I thought we said Collections was one of the triad of "Core" classes, along with Kernel and Exceptions.
That means Tools, with all of it's 30+ "Browser" classes and more must be part of that core system too?
Collections _depending_ on Tools. Something doesn't seem right here..
On Sun, Dec 29, 2013 at 4:20 PM, commits@source.squeak.org wrote:
Frank Shearar uploaded a new version of Tests to project The Trunk: http://source.squeak.org/trunk/Tests-fbs.280.mcz
==================== Summary ====================
Name: Tests-fbs.280 Author: fbs Time: 29 December 2013, 10:20:48.376 pm UUID: 6db4237c-c2f3-3440-9357-b77b8cb6f576 Ancestors: Tests-fbs.279
Collections now depends on Tools, because of the recent move of #browseWithPrettyPrint preference from Preferences to SystemBrowser. (Eventually this dependency should disappear.)
=============== Diff against Tests-fbs.279 ===============
Item was changed: ----- Method: PackageDependencyTest>>testCollections (in category 'tests') ----- testCollections self testPackage: 'Collections' dependsExactlyOn: #( Compiler Kernel Files Graphics Multilingual System 'ToolBuilder-Kernel'
Tools ).!
On 30 December 2013 21:18, Chris Muller asqueaker@gmail.com wrote:
I thought we said Collections was one of the triad of "Core" classes, along with Kernel and Exceptions.
Yep, at least until it's ripped apart. (Remember, we'd discussed moving those parts of Collection that Kernel absolutely requires (Kernel itself needs to go on a diet too, after all) into Kernel, and keep Collections containing the other generally-useful-but-not-Kernel stuff.)
That means Tools, with all of it's 30+ "Browser" classes and more must be part of that core system too?
:) No, of course not.
Collections _depending_ on Tools. Something doesn't seem right here..
Hence "(Eventually this dependency should disappear.)" But since this has stirred up conversation, let's figure out where this method _should_ go. It's hooked into compilation stuff, so maybe it should go into Compiler, or a Compiler-related package rather, along with ClassDescription >> #compile:classified:withStamp:notifying:logSource: and all the myriad other bits that use a Compiler.
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
frank
On Sun, Dec 29, 2013 at 4:20 PM, commits@source.squeak.org wrote:
Frank Shearar uploaded a new version of Tests to project The Trunk: http://source.squeak.org/trunk/Tests-fbs.280.mcz
==================== Summary ====================
Name: Tests-fbs.280 Author: fbs Time: 29 December 2013, 10:20:48.376 pm UUID: 6db4237c-c2f3-3440-9357-b77b8cb6f576 Ancestors: Tests-fbs.279
Collections now depends on Tools, because of the recent move of #browseWithPrettyPrint preference from Preferences to SystemBrowser. (Eventually this dependency should disappear.)
=============== Diff against Tests-fbs.279 ===============
Item was changed: ----- Method: PackageDependencyTest>>testCollections (in category 'tests') ----- testCollections self testPackage: 'Collections' dependsExactlyOn: #( Compiler Kernel Files Graphics Multilingual System 'ToolBuilder-Kernel'
Tools ).!
This ties into my question about #ignoreStyleIfOnlyBold in the other thread.
This new dependency Collections has on Tools (!) is due to one single method, Text>>#askIfAddStyle:req:.
But who are the senders of _that_ method? The only one is ClassDescription>>#logMethodSource:forMethodWithNode:inCategory:withStamp:notifying:.
But look at the condition on which it calls askIfAddStyle:req:. It's only if the #confirmFirstUseOfStyle is set! Something which is now outdated and no longer useful thanks to Shout.
So a better option, to me, would be to eliminate that preference and then the #askIfAddStyle:req: method entirely.
On Mon, Dec 30, 2013 at 4:47 PM, Frank Shearar frank.shearar@gmail.com wrote:
On 30 December 2013 21:18, Chris Muller asqueaker@gmail.com wrote:
I thought we said Collections was one of the triad of "Core" classes, along with Kernel and Exceptions.
Yep, at least until it's ripped apart. (Remember, we'd discussed moving those parts of Collection that Kernel absolutely requires (Kernel itself needs to go on a diet too, after all) into Kernel, and keep Collections containing the other generally-useful-but-not-Kernel stuff.)
That means Tools, with all of it's 30+ "Browser" classes and more must be part of that core system too?
:) No, of course not.
Collections _depending_ on Tools. Something doesn't seem right here..
Hence "(Eventually this dependency should disappear.)"
Ok, I wasn't sure whether this was a new permanent dependency or just a temporary one. I guess you're saying it's a temporary one.
But since this has stirred up conversation, let's figure out where this method _should_ go. It's hooked into compilation stuff, so maybe it should go into Compiler, or a Compiler-related package rather, along with ClassDescription >> #compile:classified:withStamp:notifying:logSource: and all the myriad other bits that use a Compiler.
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
frank
On Sun, Dec 29, 2013 at 4:20 PM, commits@source.squeak.org wrote:
Frank Shearar uploaded a new version of Tests to project The Trunk: http://source.squeak.org/trunk/Tests-fbs.280.mcz
==================== Summary ====================
Name: Tests-fbs.280 Author: fbs Time: 29 December 2013, 10:20:48.376 pm UUID: 6db4237c-c2f3-3440-9357-b77b8cb6f576 Ancestors: Tests-fbs.279
Collections now depends on Tools, because of the recent move of #browseWithPrettyPrint preference from Preferences to SystemBrowser. (Eventually this dependency should disappear.)
=============== Diff against Tests-fbs.279 ===============
Item was changed: ----- Method: PackageDependencyTest>>testCollections (in category 'tests') ----- testCollections self testPackage: 'Collections' dependsExactlyOn: #( Compiler Kernel Files Graphics Multilingual System 'ToolBuilder-Kernel'
Tools ).!
On 30-12-2013, at 2:47 PM, Frank Shearar frank.shearar@gmail.com wrote:
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
Move to category ‘trashcan’.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Oxymorons: Soft rock
On 31 Dec 2013, at 1:51, tim Rowledge tim@rowledge.org wrote:
On 30-12-2013, at 2:47 PM, Frank Shearar frank.shearar@gmail.com wrote:
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
Move to category ‘trashcan’.
David at least uses <curmudgeonAlert> tags :p
More seriously, pretty printing implies a lot more than syntax highlighting: it implies a canonical text representation. That's not relevant to Chris' question, but is relevant to the preference, and what it does.
frank
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Oxymorons: Soft rock
On 31 December 2013 09:40, Frank Shearar frank.shearar@gmail.com wrote:
On 31 Dec 2013, at 1:51, tim Rowledge tim@rowledge.org wrote:
On 30-12-2013, at 2:47 PM, Frank Shearar frank.shearar@gmail.com wrote:
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
Move to category ‘trashcan’.
David at least uses <curmudgeonAlert> tags :p
More seriously, pretty printing implies a lot more than syntax highlighting: it implies a canonical text representation. That's not relevant to Chris' question, but is relevant to the preference, and what it does.
Oh hey, look: #colorWhenPrettyPrinting! It looks like this was removed at some point, but without `Preferences removePreference: #colorWhenPrettyPrinting`.
frank
frank
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Oxymorons: Soft rock
On 31-12-2013, at 1:40 AM, Frank Shearar frank.shearar@gmail.com wrote:
On 31 Dec 2013, at 1:51, tim Rowledge tim@rowledge.org wrote:
On 30-12-2013, at 2:47 PM, Frank Shearar frank.shearar@gmail.com wrote:
But hey, if there's a better place than SystemBrowser for a preference called "browseWithPrettyPrint", I'm happy to move it.
Move to category ‘trashcan’.
David at least uses <curmudgeonAlert> tags :p
More seriously, pretty printing implies a lot more than syntax highlighting: it implies a canonical text representation. That's not relevant to Chris' question, but is relevant to the preference, and what it does.
That’s cos Dave is much nicer than I am.
But seriously, that preference - so far as I can work out - seems redundant in the face of Shout being part of the system. Yes, Shout doesn’t re-format code (unless there is yet another preference somewhere?) but what it does quite well (even though I really don’t like it being done to my code) is of a similar function & purpose. The option to pretty print is still there in the menu(s).
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: MET: Misread and Eat Tape
squeak-dev@lists.squeakfoundation.org