tim Rowledge uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-tpr.813.mcz ==================== Summary ==================== Name: Tools-tpr.813 Author: tpr Time: 10 May 2018, 2:54:14.550558 pm UUID: 01e3ab07-6692-4a92-a287-da9d4452de56 Ancestors: Tools-cmm.812 Remove the non-working (and too much work to be worth it) #browseClass send. Opening a full class browser is as useful and actually works. Rename #classHierarchy to #browseClassHiearchy because it actually makes sense to name similar methods similarly rather than with wildy unconnected names that just makes life more annoying that is strictly required. Deprecate the old version. (See 60Deprecated-tpr.19) =============== Diff against Tools-cmm.812 =============== Item was added: + ----- Method: ObjectExplorer>>browseClassHierarchy (in category 'menus - actions') ----- + browseClassHierarchy + "Create and schedule a class list browser on the receiver's hierarchy." + + self systemNavigation + spawnHierarchyForClass: self selectedClass + selector: nil + ! Item was removed: - ----- Method: ObjectExplorer>>classHierarchy (in category 'menus - actions') ----- - classHierarchy - "Create and schedule a class list browser on the receiver's hierarchy." - - self systemNavigation - spawnHierarchyForClass: self selectedClass - selector: nil - ! Item was changed: ----- Method: ObjectExplorer>>explorerKey:from:event: (in category 'menus') ----- explorerKey: aChar from: view event: event event anyModifierKeyPressed ifFalse: [^ false]. self object ifNotNil: [ aChar == $i ifTrue: [self inspectSelection. ^ true]. aChar == $I ifTrue: [self exploreSelection. ^ true]. aChar == $b ifTrue: [self browseFull. ^ true]. + aChar == $h ifTrue: [self browseClassHierarchy. ^ true]. - aChar == $h ifTrue: [self classHierarchy. ^ true]. aChar == $c ifTrue: [self copyName. ^ true]. aChar == $p ifTrue: [self browseFullProtocol. ^ true]. aChar == $N ifTrue: [self browseClassRefs. ^ true]. aChar == $v ifTrue: [self viewerForValue. ^ true]]. ^ false! Item was changed: ----- Method: ObjectExplorer>>mainTreeListMenu: (in category 'menus') ----- mainTreeListMenu: aMenu <treeListMenu> aMenu addStayUpItemSpecial. aMenu addList: #( ('inspect (i)' inspectSelection) ('explore (I)' exploreSelection) - ('method refs to this inst var' referencesToSelection) ('methods storing into this inst var' defsOfSelection) ('objects pointing to this value' objectReferencesToSelection) ('chase pointers' chasePointersForSelection) + ('explore pointers' explorePointersForSelection) - ('explore pointers' explorePointersForSelection) - ('browse full (b)' browseFull) + ('browse hierarchy (h)' browseClassHierarchy) - ('browse class' browseClass) - ('browse hierarchy (h)' classHierarchy) ('browse protocol (p)' browseFullProtocol) - ('references... (r)' browseVariableReferences) + ('assignments... (a)' browseVariableAssignments) - ('assignments... (a)' browseVariableAssignments) ('class refs (N)' browseClassRefs) - ('copy name (c)' copyName) ('copy value' copyValue) ('basic inspect' inspectBasic)). Smalltalk isMorphic ifTrue: [ aMenu addList: #( - ('viewer for this value (v)' viewerForValue))]. aMenu addList: #( - ('monitor changes' monitorSelection)). monitorList isEmptyOrNil ifFalse: [ aMenu addList: #( ('stop monitoring all' stopMonitoring))]. ^ aMenu! |
Since we're scrutinizing naming consistency, how about renaming
"browse full" to "browse package"? A Hierarchy Browser is spawned when the users selects.... (drum roll please)... 'browse hierarchy'! A Package Pane Browser is spawned when the user selects.... (drum roll please).... 'browse full"..? Wha? Not "browse package?" On Thu, May 10, 2018 at 4:54 PM, <[hidden email]> wrote: > tim Rowledge uploaded a new version of Tools to project The Trunk: > http://source.squeak.org/trunk/Tools-tpr.813.mcz > > ==================== Summary ==================== > > Name: Tools-tpr.813 > Author: tpr > Time: 10 May 2018, 2:54:14.550558 pm > UUID: 01e3ab07-6692-4a92-a287-da9d4452de56 > Ancestors: Tools-cmm.812 > > Remove the non-working (and too much work to be worth it) #browseClass send. Opening a full class browser is as useful and actually works. > Rename #classHierarchy to #browseClassHiearchy because it actually makes sense to name similar methods similarly rather than with wildy unconnected names that just makes life more annoying that is strictly required. Deprecate the old version. (See 60Deprecated-tpr.19) > > =============== Diff against Tools-cmm.812 =============== > > Item was added: > + ----- Method: ObjectExplorer>>browseClassHierarchy (in category 'menus - actions') ----- > + browseClassHierarchy > + "Create and schedule a class list browser on the receiver's hierarchy." > + > + self systemNavigation > + spawnHierarchyForClass: self selectedClass > + selector: nil > + ! > > Item was removed: > - ----- Method: ObjectExplorer>>classHierarchy (in category 'menus - actions') ----- > - classHierarchy > - "Create and schedule a class list browser on the receiver's hierarchy." > - > - self systemNavigation > - spawnHierarchyForClass: self selectedClass > - selector: nil > - ! > > Item was changed: > ----- Method: ObjectExplorer>>explorerKey:from:event: (in category 'menus') ----- > explorerKey: aChar from: view event: event > > event anyModifierKeyPressed ifFalse: [^ false]. > > self object ifNotNil: [ > aChar == $i ifTrue: [self inspectSelection. ^ true]. > aChar == $I ifTrue: [self exploreSelection. ^ true]. > > aChar == $b ifTrue: [self browseFull. ^ true]. > + aChar == $h ifTrue: [self browseClassHierarchy. ^ true]. > - aChar == $h ifTrue: [self classHierarchy. ^ true]. > aChar == $c ifTrue: [self copyName. ^ true]. > aChar == $p ifTrue: [self browseFullProtocol. ^ true]. > aChar == $N ifTrue: [self browseClassRefs. ^ true]. > aChar == $v ifTrue: [self viewerForValue. ^ true]]. > > ^ false! > > Item was changed: > ----- Method: ObjectExplorer>>mainTreeListMenu: (in category 'menus') ----- > mainTreeListMenu: aMenu > <treeListMenu> > > aMenu addStayUpItemSpecial. > > aMenu addList: #( > ('inspect (i)' inspectSelection) > ('explore (I)' exploreSelection) > - > ('method refs to this inst var' referencesToSelection) > ('methods storing into this inst var' defsOfSelection) > ('objects pointing to this value' objectReferencesToSelection) > ('chase pointers' chasePointersForSelection) > + ('explore pointers' explorePointersForSelection) > - ('explore pointers' explorePointersForSelection) > - > ('browse full (b)' browseFull) > + ('browse hierarchy (h)' browseClassHierarchy) > - ('browse class' browseClass) > - ('browse hierarchy (h)' classHierarchy) > ('browse protocol (p)' browseFullProtocol) > - > ('references... (r)' browseVariableReferences) > + ('assignments... (a)' browseVariableAssignments) > - ('assignments... (a)' browseVariableAssignments) > ('class refs (N)' browseClassRefs) > - > ('copy name (c)' copyName) > ('copy value' copyValue) > ('basic inspect' inspectBasic)). > > Smalltalk isMorphic ifTrue: [ > aMenu addList: #( > - > ('viewer for this value (v)' viewerForValue))]. > > aMenu addList: #( > - > ('monitor changes' monitorSelection)). > > monitorList isEmptyOrNil ifFalse: [ > aMenu addList: #( > ('stop monitoring all' stopMonitoring))]. > > ^ aMenu! > > |
That's a preference variable, is it not? On 5/11/18 3:07 PM, Chris Muller wrote:
Since we're scrutinizing naming consistency, how about renaming "browse full" to "browse package"? A Hierarchy Browser is spawned when the users selects.... (drum roll please)... 'browse hierarchy'! A Package Pane Browser is spawned when the user selects.... (drum roll please).... 'browse full"..? Wha? Not "browse package?" On Thu, May 10, 2018 at 4:54 PM, [hidden email] wrote:tim Rowledge uploaded a new version of Tools to project The Trunk: http://source.squeak.org/trunk/Tools-tpr.813.mcz ==================== Summary ==================== Name: Tools-tpr.813 Author: tpr Time: 10 May 2018, 2:54:14.550558 pm UUID: 01e3ab07-6692-4a92-a287-da9d4452de56 Ancestors: Tools-cmm.812 Remove the non-working (and too much work to be worth it) #browseClass send. Opening a full class browser is as useful and actually works. Rename #classHierarchy to #browseClassHiearchy because it actually makes sense to name similar methods similarly rather than with wildy unconnected names that just makes life more annoying that is strictly required. Deprecate the old version. (See 60Deprecated-tpr.19) =============== Diff against Tools-cmm.812 =============== Item was added: + ----- Method: ObjectExplorer>>browseClassHierarchy (in category 'menus - actions') ----- + browseClassHierarchy + "Create and schedule a class list browser on the receiver's hierarchy." + + self systemNavigation + spawnHierarchyForClass: self selectedClass + selector: nil + ! Item was removed: - ----- Method: ObjectExplorer>>classHierarchy (in category 'menus - actions') ----- - classHierarchy - "Create and schedule a class list browser on the receiver's hierarchy." - - self systemNavigation - spawnHierarchyForClass: self selectedClass - selector: nil - ! Item was changed: ----- Method: ObjectExplorer>>explorerKey:from:event: (in category 'menus') ----- explorerKey: aChar from: view event: event event anyModifierKeyPressed ifFalse: [^ false]. self object ifNotNil: [ aChar == $i ifTrue: [self inspectSelection. ^ true]. aChar == $I ifTrue: [self exploreSelection. ^ true]. aChar == $b ifTrue: [self browseFull. ^ true]. + aChar == $h ifTrue: [self browseClassHierarchy. ^ true]. - aChar == $h ifTrue: [self classHierarchy. ^ true]. aChar == $c ifTrue: [self copyName. ^ true]. aChar == $p ifTrue: [self browseFullProtocol. ^ true]. aChar == $N ifTrue: [self browseClassRefs. ^ true]. aChar == $v ifTrue: [self viewerForValue. ^ true]]. ^ false! Item was changed: ----- Method: ObjectExplorer>>mainTreeListMenu: (in category 'menus') ----- mainTreeListMenu: aMenu <treeListMenu> aMenu addStayUpItemSpecial. aMenu addList: #( ('inspect (i)' inspectSelection) ('explore (I)' exploreSelection) - ('method refs to this inst var' referencesToSelection) ('methods storing into this inst var' defsOfSelection) ('objects pointing to this value' objectReferencesToSelection) ('chase pointers' chasePointersForSelection) + ('explore pointers' explorePointersForSelection) - ('explore pointers' explorePointersForSelection) - ('browse full (b)' browseFull) + ('browse hierarchy (h)' browseClassHierarchy) - ('browse class' browseClass) - ('browse hierarchy (h)' classHierarchy) ('browse protocol (p)' browseFullProtocol) - ('references... (r)' browseVariableReferences) + ('assignments... (a)' browseVariableAssignments) - ('assignments... (a)' browseVariableAssignments) ('class refs (N)' browseClassRefs) - ('copy name (c)' copyName) ('copy value' copyValue) ('basic inspect' inspectBasic)). Smalltalk isMorphic ifTrue: [ aMenu addList: #( - ('viewer for this value (v)' viewerForValue))]. aMenu addList: #( - ('monitor changes' monitorSelection)). monitorList isEmptyOrNil ifFalse: [ aMenu addList: #( ('stop monitoring all' stopMonitoring))]. ^ aMenu! |
In reply to this post by Chris Muller-3
> On 11-05-2018, at 12:07 PM, Chris Muller <[hidden email]> wrote: > > Since we're scrutinizing naming consistency, how about renaming > "browse full" to "browse package"? As Bob pointed out there is actually a preference for that; personally I don't much care for the extra list and so don't set that pref. So far as I can see we use 'browse full' fairly pervasively around the image to mean 'big dog main system browser' so it's probably best to keep that. I agree it's not entirely wonderful as a phrase but it's a real bugger trying to name things well. The #classHierarchy thing annoys me a bit though - given we already had browseXXXX all over the place, how did we end up with the utterly unconnected #classHierarchy in order to browseClassHierarchy ? Madness I tell you. I did my best to change that neatly in the ObjectExplorer locale but it's there again in all sorts of StringHolder subclasses and I just don't have enough time to track it all down today. Must resist rant about insanity of StringHolder subclasses.......... I dare say somebody will point me to a browser that does this, but it occurs to me that being able to swap the view-point of a currently open browser between system-full-broswer and hierarchical (and maybe others, who knows) might be nice. And maybe the default effect of find-senders/implementors in a hierarchical focussed browser should be the 'local XXX' instead of global - not removing the global, just swapping the ordering of the two options or some similar approach. tim -- tim Rowledge; [hidden email]; http://www.rowledge.org/tim Useful random insult:- Trying out for the javelin retrieval team. |
The preference you speak of is 'Browser shows package pane'. Notice,
"Shows" (vs. "not shows"), so it's talking about a pane of the same browser -- a "Package Browser". The preference just splits off an extra pane based on the first dash "-" to delineate 'Package's. But this causes it to show the user utterly wrong and confusing information. For example, from these packages, Help-Squeak-CorePackages Help-Squeak-Project Help-Squeak-SWiki Help-Squeak-TerseGuide It makes the user think we have a package called "Help" and puts the four "Squeak-*" above into it, as if there is some sort of object delineation there, even though there isn't, not even conceptually!!! So, let's not let this confusion-causing preference of utter nonsense get in the way of improving the clarity of our IDE objects and their descriptions in the various menus. There is no 'full', it should simply be 'browse package' and, for now, the preference setting will decide whether the confusing extra pane shows or not. After that, I would support asking the list about getting rid of that Pref and "extra pane" browser altogether. On Fri, May 11, 2018 at 5:04 PM, tim Rowledge <[hidden email]> wrote: > > >> On 11-05-2018, at 12:07 PM, Chris Muller <[hidden email]> wrote: >> >> Since we're scrutinizing naming consistency, how about renaming >> "browse full" to "browse package"? > > As Bob pointed out there is actually a preference for that; personally I don't much care for the extra list and so don't set that pref. So far as I can see we use 'browse full' fairly pervasively around the image to mean 'big dog main system browser' so it's probably best to keep that. I agree it's not entirely wonderful as a phrase but it's a real bugger trying to name things well. > > The #classHierarchy thing annoys me a bit though - given we already had browseXXXX all over the place, how did we end up with the utterly unconnected #classHierarchy in order to browseClassHierarchy ? Madness I tell you. I did my best to change that neatly in the ObjectExplorer locale but it's there again in all sorts of StringHolder subclasses and I just don't have enough time to track it all down today. Must resist rant about insanity of StringHolder subclasses.......... > > I dare say somebody will point me to a browser that does this, but it occurs to me that being able to swap the view-point of a currently open browser between system-full-broswer and hierarchical (and maybe others, who knows) might be nice. And maybe the default effect of find-senders/implementors in a hierarchical focussed browser should be the 'local XXX' instead of global - not removing the global, just swapping the ordering of the two options or some similar approach. > > tim > -- > tim Rowledge; [hidden email]; http://www.rowledge.org/tim > Useful random insult:- Trying out for the javelin retrieval team. > > > |
Perhaps people new to Squeak will understand
that, but for others, a Package Browser and a System Browser are
different and invoking a System Browser by means of "browse
package" will seem confusing. Especially if the user has no
interest in the "package" concept, but wishes to see all classes
rather than some subset (which also might relate to packages). On 5/11/18 6:58 PM, Chris Muller wrote:
The preference you speak of is 'Browser shows package pane'. Notice, "Shows" (vs. "not shows"), so it's talking about a pane of the same browser -- a "Package Browser". The preference just splits off an extra pane based on the first dash "-" to delineate 'Package's. But this causes it to show the user utterly wrong and confusing information. For example, from these packages, Help-Squeak-CorePackages Help-Squeak-Project Help-Squeak-SWiki Help-Squeak-TerseGuide It makes the user think we have a package called "Help" and puts the four "Squeak-*" above into it, as if there is some sort of object delineation there, even though there isn't, not even conceptually!!! So, let's not let this confusion-causing preference of utter nonsense get in the way of improving the clarity of our IDE objects and their descriptions in the various menus. There is no 'full', it should simply be 'browse package' and, for now, the preference setting will decide whether the confusing extra pane shows or not. After that, I would support asking the list about getting rid of that Pref and "extra pane" browser altogether. On Fri, May 11, 2018 at 5:04 PM, tim Rowledge [hidden email] wrote:On 11-05-2018, at 12:07 PM, Chris Muller [hidden email] wrote: Since we're scrutinizing naming consistency, how about renaming "browse full" to "browse package"?As Bob pointed out there is actually a preference for that; personally I don't much care for the extra list and so don't set that pref. So far as I can see we use 'browse full' fairly pervasively around the image to mean 'big dog main system browser' so it's probably best to keep that. I agree it's not entirely wonderful as a phrase but it's a real bugger trying to name things well. The #classHierarchy thing annoys me a bit though - given we already had browseXXXX all over the place, how did we end up with the utterly unconnected #classHierarchy in order to browseClassHierarchy ? Madness I tell you. I did my best to change that neatly in the ObjectExplorer locale but it's there again in all sorts of StringHolder subclasses and I just don't have enough time to track it all down today. Must resist rant about insanity of StringHolder subclasses.......... I dare say somebody will point me to a browser that does this, but it occurs to me that being able to swap the view-point of a currently open browser between system-full-broswer and hierarchical (and maybe others, who knows) might be nice. And maybe the default effect of find-senders/implementors in a hierarchical focussed browser should be the 'local XXX' instead of global - not removing the global, just swapping the ordering of the two options or some similar approach. tim -- tim Rowledge; [hidden email]; http://www.rowledge.org/tim Useful random insult:- Trying out for the javelin retrieval team. |
In reply to this post by Chris Muller-3
Hi,
If I read "browse package" I would expect a browser for just one package. At least in VA Smalltalk there exists such a browser, although it is named application browser because what is called package in Squeak is called appliation there. I use it most of the time instead of a full browser. So I would not use "browse package" for opening a system browser. Bernhard > Am 12.05.2018 um 00:58 schrieb Chris Muller <[hidden email]>: > > The preference you speak of is 'Browser shows package pane'. Notice, > "Shows" (vs. "not shows"), so it's talking about a pane of the same > browser -- a "Package Browser". > > The preference just splits off an extra pane based on the first dash > "-" to delineate 'Package's. But this causes it to show the user > utterly wrong and confusing information. For example, from these > packages, > > Help-Squeak-CorePackages > Help-Squeak-Project > Help-Squeak-SWiki > Help-Squeak-TerseGuide > > It makes the user think we have a package called "Help" and puts the > four "Squeak-*" above into it, as if there is some sort of object > delineation there, even though there isn't, not even conceptually!!! > > So, let's not let this confusion-causing preference of utter nonsense > get in the way of improving the clarity of our IDE objects and their > descriptions in the various menus. There is no 'full', it should > simply be 'browse package' and, for now, the preference setting will > decide whether the confusing extra pane shows or not. > > After that, I would support asking the list about getting rid of that > Pref and "extra pane" browser altogether. > > > On Fri, May 11, 2018 at 5:04 PM, tim Rowledge <[hidden email]> wrote: >> >> >>> On 11-05-2018, at 12:07 PM, Chris Muller <[hidden email]> wrote: >>> >>> Since we're scrutinizing naming consistency, how about renaming >>> "browse full" to "browse package"? >> >> As Bob pointed out there is actually a preference for that; personally I don't much care for the extra list and so don't set that pref. So far as I can see we use 'browse full' fairly pervasively around the image to mean 'big dog main system browser' so it's probably best to keep that. I agree it's not entirely wonderful as a phrase but it's a real bugger trying to name things well. >> >> The #classHierarchy thing annoys me a bit though - given we already had browseXXXX all over the place, how did we end up with the utterly unconnected #classHierarchy in order to browseClassHierarchy ? Madness I tell you. I did my best to change that neatly in the ObjectExplorer locale but it's there again in all sorts of StringHolder subclasses and I just don't have enough time to track it all down today. Must resist rant about insanity of StringHolder subclasses.......... >> >> I dare say somebody will point me to a browser that does this, but it occurs to me that being able to swap the view-point of a currently open browser between system-full-broswer and hierarchical (and maybe others, who knows) might be nice. And maybe the default effect of find-senders/implementors in a hierarchical focussed browser should be the 'local XXX' instead of global - not removing the global, just swapping the ordering of the two options or some similar approach. >> >> tim >> -- >> tim Rowledge; [hidden email]; http://www.rowledge.org/tim >> Useful random insult:- Trying out for the javelin retrieval team. >> >> >> > |
Free forum by Nabble | Edit this page |