[squeak-dev] The Trunk: Morphic-mt.1769.mczNicolas Cellier nicolas.cellier.aka.nice at gmail.comMon May 17 16:53:20 UTC 2021
Hi all, Virtual Machine implementation is a detail. SmallInteger are two complement, while LargeNegativeInteger are sign magnitude (with sign encoded in the class). From bit perspective (bitAnd: bitOr: bitAt: etc...), all Integer behave as two complement with an infinite sequence of 1 for negative, and infinite sequence of 0 for positive. We could express this in inspectors with a bits field and a leading ... 00000000 or ... 11111111 digit (yes digits are 1 byte long currently). Le lun. 17 mai 2021 à 15:35, Marcel Taeumel <marcel.taeumel at hpi.de> a écrit : > > The decimal representation is just one of many equivalent > representations. > > The problem being that the basic (machine) representation is not clear. I > think its 2's complement. HEX and OCT -- I suppose -- are very close to the > machine representation. > > That's not an issue for positive numbers. :-) > > Best, > Marcel > > Am 17.05.2021 14:00:36 schrieb Thiede, Christoph < > christoph.thiede at student.hpi.uni-potsdam.de>: > > > > But this would not really match the actual object layout of integers > in Squeak which can have arbitrary sizes ... > > > > So, the binary representation does not make sense, right? And neither > does the HEX or OCT. :-) > > Why do you think so? The object layout of Squeak Integer is not something > like "list of decimal digits". The decimal representation is just one of > many equivalent representations. On the other hand, any complement semantic > adds a completely new concept because it neglects the infinite size of > integers. > > Best, > Christoph > > ------------------------------ > *Von:* Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org> im > Auftrag von Taeumel, Marcel > *Gesendet:* Montag, 17. Mai 2021 13:08:47 > *An:* squeak-dev > *Betreff:* Re: [squeak-dev] The Trunk: Morphic-mt.1769.mcz > > > But this would not really match the actual object layout of integers in > Squeak which can have arbitrary sizes ... > > So, the binary representation does not make sense, right? And neither does > the HEX or OCT. :-) > > Best, > Marcel > > Am 17.05.2021 13:07:28 schrieb Thiede, Christoph < > christoph.thiede at student.hpi.uni-potsdam.de>: > > > And I would have expected a 2's complement, right? > > > Ah, now I see. :-) But this would not really match the actual object > layout of integers in Squeak which can have arbitrary sizes ... > > > In my opinion, we should just leave it as is (just maybe without the > space), I don't really like such artificial limitations. :-) > > > Best, > > Christoph > ------------------------------ > *Von:* Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org> im > Auftrag von Taeumel, Marcel > *Gesendet:* Montag, 17. Mai 2021 12:52:23 > *An:* squeak-dev > *Betreff:* Re: [squeak-dev] The Trunk: Morphic-mt.1769.mcz > > > Apart from that, I would have expected '-02' instead of '- 02'. :-) > > And I would have expected a 2's complement, right? Or even drop this > representation for negative numbers altogether? > > Best, > Marcel > > Am 17.05.2021 12:49:51 schrieb Thiede, Christoph < > christoph.thiede at student.hpi.uni-potsdam.de>: > > Nice catch! :-) > > > For anyone else wondering about this glitch: > > Before: > > > After: > > > IMO our printing protocol on Number should have a padded argument similar > to the (post-comma) decimal places. > Apart from that, I would have expected '-02' instead of '- 02'. :-) > > Best, > Christoph > ------------------------------ > *Von:* Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org> im > Auftrag von commits at source.squeak.org <commits at source.squeak.org> > *Gesendet:* Donnerstag, 6. Mai 2021 19:05:29 > *An:* squeak-dev at lists.squeakfoundation.org; > packages at lists.squeakfoundation.org > *Betreff:* [squeak-dev] The Trunk: Morphic-mt.1769.mcz > > Marcel Taeumel uploaded a new version of Morphic to project The Trunk: > http://source.squeak.org/trunk/Morphic-mt.1769.mcz > > ==================== Summary ==================== > > Name: Morphic-mt.1769 > Author: mt > Time: 6 May 2021, 7:05:24.828981 pm > UUID: 8109f744-f1f3-7048-8646-fed21a8e23d4 > Ancestors: Morphic-ct.1768 > > Fixes minor glitch in objext explorers on integers. > > This makes me wonder ... what are the expectations for negative integers > here? > > =============== Diff against Morphic-ct.1768 =============== > > Item was changed: > ----- Method: Integer>>explorerContents (in category > '*Morphic-Explorer') ----- > explorerContents > > ^#( > ('hexadecimal' 16 2) > ('octal' 8 3) > ('binary' 2 4)) collect: [ :each | | label group | > group := each third. > + label := self abs printStringBase: each second. > - label := self printStringBase: each second. > label := label padded: #left to: (label size > roundUpTo: group) with: $0. > label := String streamContents: [:s | > + self negative ifTrue: [s nextPutAll: '- > ']. > (1 to: label size by: group) > do: [:index | > 1 to: group do: [:gIndex | > s nextPut: (label > at: index + gIndex - 1)]] > separatedBy: [s space]]. > > ObjectExplorerWrapper > with: label > name: each first translated > model: self ]! > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20210517/655e37e3/attachment-0001.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: pastedImage.png Type: image/png Size: 8037 bytes Desc: not available URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20210517/655e37e3/attachment-0003.png> -------------- next part -------------- A non-text attachment was scrubbed... Name: pastedImage.png Type: image/png Size: 7983 bytes Desc: not available URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20210517/655e37e3/attachment-0004.png> -------------- next part -------------- A non-text attachment was scrubbed... Name: pastedImage.png Type: image/png Size: 7983 bytes Desc: not available URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20210517/655e37e3/attachment-0005.png>
More information about the Squeak-dev mailing list |
