- Pharo-users - lists.pharo.org

Re: [Pharo-users] Zinc downloadTo: help required

by Sven Van Caekenberghe

Hi Chris, Thanks again for testing and for suggesting the fix. ZnUtils>>#streamFrom:to: without the size is a new method that was not yet covered by a unit test, I added one now. Stef, I will update the issue: Zn can be updated now in Pharo 2.0 (and maybe in 1.4 as well). Sven On 03 Aug 2012, at 18:19, Chris <cpmbailey(a)btinternet.com> wrote: > Almost! > > I had to change ZnUtils class>>#streamFrom:to: to say > > readCount := inputStream readInto: buffer startingAt: 1 count: buffer size rather than startingAt: 0. > > Otherwise ZnChunkedReadStream>>#readInto:startingAt:count: had an offset and read of 0 first time in and failed on the ByteArray replace. Seems a bit fundamental but worked fine after I changed it. > >> If Chris confirms it fixes his problem, we probably should. >> >> On 02 Aug 2012, at 19:43, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >> >>> sven should we do an update in zinc for 2.0? >>> >>> Stef >>> >>> On Aug 2, 2012, at 11:39 AM, Sven Van Caekenberghe wrote: >>> >>>> Hi Chris, >>>> >>>> On 21 Jul 2012, at 17:52, Chris <cpmbailey(a)btinternet.com> wrote: >>>> >>>>> On 20/07/2012 18:30, Sven Van Caekenberghe wrote: >>>>>> On 20 Jul 2012, at 18:22, Chris wrote: >>>>>> >>>>>>> Thanks for that. I'm actually still having a bit of trouble when the file is bigger than the chunk size. ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable which is bigger than the collection and requestedCount. >>>>>> I won't be able to do anything until the beginning of August. >>>>>> >>>>>> It would be really helpful if you could provide me with an actual test case that fails. >>>>> Okay thanks Sven, I'll see what I can do. I am just working with a third party server which is sending chunks of 1mb (which I assume is valid) and I think Zinc just needs to support chunks bigger than the 16k in the above method >>>>> >>>>> Regards, >>>>> Chris >>>> The 16K buffer used in ZnUtils>>#streamFrom:to:[size:] is independent from the chunk buffer used in ZnChunkedReadStream, BUT you did find a serious problem. Actually the code of ZnChunkedReadStream>>#readInto:startingAt:count: was embarrassingly bad and wrong although it did work in all cases thrown to it up until now. >>>> >>>> I added an extra test for the case where the buffer being used is smaller than the chunk size: >>>> >>>> ZnChunkedReadStream>>#testReadingBuffered >>>> | data chunked plain buffer readStream | >>>> data := String withAll: ($a to: $z), ($A to: $Z). >>>> chunked := String streamContents: [ :stream | >>>> ZnUtils nextPutAll: data on: stream chunked: 16 ]. >>>> readStream := ZnChunkedReadStream on: chunked readStream. >>>> buffer := String new: 11. >>>> plain := String streamContents: [ :output | | readCount | >>>> [ readStream atEnd ] whileFalse: [ >>>> readCount := readStream readInto: buffer startingAt: 1 count: buffer size. >>>> output next: readCount putAll: buffer ] ]. >>>> self assert: plain equals: data >>>> >>>> As usual, load the latest code to get the fix. I hope it now works for your particular case. >>>> >>>> Thanks again for reporting this, being persistent about it and for pointing me in the right direction ! >>>> >>>> Regards, >>>> >>>> Sven >>>> >>>> Here are the actual commits: >>>> >>>> === >>>> Name: Zinc-HTTP-SvenVanCaekenberghe.291 >>>> Author: SvenVanCaekenberghe >>>> Time: 2 August 2012, 11:26:02 am >>>> UUID: 3d8c50cd-2d7b-459f-89f3-b77a23dccfdd >>>> Ancestors: Zinc-HTTP-SvenVanCaekenberghe.290 >>>> >>>> various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem); >>>> added ZnLimitedReadStream>>#nextInto: as it is used by Fuel >>>> === >>>> Name: Zinc-Tests-SvenVanCaekenberghe.151 >>>> Author: SvenVanCaekenberghe >>>> Time: 2 August 2012, 11:27:58 am >>>> UUID: 3da15e83-c0ca-4066-a496-71d91393db01 >>>> Ancestors: Zinc-Tests-SvenVanCaekenberghe.150 >>>> >>>> added new ZnChunkedReadStreamTests>>#testReadingBuffered to validate various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem) >>>> === >>>> >>>> -- >>>> Sven Van Caekenberghe >>>> http://stfx.eu >>>> Smalltalk is the Red Pill >>>> >>>> >>>> >>> >> >> > >

Aug. 3, 2012

Re: [Pharo-users] Zinc downloadTo: help required

by Chris

Almost! I had to change ZnUtils class>>#streamFrom:to: to say readCount := inputStream readInto: buffer startingAt: 1 count: buffer size rather than startingAt: 0. Otherwise ZnChunkedReadStream>>#readInto:startingAt:count: had an offset and read of 0 first time in and failed on the ByteArray replace. Seems a bit fundamental but worked fine after I changed it. > If Chris confirms it fixes his problem, we probably should. > > On 02 Aug 2012, at 19:43, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > >> sven should we do an update in zinc for 2.0? >> >> Stef >> >> On Aug 2, 2012, at 11:39 AM, Sven Van Caekenberghe wrote: >> >>> Hi Chris, >>> >>> On 21 Jul 2012, at 17:52, Chris <cpmbailey(a)btinternet.com> wrote: >>> >>>> On 20/07/2012 18:30, Sven Van Caekenberghe wrote: >>>>> On 20 Jul 2012, at 18:22, Chris wrote: >>>>> >>>>>> Thanks for that. I'm actually still having a bit of trouble when the file is bigger than the chunk size. ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable which is bigger than the collection and requestedCount. >>>>> I won't be able to do anything until the beginning of August. >>>>> >>>>> It would be really helpful if you could provide me with an actual test case that fails. >>>> Okay thanks Sven, I'll see what I can do. I am just working with a third party server which is sending chunks of 1mb (which I assume is valid) and I think Zinc just needs to support chunks bigger than the 16k in the above method >>>> >>>> Regards, >>>> Chris >>> The 16K buffer used in ZnUtils>>#streamFrom:to:[size:] is independent from the chunk buffer used in ZnChunkedReadStream, BUT you did find a serious problem. Actually the code of ZnChunkedReadStream>>#readInto:startingAt:count: was embarrassingly bad and wrong although it did work in all cases thrown to it up until now. >>> >>> I added an extra test for the case where the buffer being used is smaller than the chunk size: >>> >>> ZnChunkedReadStream>>#testReadingBuffered >>> | data chunked plain buffer readStream | >>> data := String withAll: ($a to: $z), ($A to: $Z). >>> chunked := String streamContents: [ :stream | >>> ZnUtils nextPutAll: data on: stream chunked: 16 ]. >>> readStream := ZnChunkedReadStream on: chunked readStream. >>> buffer := String new: 11. >>> plain := String streamContents: [ :output | | readCount | >>> [ readStream atEnd ] whileFalse: [ >>> readCount := readStream readInto: buffer startingAt: 1 count: buffer size. >>> output next: readCount putAll: buffer ] ]. >>> self assert: plain equals: data >>> >>> As usual, load the latest code to get the fix. I hope it now works for your particular case. >>> >>> Thanks again for reporting this, being persistent about it and for pointing me in the right direction ! >>> >>> Regards, >>> >>> Sven >>> >>> Here are the actual commits: >>> >>> === >>> Name: Zinc-HTTP-SvenVanCaekenberghe.291 >>> Author: SvenVanCaekenberghe >>> Time: 2 August 2012, 11:26:02 am >>> UUID: 3d8c50cd-2d7b-459f-89f3-b77a23dccfdd >>> Ancestors: Zinc-HTTP-SvenVanCaekenberghe.290 >>> >>> various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem); >>> added ZnLimitedReadStream>>#nextInto: as it is used by Fuel >>> === >>> Name: Zinc-Tests-SvenVanCaekenberghe.151 >>> Author: SvenVanCaekenberghe >>> Time: 2 August 2012, 11:27:58 am >>> UUID: 3da15e83-c0ca-4066-a496-71d91393db01 >>> Ancestors: Zinc-Tests-SvenVanCaekenberghe.150 >>> >>> added new ZnChunkedReadStreamTests>>#testReadingBuffered to validate various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem) >>> === >>> >>> -- >>> Sven Van Caekenberghe >>> http://stfx.eu >>> Smalltalk is the Red Pill >>> >>> >>> >> > >

Aug. 3, 2012

Re: [Pharo-users] iterating through an ascii file line by line

by S Krish

fileStream := ... stream on the ascii file... [ fileStream atEnd ] whileFalse: [ str := fileStream upTo: Character cr . ... ] upTo: Character lf *or *upToAll: String crlf as per OS.. or make it even more generic to work across all OS ... On Tue, Jul 31, 2012 at 10:00 PM, Paul DeBruicker <pdebruic(a)gmail.com>wrote: > Sorry forgot to mention I'm using Pharo 1.3 on linux. > > >

Aug. 3, 2012

Re: [Pharo-users] iterating through an ascii file by line

by HernÃ¡n Morales Durand

Hi Sven, I didn't tried in Pharo 2.0 because is not stable. In Pharo 1.4 this works FileStream fileNamed: 'd:\Test\filter.csv' do: [ :readStream | | lineReader | lineReader := ZnLineReader on: readStream. Array streamContents: [ :output | | line | [ (line := lineReader nextLine) isEmpty ] whileFalse: [ output nextPut: line ] ] ]. but this script fails: 'd:\Test\filter.csv' asFileReference readStreamDo: [ :readStream | | lineReader | lineReader := ZnLineReader on: readStream. Array streamContents: [ :output | | line | [ (line := lineReader nextLine) isEmpty ] whileFalse: [ output nextPut: line ] ] ]. probably a problem with #asFileReference Cheers, HernÃ¡n 2012/8/2 Sven Van Caekenberghe <sven(a)beta9.be> > HernÃ¡n, > > On 01 Aug 2012, at 20:50, HernÃ¡n Morales Durand <hernan.morales(a)gmail.com> > wrote: > > > > > 2012/8/1 Sven Van Caekenberghe <sven(a)beta9.be> > > Paul, > > > > On 31 Jul 2012, at 18:25, Paul DeBruicker <pdebruic(a)gmail.com> wrote: > > > > > Is there an easy/standard way to open an ascii file and iterate > through it line by line? > > > > > > > > > I want to put several files contents into their own OrderedCollections > and then do awful things to them. The files are ~500MB each > > > > This uses code that is already in Pharo: > > > > '/tmp/foo.txt' asFileReference readStreamDo: [ :readStream | > > | lineReader | > > lineReader := ZnLineReader on: readStream. > > Array streamContents: [ :output | > > | line | > > [ (line := lineReader nextLine) isEmpty ] > > whileFalse: [ output nextPut: line ] ] ] > > > > Works with any line end convention. Check the class comment of > ZnLineReader. > > > > > > No, it doesn't work in latest Pharo 1.4 (14557) > > MessageNotUnderstood: FileSystemReadStream>>isBinary > > > > Cheers, > > > > HernÃ¡n > > I did try the code snippet on Pharo 2.0 before posting it, it worked. > > I just tried the following on Pharo 1.4, and it works as well. > > FileStream fileNamed: '/tmp/foo.txt' do: [ :readStream | > | lineReader | > lineReader := ZnLineReader on: readStream. > Array streamContents: [ :output | > | line | > [ (line := lineReader nextLine) isEmpty ] > whileFalse: [ output nextPut: line ] ] ] > > Maybe you tried using some FileSystem port on 1.4 that is not as complete > as the code that is by default in 2.0 ? > But ZnLineReader is independent of that, it has been part of Zn since > years, in Pharo 1.3 and up. > > Sven >

Aug. 3, 2012

Re: [Pharo-users] Zinc downloadTo: help required

by StÃ©phane Ducasse

oki http://code.google.com/p/pharo/issues/detail?id=6510 Stef

Aug. 2, 2012

Re: [Pharo-users] Zinc downloadTo: help required

by Sven Van Caekenberghe

If Chris confirms it fixes his problem, we probably should. On 02 Aug 2012, at 19:43, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > sven should we do an update in zinc for 2.0? > > Stef > > On Aug 2, 2012, at 11:39 AM, Sven Van Caekenberghe wrote: > >> Hi Chris, >> >> On 21 Jul 2012, at 17:52, Chris <cpmbailey(a)btinternet.com> wrote: >> >>> On 20/07/2012 18:30, Sven Van Caekenberghe wrote: >>>> On 20 Jul 2012, at 18:22, Chris wrote: >>>> >>>>> Thanks for that. I'm actually still having a bit of trouble when the file is bigger than the chunk size. ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable which is bigger than the collection and requestedCount. >>>> I won't be able to do anything until the beginning of August. >>>> >>>> It would be really helpful if you could provide me with an actual test case that fails. >>> >>> Okay thanks Sven, I'll see what I can do. I am just working with a third party server which is sending chunks of 1mb (which I assume is valid) and I think Zinc just needs to support chunks bigger than the 16k in the above method >>> >>> Regards, >>> Chris >> >> The 16K buffer used in ZnUtils>>#streamFrom:to:[size:] is independent from the chunk buffer used in ZnChunkedReadStream, BUT you did find a serious problem. Actually the code of ZnChunkedReadStream>>#readInto:startingAt:count: was embarrassingly bad and wrong although it did work in all cases thrown to it up until now. >> >> I added an extra test for the case where the buffer being used is smaller than the chunk size: >> >> ZnChunkedReadStream>>#testReadingBuffered >> | data chunked plain buffer readStream | >> data := String withAll: ($a to: $z), ($A to: $Z). >> chunked := String streamContents: [ :stream | >> ZnUtils nextPutAll: data on: stream chunked: 16 ]. >> readStream := ZnChunkedReadStream on: chunked readStream. >> buffer := String new: 11. >> plain := String streamContents: [ :output | | readCount | >> [ readStream atEnd ] whileFalse: [ >> readCount := readStream readInto: buffer startingAt: 1 count: buffer size. >> output next: readCount putAll: buffer ] ]. >> self assert: plain equals: data >> >> As usual, load the latest code to get the fix. I hope it now works for your particular case. >> >> Thanks again for reporting this, being persistent about it and for pointing me in the right direction ! >> >> Regards, >> >> Sven >> >> Here are the actual commits: >> >> === >> Name: Zinc-HTTP-SvenVanCaekenberghe.291 >> Author: SvenVanCaekenberghe >> Time: 2 August 2012, 11:26:02 am >> UUID: 3d8c50cd-2d7b-459f-89f3-b77a23dccfdd >> Ancestors: Zinc-HTTP-SvenVanCaekenberghe.290 >> >> various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem); >> added ZnLimitedReadStream>>#nextInto: as it is used by Fuel >> === >> Name: Zinc-Tests-SvenVanCaekenberghe.151 >> Author: SvenVanCaekenberghe >> Time: 2 August 2012, 11:27:58 am >> UUID: 3da15e83-c0ca-4066-a496-71d91393db01 >> Ancestors: Zinc-Tests-SvenVanCaekenberghe.150 >> >> added new ZnChunkedReadStreamTests>>#testReadingBuffered to validate various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem) >> === >> >> -- >> Sven Van Caekenberghe >> http://stfx.eu >> Smalltalk is the Red Pill >> >> >> > >

Aug. 2, 2012

Re: [Pharo-users] Zinc downloadTo: help required

by StÃ©phane Ducasse

sven should we do an update in zinc for 2.0? Stef On Aug 2, 2012, at 11:39 AM, Sven Van Caekenberghe wrote: > Hi Chris, > > On 21 Jul 2012, at 17:52, Chris <cpmbailey(a)btinternet.com> wrote: > >> On 20/07/2012 18:30, Sven Van Caekenberghe wrote: >>> On 20 Jul 2012, at 18:22, Chris wrote: >>> >>>> Thanks for that. I'm actually still having a bit of trouble when the file is bigger than the chunk size. ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable which is bigger than the collection and requestedCount. >>> I won't be able to do anything until the beginning of August. >>> >>> It would be really helpful if you could provide me with an actual test case that fails. >> >> Okay thanks Sven, I'll see what I can do. I am just working with a third party server which is sending chunks of 1mb (which I assume is valid) and I think Zinc just needs to support chunks bigger than the 16k in the above method >> >> Regards, >> Chris > > The 16K buffer used in ZnUtils>>#streamFrom:to:[size:] is independent from the chunk buffer used in ZnChunkedReadStream, BUT you did find a serious problem. Actually the code of ZnChunkedReadStream>>#readInto:startingAt:count: was embarrassingly bad and wrong although it did work in all cases thrown to it up until now. > > I added an extra test for the case where the buffer being used is smaller than the chunk size: > > ZnChunkedReadStream>>#testReadingBuffered > | data chunked plain buffer readStream | > data := String withAll: ($a to: $z), ($A to: $Z). > chunked := String streamContents: [ :stream | > ZnUtils nextPutAll: data on: stream chunked: 16 ]. > readStream := ZnChunkedReadStream on: chunked readStream. > buffer := String new: 11. > plain := String streamContents: [ :output | | readCount | > [ readStream atEnd ] whileFalse: [ > readCount := readStream readInto: buffer startingAt: 1 count: buffer size. > output next: readCount putAll: buffer ] ]. > self assert: plain equals: data > > As usual, load the latest code to get the fix. I hope it now works for your particular case. > > Thanks again for reporting this, being persistent about it and for pointing me in the right direction ! > > Regards, > > Sven > > Here are the actual commits: > > === > Name: Zinc-HTTP-SvenVanCaekenberghe.291 > Author: SvenVanCaekenberghe > Time: 2 August 2012, 11:26:02 am > UUID: 3d8c50cd-2d7b-459f-89f3-b77a23dccfdd > Ancestors: Zinc-HTTP-SvenVanCaekenberghe.290 > > various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem); > added ZnLimitedReadStream>>#nextInto: as it is used by Fuel > === > Name: Zinc-Tests-SvenVanCaekenberghe.151 > Author: SvenVanCaekenberghe > Time: 2 August 2012, 11:27:58 am > UUID: 3da15e83-c0ca-4066-a496-71d91393db01 > Ancestors: Zinc-Tests-SvenVanCaekenberghe.150 > > added new ZnChunkedReadStreamTests>>#testReadingBuffered to validate various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem) > === > > -- > Sven Van Caekenberghe > http://stfx.eu > Smalltalk is the Red Pill > > >

Aug. 2, 2012

Re: [Pharo-users] Zinc downloadTo: help required

by Sven Van Caekenberghe

Hi Chris, On 21 Jul 2012, at 17:52, Chris <cpmbailey(a)btinternet.com> wrote: > On 20/07/2012 18:30, Sven Van Caekenberghe wrote: >> On 20 Jul 2012, at 18:22, Chris wrote: >> >>> Thanks for that. I'm actually still having a bit of trouble when the file is bigger than the chunk size. ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable which is bigger than the collection and requestedCount. >> I won't be able to do anything until the beginning of August. >> >> It would be really helpful if you could provide me with an actual test case that fails. > > Okay thanks Sven, I'll see what I can do. I am just working with a third party server which is sending chunks of 1mb (which I assume is valid) and I think Zinc just needs to support chunks bigger than the 16k in the above method > > Regards, > Chris The 16K buffer used in ZnUtils>>#streamFrom:to:[size:] is independent from the chunk buffer used in ZnChunkedReadStream, BUT you did find a serious problem. Actually the code of ZnChunkedReadStream>>#readInto:startingAt:count: was embarrassingly bad and wrong although it did work in all cases thrown to it up until now. I added an extra test for the case where the buffer being used is smaller than the chunk size: ZnChunkedReadStream>>#testReadingBuffered | data chunked plain buffer readStream | data := String withAll: ($a to: $z), ($A to: $Z). chunked := String streamContents: [ :stream | ZnUtils nextPutAll: data on: stream chunked: 16 ]. readStream := ZnChunkedReadStream on: chunked readStream. buffer := String new: 11. plain := String streamContents: [ :output | | readCount | [ readStream atEnd ] whileFalse: [ readCount := readStream readInto: buffer startingAt: 1 count: buffer size. output next: readCount putAll: buffer ] ]. self assert: plain equals: data As usual, load the latest code to get the fix. I hope it now works for your particular case. Thanks again for reporting this, being persistent about it and for pointing me in the right direction ! Regards, Sven Here are the actual commits: === Name: Zinc-HTTP-SvenVanCaekenberghe.291 Author: SvenVanCaekenberghe Time: 2 August 2012, 11:26:02 am UUID: 3d8c50cd-2d7b-459f-89f3-b77a23dccfdd Ancestors: Zinc-HTTP-SvenVanCaekenberghe.290 various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem); added ZnLimitedReadStream>>#nextInto: as it is used by Fuel === Name: Zinc-Tests-SvenVanCaekenberghe.151 Author: SvenVanCaekenberghe Time: 2 August 2012, 11:27:58 am UUID: 3da15e83-c0ca-4066-a496-71d91393db01 Ancestors: Zinc-Tests-SvenVanCaekenberghe.150 added new ZnChunkedReadStreamTests>>#testReadingBuffered to validate various fixes to ZnChunkedReadStream>>#readInto:startingAt:count: (thx Chris Bailey for reporting the problem) === -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Aug. 2, 2012

Re: [Pharo-users] iterating through an ascii file by line

by Sven Van Caekenberghe

HernÃ¡n, On 01 Aug 2012, at 20:50, HernÃ¡n Morales Durand <hernan.morales(a)gmail.com> wrote: > > 2012/8/1 Sven Van Caekenberghe <sven(a)beta9.be> > Paul, > > On 31 Jul 2012, at 18:25, Paul DeBruicker <pdebruic(a)gmail.com> wrote: > > > Is there an easy/standard way to open an ascii file and iterate through it line by line? > > > > > > I want to put several files contents into their own OrderedCollections and then do awful things to them. The files are ~500MB each > > This uses code that is already in Pharo: > > '/tmp/foo.txt' asFileReference readStreamDo: [ :readStream | > | lineReader | > lineReader := ZnLineReader on: readStream. > Array streamContents: [ :output | > | line | > [ (line := lineReader nextLine) isEmpty ] > whileFalse: [ output nextPut: line ] ] ] > > Works with any line end convention. Check the class comment of ZnLineReader. > > > No, it doesn't work in latest Pharo 1.4 (14557) > MessageNotUnderstood: FileSystemReadStream>>isBinary > > Cheers, > > HernÃ¡n I did try the code snippet on Pharo 2.0 before posting it, it worked. I just tried the following on Pharo 1.4, and it works as well. FileStream fileNamed: '/tmp/foo.txt' do: [ :readStream | | lineReader | lineReader := ZnLineReader on: readStream. Array streamContents: [ :output | | line | [ (line := lineReader nextLine) isEmpty ] whileFalse: [ output nextPut: line ] ] ] Maybe you tried using some FileSystem port on 1.4 that is not as complete as the code that is by default in 2.0 ? But ZnLineReader is independent of that, it has been part of Zn since years, in Pharo 1.3 and up. Sven

Aug. 2, 2012

Re: [Pharo-users] How do you guys navigate while editing code?

by Thomas Worthington

On Wed, 01 Aug 2012 08:31:53 +0100, Patrik Sundberg <patrik.sundberg(a)gmail.com> wrote: > On Aug 1, 2012 2:47 AM, "Sean P. DeNigris" <sean(a)clipperadams.com> wrote: >> >> >> Patrik Sundberg wrote >> > >> > I'm having vim withdrawal symptoms :) >> > >> >> Ha ha. I know the feeling. Now that we have Keymapping, I finally >> started >> vim bindings for Pharo tools. The project is at >> http://ss3.gemstone.com/ss/VimPharo.html ( until I move it to github ;-) > ). >> I'm very close to having most of the basic navigation shortcuts in >> normal >> mode implemented. >> >> In the long term, I think what would be the most powerful for Pharo is >> to >> reimagine what the vim shortcuts would be if they operated on live >> objects >> (AST) instead of text snippets... >> >> HTH, Sean > > indeed, I've noted the project and I'm following it :) > > all the prerequisites to do something very powerful is there. I can even > imagine something as controversial as an input/edit/refactoring service > run > by image and connected to (even remotely) by vim/emacs/whatever same way > that slime works for lisp and emacs. While this sounds nice I do wonder if there is some simpler solution to common issues. When I create a top level category, for example, the logical thing would be for focus to go to the Class template. Similarly, when I click on the class category and the system presents the method template it seems fairly logical to put the focus on the text there. As it is, unless I've missed something, when a method template is automatically created in the source panel it is highlighted as if for editing but focus has to be switched to start editing, which removes the highlighting and so I have to re-create it with ctl-a! All this involves a lot of moving from keyboard to mouse and back again. Is there some shortcut for simply changing focus from panel to panel in the browser? Thomas Worthington

Aug. 1, 2012