: #menus
> default: true
> balloonHelp: 'If true, the menus will include some icons.'.
> Preferences
> addPreference: #gradientMenu
> category: #menus
> default: true
> balloonHelp: 'If true, the menus will have a gradient look.'.
> Preferences
> addPreference: #menuAppearance3d
> category: #menus
> default: true
> balloonHelp: 'If true, the menus will be painted with a appearance
> 3d.'.
> Preferences
> addPreference: #gradientScrollBars
> categories: #(#scrolling)
> default: true
> balloonHelp: 'If true, the scrollbars will use a gradient look.'
> projectLocal: false
> changeInformee: ScrollBar
> changeSelector: #changesInPreferences.
> Preferences
> addPreference: #alternativeButtonsInScrollBars
> categories: #(#scrolling)
> default: false
> balloonHelp: 'If true, the buttons of scrollbars will use an
> alternative set of icons.'
> projectLocal: false
> changeInformee: ScrollBar
> changeSelector: #changesInPreferences.
>
> 5537AddTargetMenuToSBM-sd -- stephane ducasse -- 31 July 2003
> Add the possibility to enter the target of a simpleButtonMorph as an
> expression"
>
> 5538BrowserInstanceCreation -- Joshua Gargus -- 29 October 2003
> Most of the methods in the 'instance creation' category of Browser did
> not answer instances of the class. Instead, they answered the class
> itself. This changeset fixes all of these methods so that they answer
> the newly created Browser."
>
> 5539PlgblListItemWrapper-ar -- Andreas Raab -- 14 October 2003
> luggableListItemWrapper makes it more easy for clients to use
> hierarchical lists. Rather than having to write a subclass of
> ListItemWrapper, a PluggableListItemWrapper can be used to provide the
> appropriate information straight from the model:
> string - an explicit string representation (contrary to the 'item'
> which contains any kind of object)
> getStringSelector - a message invoked to retrieve the sting
> representation of its item dynamically from its model (when a constant
> representation is undesirable)
> hasContentsSelector - a message invoked in the model to answer whether
> the item has any children or not.
> getContentsSelector - a message invoked in the model to retrieve the
> contents for its item.
> All callback selectors can have zero, one or two arguments with the  
> item
> and the wrapper as first and second argument.
>
> 5540CodeFont-nk -- Ned Konz -- 28 October 2003
> This makes the Browsers use a standardCodeFont.
> And it (re-) adds a menu item to change that font to the Preferences
> menu for fonts.
>
> 5541SketchMorphEditorOpen-nk -- Ned Konz -- 9 November 2003
> This adds a construction method to SketchMorph that opens an editor on  
> a
> blank SketchMorph."
>
> 5542MetaclassInitFix-nk -- Ned Konz -- 9 November 2003
> The recent changes to new/initialize cause superclasses to be
> re-initialized upon subclassing.
> This CS fixes that problem.
>
> 5543declareInstVarFIX-md -- Marcus Denker -- 12 November 2003
> The *very nice* feature introduced by [ENH] DeclareInstVar (5514) does
> not work correctly. This is just disables the menu-entry for now, will
> look into this later.
> Lukas' comment:
> | I was wondering if this ever worked?
> |
> ||yes, it works. But some cases (where the class has got super- and
> subclasses with i-vars) you ||might get a mess with the indexes of the
> i-vars. I don't know how to fix that problem as the ||Squeak parser is
> really messy.
>
> 5544BlockFullPrintOn-LC-dew -- Leandro Caniglia and Doug Way -- 6
> January 2002
> dew: V2, modifies BlockContext>>printOn: as well so that a truncated
> decompiled block appears in all inspectors, etc.
> md: Markus Gaelli digged this up from the past. I have not yet looked  
> at
> it
> at all, but I think something like this would be nice to add. ns is
> Sheldon
> Nicholl, see
>
> http://lists.squeakfoundation.org/pipermail/squeak-dev/1998-January/ 
> 0069
> 47.html
> This change set implements the BlockContext>>fullPrintOn: method. It is
> especially useful when combined with the fullPrintOnInspector GOODIE,
> that replaces the use of printString in inspectors with that of
> fullPrintString.
> Full credit to 'ns' (don't remember who's she/he?) for the
> decompileBlock: method.
>
> 5545SystDictDeprecatedFix-nk -- Ned Konz -- 12 November 2003
> This fixes several problems with the recent deprecation of various
> methods in SystemDictionary:
> * in several places, methods that had been primitives got changed so
> they would always call primitiveError.
> * why bother renaming methods as deprecated, if you move them? Instead,
> just call the moved method per suggestion.
> * having done that, there were some deprecated (and moved) methods that
> were no longer called. We can remove these.
> * some of the recently deprecated methods sent the recently deprecated
> #deprecated:explanation: message, which would cause a double  
> Deprecation
> warning.
>
>
>
>
> <This automated message brought to you by the UpdateIncorporationTool.>
> <See the UpdateIncorporationTool package on SqueakMap for details.>
>
>