[Ann][Cuis] New Cuis release
Locked
 
|
Hi Folks,
Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . New in Cuis 2.5 - Compatibility with the Cog VM - Fix the sequence of click - wait - double click in text editors (old Squeak bug) - Fix formatting bugs in a trailing empty line in text editors (old Squeak bug) - A few updates from Squeak - Several minor fixes and enhancements Cheers, Juan Vuletich |
Re: [Ann][Cuis] New Cuis release
|
|
On Mon, 9 Aug 2010, Juan Vuletich wrote:
> Hi Folks, > > Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . Cool. > > New in Cuis 2.5 > - Compatibility with the Cog VM > - Fix the sequence of click - wait - double click in text editors (old Squeak > bug) > - Fix formatting bugs in a trailing empty line in text editors (old Squeak > bug) > - A few updates from Squeak > - Several minor fixes and enhancements A few notes: - I tried to run the tests, during that debuggers popped up with "Code simulation error". After fighting them, i got "230 run, 165 passed, 15 failed, 50 errors". - While I was fighting with the debuggers, I tried to use the "return value" feature, but it's sending #default to FillInTheBlank, which is missing, so another debugger opens. - So I decided to find the methods which send #FillInTheBlank and #default, but [ Smalltalk browseAllCallsOn: #FillInTheBlank and: #default. ] doesn't find any senders, so I used this snippet: [ | a b | a := Smalltalk allCallsOn: #FillInTheBlank. b := Smalltalk allCallsOn: #default. Smalltalk browseMessageList: (a intersection: b) name: 'Senders' autoSelect: 'FillInTheBlank default' ]. The following methods send #default to FillInTheBlank: Debugger >> #returnValue Parser >> #defineClass: - I tried to sort the result of (a intersection: b) by sending #sort to it, but OrderedCollection doesn't understand it. - And I found that FFT >> #transformForward: doesn't use the primitive. :) Levente > > Cheers, > Juan Vuletich > > |
Re: [Ann][Cuis] New Cuis release
|
|
In reply to this post by Juan Vuletich-4
On 8/9/10 2:26 PM, "Juan Vuletich" <[hidden email]> wrote: > Hi Folks, > > Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . > > New in Cuis 2.5 > - Compatibility with the Cog VM > - Fix the sequence of click - wait - double click in text editors (old > Squeak bug) > - Fix formatting bugs in a trailing empty line in text editors (old > Squeak bug) > - A few updates from Squeak > - Several minor fixes and enhancements > > Cheers, > Juan Vuletich This is very good news. I wish know... Cuis 2.5 Cog vs non Cog majorBenchmarks numbers ? Any chance to OneClickCuis with this Cuis 2.5 Cog ? Thanks Juan, keep good work Edgar |
Re: [Ann][Cuis] New Cuis release
|
|
In reply to this post by Levente Uzonyi-2
Hi Levente,
Thanks for the great feedback! A new release, 2.6 is out now. It addresses all the issues you raise. You might take a look at change set 540, for FFT. I did a bit of refactoring in FFT, and now the user is able to choose between high performance, 32 bit Float precision (the new default) and slow, 64 bit Double precision. The sample "test" methods reflect that too. Cheers, Juan Vuletich Levente Uzonyi wrote: > On Mon, 9 Aug 2010, Juan Vuletich wrote: > >> Hi Folks, >> >> Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . > > Cool. > >> >> New in Cuis 2.5 >> - Compatibility with the Cog VM >> - Fix the sequence of click - wait - double click in text editors >> (old Squeak bug) >> - Fix formatting bugs in a trailing empty line in text editors (old >> Squeak bug) >> - A few updates from Squeak >> - Several minor fixes and enhancements > > A few notes: > - I tried to run the tests, during that debuggers popped up with "Code > simulation error". After fighting them, i got "230 run, 165 passed, 15 > failed, 50 errors". > - While I was fighting with the debuggers, I tried to use the "return > value" feature, but it's sending #default to FillInTheBlank, which is > missing, so another debugger opens. > - So I decided to find the methods which send #FillInTheBlank and > #default, but [ Smalltalk browseAllCallsOn: #FillInTheBlank and: > #default. ] doesn't find any senders, so I used this snippet: > [ > | a b | > a := Smalltalk allCallsOn: #FillInTheBlank. > b := Smalltalk allCallsOn: #default. > Smalltalk > browseMessageList: (a intersection: b) > name: 'Senders' > autoSelect: 'FillInTheBlank default' ]. > The following methods send #default to FillInTheBlank: > Debugger >> #returnValue > Parser >> #defineClass: > - I tried to sort the result of (a intersection: b) by sending #sort > to it, but OrderedCollection doesn't understand it. > - And I found that FFT >> #transformForward: doesn't use the > primitive. :) > > > Levente > >> >> Cheers, >> Juan Vuletich >> >> > ------------------------------------------------------------------------ > > > No virus found in this incoming message. > Checked by AVG - www.avg.com > Version: 9.0.851 / Virus Database: 271.1.1/3062 - Release Date: 08/10/10 03:35:00 > > |
Re: [Ann][Cuis] New Cuis release
|
|
In reply to this post by Edgar De Cleene
Edgar J. De Cleene wrote:
> > On 8/9/10 2:26 PM, "Juan Vuletich" <[hidden email]> wrote: > > >> Hi Folks, >> >> Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . >> >> New in Cuis 2.5 >> - Compatibility with the Cog VM >> - Fix the sequence of click - wait - double click in text editors (old >> Squeak bug) >> - Fix formatting bugs in a trailing empty line in text editors (old >> Squeak bug) >> - A few updates from Squeak >> - Several minor fixes and enhancements >> >> Cheers, >> Juan Vuletich >> > > > This is very good news. > > I wish know... > > Cuis 2.5 Cog vs non Cog majorBenchmarks numbers ? > Any chance to OneClickCuis with this Cuis 2.5 Cog ? > > > Thanks Juan, keep good work > > Edgar > Hi Edgar Not sure what benchmarks you refer to, but feel free to experiment and post the numbers you get. WRT the OneClick distribution, anyone can help! Cheers, Juan Vuletich |
Re: [Ann][Cuis] New Cuis release
|
|
In reply to this post by Juan Vuletich-4
On Tue, 10 Aug 2010, Juan Vuletich wrote:
> Hi Levente, > > Thanks for the great feedback! A new release, 2.6 is out now. It addresses > all the issues you raise. That was a quick release. I can confirm that all tests pass, and all issues are solved. > > You might take a look at change set 540, for FFT. I did a bit of refactoring > in FFT, and now the user is able to choose between high performance, 32 bit > Float precision (the new default) and slow, 64 bit Double precision. The > sample "test" methods reflect that too. Thanks, I'll check it out. Levente > > Cheers, > Juan Vuletich > > Levente Uzonyi wrote: >> On Mon, 9 Aug 2010, Juan Vuletich wrote: >> >>> Hi Folks, >>> >>> Cuis 2.5 is available at www.jvuletich.org/Cuis/Index.html . >> >> Cool. >> >>> >>> New in Cuis 2.5 >>> - Compatibility with the Cog VM >>> - Fix the sequence of click - wait - double click in text editors (old >>> Squeak bug) >>> - Fix formatting bugs in a trailing empty line in text editors (old Squeak >>> bug) >>> - A few updates from Squeak >>> - Several minor fixes and enhancements >> >> A few notes: >> - I tried to run the tests, during that debuggers popped up with "Code >> simulation error". After fighting them, i got "230 run, 165 passed, 15 >> failed, 50 errors". >> - While I was fighting with the debuggers, I tried to use the "return >> value" feature, but it's sending #default to FillInTheBlank, which is >> missing, so another debugger opens. >> - So I decided to find the methods which send #FillInTheBlank and #default, >> but [ Smalltalk browseAllCallsOn: #FillInTheBlank and: #default. ] doesn't >> find any senders, so I used this snippet: >> [ >> | a b | >> a := Smalltalk allCallsOn: #FillInTheBlank. >> b := Smalltalk allCallsOn: #default. >> Smalltalk >> browseMessageList: (a intersection: b) >> name: 'Senders' >> autoSelect: 'FillInTheBlank default' ]. >> The following methods send #default to FillInTheBlank: >> Debugger >> #returnValue >> Parser >> #defineClass: >> - I tried to sort the result of (a intersection: b) by sending #sort to it, >> but OrderedCollection doesn't understand it. >> - And I found that FFT >> #transformForward: doesn't use the primitive. :) >> >> >> Levente >> >>> >>> Cheers, >>> Juan Vuletich >>> >>> >> ------------------------------------------------------------------------ >> >> >> No virus found in this incoming message. >> Checked by AVG - www.avg.com Version: 9.0.851 / Virus Database: >> 271.1.1/3062 - Release Date: 08/10/10 03:35:00 >> >> > > > |
«
Return to Squeak - Dev
|
1 view|%1 views
| Free forum by Nabble | Edit this page |