Squeak-dev September 2020

squeak-dev@lists.squeakfoundation.org

37 participants
267 discussions

On the rejection of Promises due to errors
by Jakob Reschke 26 Aug '23

26 Aug '23

Hi all, Tony has recently fixed our Promise implementation to make it more Promises/A+ compliant, thank you! In the discussion that lead to this fix [1], I already pointed out a difference between exceptions in Smalltalk and in JavaScript, where the Promises/A+ specification originates: Smalltalk exceptions can be resumed, while JavaScript exceptions cannot be resumed and always unroll the stack. The spec [2] says that if the onFulfilled or onRejected callback of a #then call on a promise throws an exception, then the promise returned by the #then call shall be rejected with the exception thrown. Our current Promise implementation matches this for the blocks supplied to #then:, #ifRejected: or #then:ifRejected, by catching all Errors in the blocks and rejecting the promise. But this does not allow a Squeak user to deal with exceptions in a debugger if they are signalled in the callbacks, because they are caught. The same also applies to #future promises. The latter are not really covered by the Promises/A+ spec (because it does not force the resolution or rejection of a promise that is not the result of a #then, and there is no #future in JavaScript), but futures exhibit the same problem of not being resumable in the debugger. Example: promise := 1 future / 0. "<= inspect it => promise is rejected, regardless of your actions in the debugger" Number compile: 'methodWithTypo ^ self asstring'. promise := 1 future methodWithTypo. "<= inspect it => promise is rejected, no chance to fix the misspelling of asString in the debugger and proceed" I could imagine instead letting all exceptions pass during the future or callback block evaluation, and only reject the promise if the evaluation is eventually curtailed due to the exception (be it an Error or not, think of Warning or ModificationForbidden). Example expectations: promise := 1 future / 0. "<= inspect it, press Proceed in the debugger, => promise is resolved" promise := 1 future / 0. "<= inspect it, press Abandon in the debugger, => promise is rejected" promise := 1 future methodWithTypo. "<= inspect it, fix the typo of asString in the debugger, proceed, => promise is resolved with '1'" It could be done by fulfilling a Promise about aBlock similar to this: [ self resolveWith: aBlock value ] on: Exception do: [ :ex | | resumed | resumed := false. [ | result | result := ex outer. resumed := true. ex resume: result] ifCurtailed: [resumed ifFalse: [self future rejectWith: ex]]] (Find the current implementations here: Promise>>#fulfillWith:passErrors: and Promise>>#then:ifRejected:) Note that the #outer send would only trigger handlers in the Project/World loop, or the defaultAction of the exception. The #future in front of #rejectWith: is there to avoid curtailing the unwind block context of ifCurtailed: itself if there are further errors in the rejection callbacks of the promise. The behavior of non-local exits from unwind contexts is undefined in the Smalltalk ANSI standard (just like resume: or return: in a defaultAction, or not sending resume: or return: in an on:do: exception handler at all -- VA Smalltalk interprets that as resume, while Squeak does return, for example). This implementation would also allow all deferred Notifications to pass and not reject the promise. That is because true notifications just resume silently if they are not handled. promise := [Notification signal: 'hi there'. 42] future value. "<= inspect it => Expected: resolved with 42. Actual (today): it is needlessly rejected with Notification 'hi there'" Pressing Proceed in the debugger on officially non-resumable errors (which is possible) would also not reject the promise. But further errors/debuggers are likely to appear, of which one may eventually be used to abort the execution. If the execution finishes after repeatedly pressing Proceed, then fine, resolve the promise with whatever the outcome was. promise := [self error: 'Fatal error'. 42] future value. "<= inspect it, proceed after the so-called fatal error, => Expected: resolved with 42. Actual: there is no debugger, the promise is immediately rejected." promise := [1 / 0 + 3] future value. "<= Cannot be resumed/proceeded because if ZeroDivide is resumed, it will return the exception, and ZeroDivide does not understand +, which cannot be resumed without changing the code. So you'd have to curtail the block execution => Expected: rejected with ZeroDivide or MessageNotUnderstood (depending on when you press Abandon or recompile the DoIt)." promise := [1 / 0 + 3] future value. "... or instead of changing the code or aborting, you could choose 'return entered value' in one of the debuggers, and thereby complete the evaluation of the block => Expected: resolved with whatever you entered to return in the debugger" Promises with whenRejected:/ifRejected: callbacks would no longer swallow errors, and would only be rejected when the user aborts in the debuggers, or if the future execution catches errors by itself and converts them to rejected promises, so the future promise will also be rejected. This could pose a compatibility problem for existing code. promise := (1 future / 0) then: [:result | result + 3] ifRejected: [:reason | #cancelled]. "<= inspect it => Actual: resolved with #cancelled immediately. Expected with my proposed changes: it would first show the ZeroDivide debugger, which you can abandon to resolve with #cancelled, or proceed to a MessageNotUnderstood +. If you abandon the MNU, the promise would be rejected with the MNU, not #cancelled, in accordance with the Promises/A+ spec." How to get back a catch-all->reject-immediately future under these circumstances: promise := [[1 / 0] on: Error do: [:e | e return: (Promise new rejectWith: e)]] future value. promise := [1 future + 1 then: [:n | [n / 0] on: Error do: [:e | e return: (Promise new rejectWith: e)]] future value. We could also introduce a convenience constructor for immediately-rejected promises like in JavaScript: Promise rejected: e. Or a convenience exception handler: [...] rejectOn: Error. Or [...] rejectIfCurtailed (the fullfill/then methods would probably use this as well). What do you think? As Tom Beckmann has already suggested in the last thread on the topic [1], I could also use a custom class of Promise to get just the behavior I want. But then I cannot solve it for the use of #future. At least not without patching something about the compiler in my package preamble... ;-) [1] http://lists.squeakfoundation.org/pipermail/squeak-dev/2020-April/208546.ht… [2] https://promisesaplus.com/ Kind regards, Jakob

6 17

WeakSet but no WeakIdentitySet?
by Igor Stasenko 26 Aug '23

26 Aug '23

Hello, i need to keep a bunch of CompiledMethods in WeakSet, but i am a bit afraid, that it wont work as i want to, because of CompiledMethod>>= implementation, because a code, which i currently writing doesn't changing a method's bytecode, but changing its trailer instead, and so, this could lead to undesirable effects.. If i could have a WeakIdentitySet... Meanwhile, i think i have nothing else, but just use a degenerate WeakIdentityKeyDictionary , with methods in keys and nils in values. :( -- Best regards, Igor Stasenko AKA sig.

4 3

veryDeepCopy, ImageSegments, and actionMaps
by Ned Konz 17 Aug '23

17 Aug '23

Any Object can have dependents whether or not it has an instance variable to keep track of them (these are kept in the DependentsFields class variable of Object). The only class that keeps its own dependents is Model. Any Object can have an actionMap. These are supplied by EventManager class by default, and are stored in the ActionMaps class variable of EventManager. The only class that keeps its own action maps is Morph. When we veryDeepCopy an object we also copy the global dependents (see DeepCopier>>fixDependents). However, we don't copy the actionMap in those cases where it's held by the EventManager class. When we file out an object that uses DependentsFields out into an ImageSegment, we also file out the dependents (see ImageSegment>>dependentsSave:) if it's in a project. We do this by adding a project parameter called #GlobalDependentsInProject. However, we don't file out the actionMap in those cases where it's held by the EventManager class. My question is this: should we fix these behaviors (copying and serialization) to include these "global action maps" just like we already include the "global dependents"? Or should we discourage the use of the "global action maps"? -- Ned Konz http://bike-nomad.com GPG key ID: BEEA7EFE

3 2

The Trunk: Collections-pre.857.mcz
by commits＠source.squeak.org 05 Jun '23

05 Jun '23

Patrick Rein uploaded a new version of Collections to project The Trunk: http://source.squeak.org/trunk/Collections-pre.857.mcz ==================== Summary ==================== Name: Collections-pre.857 Author: pre Time: 4 October 2019, 11:04:30.363303 am UUID: 5ef00b65-3884-c445-b276-0cc01f0b10a1 Ancestors: Collections-pre.856 Adds startOfHeader to Character, adds empty abstract implementations of scanFrom:, writeScanOn: to TextAttribute to allow for Texts which include TextAttributes which do not implement serialization to still be serialized, adds a comment to these methods. =============== Diff against Collections-pre.856 =============== Item was added: + ----- Method: Character class>>startOfHeader (in category 'accessing untypeable characters') ----- + startOfHeader + + ^ self value: 1 ! Item was added: + ----- Method: TextAttribute class>>scanFrom: (in category 'fileIn/Out') ----- + scanFrom: strm + "Read the text attribute properties from the stream. When this method has + been called the concrete TextAttribute class has already been selected via + scanCharacter. (see TextAttribute class>>#newFrom:). + For writing the format see TextAttribute>>#writeScanOn:"! Item was added: + ----- Method: TextAttribute>>writeScanOn: (in category 'fileIn/fileOut') ----- + writeScanOn: strm + "Implement this method for a text attribute to define how it it should be written + to a serialized form of a text object. The form should correspond to the source + file format, i.e. use a scan character to denote its subclass. + As TextAttributes are stored in RunArrays, this method is mostly called from RunArray>>#write scan. + For reading the written information see TextAttribute class>>#scanFrom:" + + "Do nothing because of abstract class"!

8 17

The Inbox: Morphic-ct.1641.mcz
by commits＠source.squeak.org 20 May '23

20 May '23

Christoph Thiede uploaded a new version of Morphic to project The Inbox: http://source.squeak.org/inbox/Morphic-ct.1641.mcz ==================== Summary ==================== Name: Morphic-ct.1641 Author: ct Time: 7 June 2020, 2:45:17.536284 pm UUID: b90e23e1-ce70-3d49-a8e4-f09af148ef97 Ancestors: Morphic-ct.1638 Refine #isHighFrequentEvent implementation for further event types. Fixes one possible cause of the "color depth = 0" bug. See http://forum.world.st/Image-not-startable-after-save-td5117084.html. Depends indeed on Morphic-ct.1638. =============== Diff against Morphic-ct.1638 =============== Item was added: + ----- Method: DropEvent>>isHighFrequentEvent (in category 'testing') ----- + isHighFrequentEvent + + ^ false! Item was added: + ----- Method: WindowEvent>>isHighFrequentEvent (in category 'testing') ----- + isHighFrequentEvent + + ^ (#(windowClose) includes: self type) not! Item was changed: ----- Method: WorldState>>doSafely:onErrorThat:setErrorFlag:ifFatal:afterErrorDo: (in category 'update cycle') ----- doSafely: aBlock onErrorThat: errorPredicate setErrorFlag: errorFlag ifFatal: fatalErrorBlock afterErrorDo: postErrorBlock "Evaluate aBlock and keep track of errors during morph invocations." | finished classesWithErrors | finished := false. classesWithErrors := IdentitySet new. [finished] whileFalse: [ [aBlock value. finished := true] on: Error, Halt, Warning do: [:ex | | err rcvr errCtxt errMorph | (errorPredicate cull: ex) ifFalse: [ex pass]. err := ex description. rcvr := ex receiver. errCtxt := thisContext. [ errCtxt := errCtxt sender. "Search the sender chain to find the morph causing the problem" [errCtxt notNil and: [(errCtxt receiver isMorph) not]] whileTrue: [errCtxt := errCtxt sender]. "If we're at the root of the context chain then we have a fatal problem" errCtxt ifNil: [^ fatalErrorBlock cull: err]. + errMorph := errCtxt receiver + ] doWhileTrue: [ - errMorph := errCtxt receiver. "If the morph causing the problem has already the error flag set, then search for the next morph above in the caller chain." errMorph hasProperty: errorFlag + ]. - ] whileTrue. errMorph setProperty: errorFlag toValue: true; changed. "Catch all errors, one for each receiver class." (classesWithErrors includes: rcvr class) ifFalse: [ classesWithErrors add: rcvr class. ToolSet debugException: ex]. postErrorBlock cull: err. ]].! Item was changed: ----- Method: WorldState>>processEventsSafely: (in category 'update cycle') ----- processEventsSafely: aHandMorph ^ self doSafely: [aHandMorph processEvents] + onErrorThat: [:error | self currentEvent ifNil: [true] ifNotNil: [:evt | evt isHighFrequentEvent]] - onErrorThat: [:error | ActiveEvent isNil or: [ActiveEvent isHighFrequentEvent]] setErrorFlag: #errorOnEvent ifFatal: [:error | Project current fatalEventHandlingError: error] afterErrorDo: []!

5 6

FormInspector, or also: Text>>#= and its consequences
by Thiede, Christoph 12 Jan '23

12 Jan '23

Hi all, is there any old thread about the design discussion of how Text>>#= works? (It does not consider attributes for quality.) Has this decision ever been questioned? Naively and without an overview of any existing components that could rely on this implementation, I would like to question it. Why should 'foo' asText allBold be equal to 'foo' asText addAttribute: TextURL new? With the same logic, we could also say that two dictionaries are equal iff they have got the same keys ... There is even a concrete client in the Trunk suffering from this design decision: Marcel's new FormInspector (and analogously, MorphInspector). It uses TextFontReference with a FormSetFont to display a screenshot right in the inspector pane. Unfortunately, the pane is never updated automatically because even if the screenshot changes, the text morph thinks the old text would equal the new one. I'd like to fix that without hacking any workaround into the inspectors. Even though this inspector implementation is a bit unusual, in my opinion, it shows that the current Text >> #= implementation might not be a perfect solution. I'm looking forward to your opinions. Best, Christoph

8 11

Should I put my unofficial V3 versions of Squeak 5.2 and 5.3 on files.squeak.org?
by David T. Lewis 28 Dec '22

28 Dec '22

A couple of years ago, I put a copy of a Squeak 5.2 image in V3 object memory format on the squeakvm.org server, and I linked to it from the project page at http://www.squeaksource.com/TrunkUpdateStreamV3. That link has gone dead, and I think I recall someone noticing it a recently. I still have the files, should I put them somewhere on files.squeak.org? I was thinking of putting them under the "various_images" folder: files.squeak.org/various_images/squeak_V3_images/ I also have a V3 image that is up to date with the Squeak 5.3 release, so I could put that on line as well. These images will run under an up-to-date interpreter VM or a Cog VM. They may be of interest to people doing VM work, benchmarking, or as a reference to understand the image-side differences between Spur and V3. Dave

4 5

The Trunk: Regex-Help-pre.1.mcz
by commits＠source.squeak.org 12 Oct '22

12 Oct '22

Patrick Rein uploaded a new version of Regex-Help to project The Trunk: http://source.squeak.org/trunk/Regex-Help-pre.1.mcz ==================== Summary ==================== Name: Regex-Help-pre.1 Author: pre Time: 6 July 2018, 5:14:56.533269 pm UUID: 476b203d-1709-a54e-9a96-f0dfc3a93dfd Ancestors: Converts the regex documentation from class methods to a full help topic. ==================== Snapshot ==================== SystemOrganization addCategory: #'Regex-Help'! CustomHelp subclass: #RegexHelp instanceVariableNames: '' classVariableNames: '' poolDictionaries: '' category: 'Regex-Help'! ----- Method: RegexHelp class>>bookName (in category 'as yet unclassified') ----- bookName ^ 'Regex'! ----- Method: RegexHelp class>>changelog (in category 'as yet unclassified') ----- changelog "This method was automatically generated. Edit it using:" "RegexHelp edit: #changelog" ^(HelpTopic title: 'Changelog' contents: 'VERSION 1.3.1 (September 2008) 1. Updated documentation of character classes, making clear the problems of locale - an area for future improvement VERSION 1.3 (September 2008) 1. \w now matches underscore as well as alphanumerics, in line with most other regex libraries (and our documentation!!!!). 2. \W rejects underscore as well as alphanumerics 3. added tests for this at end of testSuite 4. updated documentation and added note to old incorrect comments in version 1.1 below VERSION 1.2.3 (November 2007) 1. Regexs with ^ or $ applied to copy empty strings caused infinite loops, e.g. ('''' copyWithRegex: ''^.*$'' matchesReplacedWith: ''foo''). Applied a similar correction to that from version 1.1c, to #copyStream:to:(replacingMatchesWith:|translatingMatchesUsing:). 2. Extended RxParser testing to run each test for #copy:translatingMatchesUsing: as well as #search:. 3. Corrected #testSuite test that a dot does not match a null, which was passing by luck with Smalltalk code in a literal array. 4. Added test to end of test suite for fix 1 above. VERSION 1.2.2 (November 2006) There was no way to specify a backslash in a character set. Now [\\] is accepted. VERSION 1.2.1 (August 2006) 1. Support for returning all ranges (startIndex to: stopIndex) matching a regex - #allRangesOfRegexMatches:, #matchingRangesIn: 2. Added hint to usage documentation on how to get more information about matches when enumerating 3. Syntax description of dot corrected: matches anything but NUL since 1.1a VERSION 1.2 (May 2006) Fixed case-insensitive search for character sets. VERSION 1.1c (December 2004) Fixed the issue with #matchesOnStream:do: which caused infinite loops for matches that matched empty strings. VERSION 1.1b (November 2001) Changes valueNowOrOnUnwindDo: to ensure:, plus incorporates some earlier fixes. VERSION 1.1a (May 2001) 1. Support for keeping track of multiple subexpressions. 2. Dot (.) matches anything but NUL character, as it should per POSIX spec. 3. Some bug fixes. VERSION 1.1 (October 1999) Regular expression syntax corrections and enhancements: 1. Backslash escapes similar to those in Perl are allowed in patterns: \w any word constituent character (equivalent to [a-zA-Z0-9_]) *** underscore only since 1.3 *** \W any character but a word constituent (equivalent to [^a-xA-Z0-9_] *** underscore only since 1.3 *** \d a digit (same as [0-9]) \D anything but a digit \s a whitespace character \S anything but a whitespace character \b an empty string at a word boundary \B an empty string not at a word boundary \< an empty string at the beginning of a word \> an empty string at the end of a word For example, ''\w+'' is now a valid expression matching any word. 2. The following backslash escapes are also allowed in character sets (between square brackets): \w, \W, \d, \D, \s, and \S. 3. The following grep(1)-compatible named character classes are recognized in character sets as well: [:alnum:] [:alpha:] [:cntrl:] [:digit:] [:graph:] [:lower:] [:print:] [:punct:] [:space:] [:upper:] [:xdigit:] For example, the following patterns are equivalent: ''[[:alnum:]_]+'' ''\w+'' ''[\w]+'' ''[a-zA-Z0-9_]+'' *** underscore only since 1.3 *** 4. Some non-printable characters can be represented in regular expressions using a common backslash notation: \t tab (Character tab) \n newline (Character lf) \r carriage return (Character cr) \f form feed (Character newPage) \e escape (Character esc) 5. A dot is corectly interpreted as ''any character but a newline'' instead of ''anything but whitespace''. 6. Case-insensitive matching. The easiest access to it are new messages CharacterArray understands: #asRegexIgnoringCase, #matchesRegexIgnoringCase:, #prefixMatchesRegexIgnoringCase:. 7. The matcher (an instance of RxMatcher, the result of String>>asRegex) now provides a collection-like interface to matches in a particular string or on a particular stream, as well as substitution protocol. The interface includes the following messages: matchesIn: aString matchesIn: aString collect: aBlock matchesIn: aString do: aBlock matchesOnStream: aStream matchesOnStream: aStream collect: aBlock matchesOnStream: aStream do: aBlock copy: aString translatingMatchesUsing: aBlock copy: aString replacingMatchesWith: replacementString copyStream: aStream to: writeStream translatingMatchesUsing: aBlock copyStream: aStream to: writeStream replacingMatchesWith: aString Examples: ''\w+'' asRegex matchesIn: ''now is the time'' returns an OrderedCollection containing four strings: ''now'', ''is'', ''the'', and ''time''. ''\<t\w+'' asRegexIgnoringCase copy: ''now is the Time'' translatingMatchesUsing: [:match | match asUppercase] returns ''now is THE TIME'' (the regular expression matches words beginning with either an uppercase or a lowercase T). ACKNOWLEDGEMENTS Since the first release of the matcher, thanks to the input from several fellow Smalltalkers, I became convinced a native Smalltalk regular expression matcher was worth the effort to keep it alive. For the contributions, suggestions, and bug reports that made this release possible, I want to thank: Felix Hack Peter Hatch Alan Knight Eliot Miranda Thomas Muhr Robb Shecter David N. Smith Francis Wolinski and anyone whom I haven''t yet met or heard from, but who agrees this has not been a complete waste of time.!!' readStream nextChunkText) key: #changelog! ----- Method: RegexHelp class>>examples (in category 'as yet unclassified') ----- examples "This method was automatically generated. Edit it using:" "RegexHelp edit: #examples" ^(HelpTopic title: 'Examples' contents: 'As the introductions said, a great use for regular expressions is userinput validation. Following are a few examples of regular expressionsthat might be handy in checking input entered by the user in an inputfield. Try them out by entering something between the quotes andprint-iting. (Also, try to imagine Smalltalk code that each validationwould require if coded by hand). Most example expressions could havebeen written in alternative ways. Checking if aString may represent a nonnegative integer number: '''' matchesRegex: '':isDigit:+''or '''' matchesRegex: ''[0-9]+''or '''' matchesRegex: ''\d+'' Checking if aString may represent an integer number with an optionalsign in front: '''' matchesRegex: ''(\+|-)?\d+'' Checking if aString is a fixed-point number, with at least one digitis required after a dot: '''' matchesRegex: ''(\+|-)?\d+(\.\d+)?'' The same, but allow notation like `123.'': '''' matchesRegex: ''(\+|-)?\d+(\.\d*)?'' Recognizer for a string that might be a name: one word with firstcapital letter, no blanks, no digits. More traditional: '''' matchesRegex: ''[A-Z][A-Za-z]*'' more Smalltalkish: '''' matchesRegex: '':isUppercase::isAlphabetic:*'' A date in format MMM DD, YYYY with any number of spaces in between, inXX century: '''' matchesRegex: ''(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)[ ]+(\d\d?)[ ]*,[ ]*19(\d\d)'' Note parentheses around some components of the expression above. As`Usage'' section shows, they will allow us to obtain the actual stringsthat have matched them (i.e. month name, day number, and year number). For dessert, coming back to numbers: here is a recognizer for ageneral number format: anything like 999, or 999.999, or -999.999e+21. '''' matchesRegex: ''(\+|-)?\d+(\.\d*)?((e|E)(\+|-)?\d+)?''!!!!!!' readStream nextChunkText) key: #examples! ----- Method: RegexHelp class>>implementationNotes (in category 'as yet unclassified') ----- implementationNotes "This method was automatically generated. Edit it using:" "RegexHelp edit: #implementationNotes" ^(HelpTopic title: 'Implementation Notes' contents: 'WHAT TO LOOK AT FIRST String>>matchesRegex: -- in 90% cases this method is all you need to access the package. RxParser -- accepts a string or a stream of characters with a regular expression, and produces a syntax tree corresponding to the expression. The tree is made of instances of Rxs<whatever> classes. RxMatcher -- accepts a syntax tree of a regular expression built by the parser and compiles it into a matcher: a structure made of instances of Rxm<whatever> classes. The RxMatcher instance can test whether a string or a positionable stream of characters matches the original regular expression, or search a string or a stream for substrings matching the expression. After a match is found, the matcher can report a specific string that matched the whole expression, or any parenthesized subexpression of it. All other classes support the above functionality and are used by RxParser, RxMatcher, or both. CAVEATS The matcher is similar in spirit, but NOT in the design--let alone the code--to the original Henry Spencer''s regular expression implementation in C. The focus is on simplicity, not on efficiency. I didn''t optimize or profile anything. I may in future--or I may not: I do this in my spare time and I don''t promise anything. The matcher passes H. Spencer''s test suite (see ''test suite'' protocol), with quite a few extra tests added, so chances are good there are not too many bugs. But watch out anyway. EXTENSIONS, FUTURE, ETC. With the existing separation between the parser, the syntax tree, and the matcher, it is easy to extend the system with other matchers based on other algorithms. In fact, I have a DFA-based matcher right now, but I don''t feel it is good enough to include it here. I might add automata-based matchers later, but again I don''t promise anything. HOW TO REACH ME As of today (December 20, 2000), you can contact me at <vassili(a)parcplace.com>. If this doesn''t work, look around comp.lang.smalltalk or comp.lang.lisp. !!' readStream nextChunkText) key: #implementationNotes! ----- Method: RegexHelp class>>introduction (in category 'as yet unclassified') ----- introduction "This method was automatically generated. Edit it using:" "RegexHelp edit: #introduction" ^(HelpTopic title: 'Introduction' contents: 'A regular expression is a template specifying a class of strings. A regular expression matcher is an tool that determines whether a string belongs to a class specified by a regular expression. This is a common task of a user input validation code, and the use of regular expressions can GREATLY simplify and speed up development of such code. As an example, here is how to verify that a string is a valid hexadecimal number in Smalltalk notation, using this matcher package: aString matchesRegex: ''16r[[:xdigit:]]+'' (Coding the same "the hard way'''' is an exercise to a curious reader). This matcher is offered to the Smalltalk community in hope it will be useful. It is free in terms of money, and to a large extent -- in terms of rights of use. Refer to the "Boring Stuff" section for legalese. The "Syntax" section explains the recognized syntax of regular expressions. The "Usage" section explains matcher capabilities that go beyond what String>>matchesRegex: method offers. The "Implementation Notes" sections says a few words about what is under the hood. The "Changelog" section describes the functionality introduced in 1.1 release. Happy hacking, --Vassili Bykov <vassili(a)objectpeople.com> <vassili(a)magma.ca> !! ]style[(479 40 712),daString matchesRegex: ''16r[[:xdigit:]]+'';;,!!' readStream nextChunkText) key: #introduction! ----- Method: RegexHelp class>>license (in category 'as yet unclassified') ----- license "This method was automatically generated. Edit it using:" "RegexHelp edit: #license" ^(HelpTopic title: 'License' contents: 'The Regular Expression Matcher (``The Software'''') is Copyright (C) 1996, 1999 Vassili Bykov. It is provided to the Smalltalk community in hope it will be useful. 1. This license applies to the package as a whole, as well as to any component of it. By performing any of the activities described below, you accept the terms of this agreement. 2. The software is provided free of charge, and "as is'''', in hope that it will be useful, with ABSOLUTELY NO WARRANTY. The entire risk and all responsibility for the use of the software is with you. Under no circumstances the author may be held responsible for loss of data, loss of profit, or any other damage resulting directly or indirectly from the use of the software, even if the damage is caused by defects in the software. 3. You may use this software in any applications you build. 4. You may distribute this software provided that the software documentation and copyright notices are included and intact. 5. You may create and distribute modified versions of the software, such as ports to other Smalltalk dialects or derived work, provided that: a. any modified version is expressly marked as such and is not misrepresented as the original software; b. credit is given to the original software in the source code and documentation of the derived work; c. the copyright notice at the top of this document accompanies copyright notices of any modified version.!!' readStream nextChunkText) key: #license! ----- Method: RegexHelp class>>pages (in category 'as yet unclassified') ----- pages ^ #(introduction syntax examples usage implementationNotes license changelog)! ----- Method: RegexHelp class>>syntax (in category 'as yet unclassified') ----- syntax "This method was automatically generated. Edit it using:" "RegexHelp edit: #syntax" ^(HelpTopic title: 'Syntax' contents: 'The simplest regular expression is a single character. It matchesexactly that character. A sequence of characters matches a string withexactly the same sequence of characters: ''a'' matchesRegex: ''a'' -- true ''foobar'' matchesRegex: ''foobar'' -- true ''blorple'' matchesRegex: ''foobar'' -- false The above paragraph introduced a primitive regular expression (acharacter), and an operator (sequencing). Operators are applied toregular expressions to produce more complex regular expressions.Sequencing (placing expressions one after another) as an operator is,in a certain sense, ''invisible''--yet it is arguably the most common. A more ''visible'' operator is Kleene closure, more often simplyreferred to as ''a star''. A regular expression followed by an asteriskmatches any number (including 0) of matches of the originalexpression. For example: ''ab'' matchesRegex: ''a*b'' -- true ''aaaaab'' matchesRegex: ''a*b'' -- true ''b'' matchesRegex: ''a*b'' -- true ''aac'' matchesRegex: ''a*b'' -- false: b does not match A star''s precedence is higher than that of sequencing. A star appliesto the shortest possible subexpression that precedes it. For example,''ab*'' means ''a followed by zero or more occurrences of b'', not ''zeroor more occurrences of ab'': ''abbb'' matchesRegex: ''ab*'' -- true ''abab'' matchesRegex: ''ab*'' -- false To actually make a regex matching ''zero or more occurrences of ab'',''ab'' is enclosed in parentheses: ''abab'' matchesRegex: ''(ab)*'' -- true ''abcab'' matchesRegex: ''(ab)*'' -- false: c spoils the fun Two other operators similar to ''*'' are ''+'' and ''?''. ''+'' (positiveclosure, or simply ''plus'') matches one or more occurrences of theoriginal expression. ''?'' (''optional'') matches zero or one, but nevermore, occurrences. ''ac'' matchesRegex: ''ab*c'' -- true ''ac'' matchesRegex: ''ab+c'' -- false: need at least one b ''abbc'' matchesRegex: ''ab+c'' -- true ''abbc'' matchesRegex: ''ab?c'' -- false: too many b''s As we have seen, characters ''*'', ''+'', ''?'', ''('', and '')'' have specialmeaning in regular expressions. If one of them is to be usedliterally, it should be quoted: preceded with a backslash. (Thus,backslash is also special character, and needs to be quoted for aliteral match--as well as any other special character describedfurther). ''ab*'' matchesRegex: ''ab*'' -- false: star in the right string is special ''ab*'' matchesRegex: ''ab\*'' -- true ''a\c'' matchesRegex: ''a\\c'' -- true The last operator is ''|'' meaning ''or''. It is placed between tworegular expressions, and the resulting expression matches if one ofthe expressions matches. It has the lowest possible precedence (lowerthan sequencing). For example, ''ab*|ba*'' means ''a followed by anynumber of b''s, or b followed by any number of a''s'': ''abb'' matchesRegex: ''ab*|ba*'' -- true ''baa'' matchesRegex: ''ab*|ba*'' -- true ''baab'' matchesRegex: ''ab*|ba*'' -- false A bit more complex example is the following expression, matching thename of any of the Lisp-style ''car'', ''cdr'', ''caar'', ''cadr'',... functions: c(a|d)+r It is possible to write an expression matching an empty string, forexample: ''a|''. However, it is an error to apply ''*'', ''+'', or ''?'' tosuch expression: ''(a|)*'' is an invalid expression. So far, we have used only characters as the ''smallest'' components ofregular expressions. There are other, more ''interesting'', components. A character set is a string of characters enclosed in squarebrackets. It matches any single character if it appears between thebrackets. For example, ''[01]'' matches either ''0'' or ''1'': ''0'' matchesRegex: ''[01]'' -- true ''3'' matchesRegex: ''[01]'' -- false ''11'' matchesRegex: ''[01]'' -- false: a set matches only one character Using plus operator, we can build the following binary numberrecognizer: ''10010100'' matchesRegex: ''[01]+'' -- true ''10001210'' matchesRegex: ''[01]+'' -- false If the first character after the opening bracket is ''^'', the set isinverted: it matches any single character *not* appearing between thebrackets: ''0'' matchesRegex: ''[^01]'' -- false ''3'' matchesRegex: ''[^01]'' -- true For convenience, a set may include ranges: pairs of charactersseparated with ''-''. This is equivalent to listing all charactersbetween them: ''[0-9]'' is the same as ''[0123456789]''. Special characters within a set are ''^'', ''-'', and '']'' that closes theset. Below are the examples of how to literally use them in a set: [01^] -- put the caret anywhere except the beginning [01-] -- put the dash as the last character []01] -- put the closing bracket as the first character [^]01] (thus, empty and universal sets cannot be specified) Regular expressions can also include the following backquote escapesto refer to popular classes of characters: \w any word constituent character (same as [a-zA-Z0-9_]) \W any character but a word constituent \d a digit (same as [0-9]) \D anything but a digit \s a whitespace character (same as [:space:] below) \S anything but a whitespace character These escapes are also allowed in character classes: ''[\w+-]'' means''any character that is either a word constituent, or a plus, or aminus''. Character classes can also include the following grep(1)-compatibleelements to refer to: [:alnum:] any alphanumeric character (same as [a-zA-Z0-9]) [:alpha:] any alphabetic character (same as [a-zA-Z]) [:cntrl:] any control character. (any character with code < 32) [:digit:] any decimal digit (same as [0-9]) [:graph:] any graphical character. (any character with code >= 32). [:lower:] any lowercase character (including non-ASCII lowercase characters) [:print:] any printable character. In this version, this is the same as [:graph:] [:punct:] any punctuation character: . , !!!!!!!! ? ; : '' - ( ) '' and double quotes [:space:] any whitespace character (space, tab, CR, LF, null, form feed, Ctrl-Z, 16r2000-16r200B, 16r3000) [:upper:] any uppercase character (including non-ASCII uppercase characters) [:xdigit:] any hexadecimal character (same as [a-fA-F0-9]). Note that many of these are only as consistent or inconsistent on issuesof locale as the underlying Smalltalk implementation. Values shown hereare for VisualWorks 7.6. Note that these elements are components of the character classes,i.e. they have to be enclosed in an extra set of square brackets toform a valid regular expression. For example, a non-empty string ofdigits would be represented as ''[[:digit:]]+''. The above primitive expressions and operators are common to manyimplementations of regular expressions. The next primitive expressionis unique to this Smalltalk implementation. A sequence of characters between colons is treated as a unary selectorwhich is supposed to be understood by Characters. A character matchessuch an expression if it answers true to a message with thatselector. This allows a more readable and efficient way of specifyingcharacter classes. For example, ''[0-9]'' is equivalent to '':isDigit:'',but the latter is more efficient. Analogously to character sets,character classes can be negated: '':^isDigit:'' matches a Characterthat answers false to #isDigit, and is therefore equivalent to''[^0-9]''. As an example, so far we have seen the following equivalent ways towrite a regular expression that matches a non-empty string of digits: ''[0-9]+'' ''\d+'' ''[\d]+'' ''[[:digit:]]+'' '':isDigit:+'' The last group of special primitive expressions includes: . matching any character except a NULL; ^ matching an empty string at the beginning of a line; $ matching an empty string at the end of a line. \b an empty string at a word boundary \B an empty string not at a word boundary \< an empty string at the beginning of a word \> an empty string at the end of a word ''axyzb'' matchesRegex: ''a.+b'' -- true ''ax zb'' matchesRegex: ''a.+b'' -- true (space is matched by ''.'') ''axzb'' matchesRegex: ''a.+b'' -- true (carriage return is matched by ''.'') Again, the dot ., caret ^ and dollar $ characters are special and should be quotedto be matched literally.!! ]style[(179 21 7851),f5,!!' readStream nextChunkText) key: #syntax! ----- Method: RegexHelp class>>usage (in category 'as yet unclassified') ----- usage "This method was automatically generated. Edit it using:" "RegexHelp edit: #usage" ^(HelpTopic title: 'Usage' contents: 'The preceding section covered the syntax of regular expressions. It used the simplest possible interface to the matcher: sending #matchesRegex: message to the sample string, with regular expression string as the argument. This section explains hairier ways of using the matcher. PREFIX MATCHING AND CASE-INSENSITIVE MATCHING A CharacterArray (an EsString in VA) also understands these messages: #prefixMatchesRegex: regexString #matchesRegexIgnoringCase: regexString #prefixMatchesRegexIgnoringCase: regexString #prefixMatchesRegex: is just like #matchesRegex, except that the whole receiver is not expected to match the regular expression passed as the argument; matching just a prefix of it is enough. For example: ''abcde'' matchesRegex: ''(a|b)+'' -- false ''abcde'' prefixMatchesRegex: ''(a|b)+'' -- true The last two messages are case-insensitive versions of matching. ENUMERATION INTERFACE An application can be interested in all matches of a certain regular expression within a String. The matches are accessible using a protocol modelled after the familiar Collection-like enumeration protocol: #regex: regexString matchesDo: aBlock Evaluates a one-argument <aBlock> for every match of the regular expression within the receiver string. #regex: regexString matchesCollect: aBlock Evaluates a one-argument <aBlock> for every match of the regular expression within the receiver string. Collects results of evaluations and anwers them as a SequenceableCollection. #allRegexMatches: regexString Returns a collection of all matches (substrings of the receiver string) of the regular expression. It is an equivalent of <aString regex: regexString matchesCollect: [:each | each]>. #allRangesOfRegexMatches: regexString Returns a collection of all character ranges (startIndex to: stopIndex) that match the regular expression. REPLACEMENT AND TRANSLATION It is possible to replace all matches of a regular expression with a certain string using the message: #copyWithRegex: regexString matchesReplacedWith: aString For example: ''ab cd ab'' copyWithRegex: ''(a|b)+'' matchesReplacedWith: ''foo'' A more general substitution is match translation: #copyWithRegex: regexString matchesTranslatedUsing: aBlock This message evaluates a block passing it each match of the regular expression in the receiver string and answers a copy of the receiver with the block results spliced into it in place of the respective matches. For example: ''ab cd ab'' copyWithRegex: ''(a|b)+'' matchesTranslatedUsing: [:each | each asUppercase] All messages of enumeration and replacement protocols perform a case-sensitive match. Case-insensitive versions are not provided as part of a CharacterArray protocol. Instead, they are accessible using the lower-level matching interface. LOWER-LEVEL INTERFACE Internally, #matchesRegex: works as follows: 1. A fresh instance of RxParser is created, and the regular expression string is passed to it, yielding the expression''s syntax tree. 2. The syntax tree is passed as an initialization parameter to an instance of RxMatcher. The instance sets up some data structure that will work as a recognizer for the regular expression described by the tree. 3. The original string is passed to the matcher, and the matcher checks for a match. THE MATCHER If you repeatedly match a number of strings against the same regular expression using one of the messages defined in CharacterArray, the regular expression string is parsed and a matcher is created anew for every match. You can avoid this overhead by building a matcher for the regular expression, and then reusing the matcher over and over again. You can, for example, create a matcher at a class or instance initialization stage, and store it in a variable for future use. You can create a matcher using one of the following methods: - Sending #forString:ignoreCase: message to RxMatcher class, with the regular expression string and a Boolean indicating whether case is ignored as arguments. - Sending #forString: message. It is equivalent to <... forString: regexString ignoreCase: false>. A more convenient way is using one of the two matcher-created messages understood by CharacterArray. - <regexString asRegex> is equivalent to <RxMatcher forString: regexString>. - <regexString asRegexIgnoringCase> is equivalent to <RxMatcher forString: regexString ignoreCase: true>. Here are four examples of creating a matcher: hexRecognizer := RxMatcher forString: ''16r[0-9A-Fa-f]+'' hexRecognizer := RxMatcher forString: ''16r[0-9A-Fa-f]+'' ignoreCase: false hexRecognizer := ''16r[0-9A-Fa-f]+'' asRegex hexRecognizer := ''16r[0-9A-F]+'' asRegexIgnoringCase MATCHING The matcher understands these messages (all of them return true to indicate successful match or search, and false otherwise): matches: aString True if the whole target string (aString) matches. matchesPrefix: aString True if some prefix of the string (not necessarily the whole string) matches. search: aString Search the string for the first occurrence of a matching substring. (Note that the first two methods only try matching from the very beginning of the string). Using the above example with a matcher for `a+'', this method would answer success given a string `baaa'', while the previous two would fail. matchesStream: aStream matchesStreamPrefix: aStream searchStream: aStream Respective analogs of the first three methods, taking input from a stream instead of a string. The stream must be positionable and peekable. All these methods answer a boolean indicating success. The matcher also stores the outcome of the last match attempt and can report it: lastResult Answers a Boolean -- the outcome of the most recent match attempt. If no matches were attempted, the answer is unspecified. SUBEXPRESSION MATCHES After a successful match attempt, you can query the specifics of which part of the original string has matched which part of the whole expression. A subexpression is a parenthesized part of a regular expression, or the whole expression. When a regular expression is compiled, its subexpressions are assigned indices starting from 1, depth-first, left-to-right. For example, `((ab)+(c|d))?ef'' includes the following subexpressions with these indices: 1: ((ab)+(c|d))?ef 2: (ab)+(c|d) 3: ab 4: c|d After a successful match, the matcher can report what part of the original string matched what subexpression. It understandards these messages: subexpressionCount Answers the total number of subexpressions: the highest value that can be used as a subexpression index with this matcher. This value is available immediately after initialization and never changes. subexpression: anIndex An index must be a valid subexpression index, and this message must be sent only after a successful match attempt. The method answers a substring of the original string the corresponding subexpression has matched to. subBeginning: anIndex subEnd: anIndex Answer positions within the original string or stream where the match of a subexpression with the given index has started and ended, respectively. This facility provides a convenient way of extracting parts of input strings of complex format. For example, the following piece of code uses the ''MMM DD, YYYY'' date format recognizer example from the `Syntax'' section to convert a date to a three-element array with year, month, and day strings (you can select and evaluate it right here): | matcher | matcher := RxMatcher forString: ''(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)[ ]+(:isDigit::isDigit:?)[ ]*,[ ]*(19|20)(:isDigit::isDigit:)''. (matcher matches: ''Aug 6, 1996'') ifTrue: [Array with: (matcher subexpression: 5) with: (matcher subexpression: 2) with: (matcher subexpression: 3)] ifFalse: [''no match''] (should answer ` #(''96'' ''Aug'' ''6'')''). ENUMERATION AND REPLACEMENT The enumeration and replacement protocols exposed in CharacterArray are actually implemented by the matcher. The following messages are understood: #matchesIn: aString #matchesIn: aString do: aBlock #matchesIn: aString collect: aBlock #copy: aString replacingMatchesWith: replacementString #copy: aString translatingMatchesUsing: aBlock #matchingRangesIn: aString #matchesOnStream: aStream #matchesOnStream: aStream do: aBlock #matchesOnStream: aStream collect: aBlock #copy: sourceStream to: targetStream replacingMatchesWith: replacementString #copy: sourceStream to: targetStream translatingMatchesWith: aBlock Note that in those methods that take a block, the block may refer to the rxMatcher itself, e.g. to collect information about the position the match occurred at, or the subexpressions of the match. An example can be seen in #matchingRangesIn: ERROR HANDLING Exception signaling objects (Signals in VisualWorks, Exceptions in VisualAge) are accessible through RxParser class protocol. To handle possible errors, use the protocol described below to obtain the exception objects and use the protocol of the native Smalltalk implementation to handle them. If a syntax error is detected while parsing expression, RxParser>>syntaxErrorSignal is raised/signaled. If an error is detected while building a matcher, RxParser>>compilationErrorSignal is raised/signaled. If an error is detected while matching (for example, if a bad selector was specified using `:<selector>:'' syntax, or because of the matcher''s internal error), RxParser>>matchErrorSignal is raised RxParser>>regexErrorSignal is the parent of all three. Since any of the three signals can be raised within a call to #matchesRegex:, it is handy if you want to catch them all. For example: VisualWorks: RxParser regexErrorSignal handle: [:ex | ex returnWith: nil] do: [''abc'' matchesRegex: ''))garbage[''] VisualAge: [''abc'' matchesRegex: ''))garbage[''] when: RxParser regexErrorSignal do: [:signal | signal exitWith: nil]!!' readStream nextChunkText) key: #usage!

2 1

ResumableTestFailure
by Thiede, Christoph 13 Sep '22

13 Sep '22

Hi all! :-) I was just working a bit with Squeak by Example (the book) and found out that resumable assertions as described in the book (see chapter SUnit, Advanced, Continuing after a failure) do not any longer work in a trunk image. If you write a test case with some resumable assertions, turn on logging, and run the test, only the first failure will be logged, not the rest, too. (Methods to reproduce: isLogging ^ true "You will need to open a Transcript" testResumable #(1 2 3 4 5) do: [:i | self assert: i even description: ['{1} is not even' format: {i}] resumable: true] ) A second issue I experienced frequently when I felt the desire to work with resumable assertions was that they are basically only available for two selectors: #assert:description:resumable: and #deny:description:identical:. I often would like to insert a resumable into any test without being that verbose, and nearly redoubling the number of assertion selectors by adding a resumable argument to each of them did not feel right, either. Please find the attached changeset, which fixes both issues. * In TestResult >> #runCase:, resumable assertions are identified and - finally - resumed, after marking the test as failed. Unfortunately, other test runners (yes, I'm thinking of smalltalkCI) will need to reimplement this feature unless they reuse the TestResult implementation. * In TestCase, I refactored some methods that raise assertion failures and added the most basic #signalFailure:resumable: as the single source of signalization. Instead of setting the resumable bit manually for each assertion, you now can also make the whole test execution resumable by: * overriding #isResumable, * setting a resumable flag using #resumable:, * setting a resumable flag temporarily using #beResumable:during:, or * adding a <resumable> or <resumable: true> pragma to the test method. * In addition, I tested the added and fixed behavior in ResumableTestFailureTestCase. Looking forward to your review! :-) Best, Christoph

2 2

The Inbox: Kernel-ct.1294.mcz
by commits＠source.squeak.org 09 Sep '22

09 Sep '22

A new version of Kernel was added to project The Inbox: http://source.squeak.org/inbox/Kernel-ct.1294.mcz ==================== Summary ==================== Name: Kernel-ct.1294 Author: ct Time: 9 January 2020, 3:42:36.672555 pm UUID: 7409dcc3-160e-9d44-9712-f266e0f25973 Ancestors: Kernel-nice.1292 Proposal: Add #haltOnceIf: to Object, which combines the strengths of #haltOnce and #haltIf:. Especially useful in performance-critical situations: It HaltOnce is disabled, the condition will not be evaluated at all. =============== Diff against Kernel-nice.1292 =============== Item was changed: ----- Method: Object>>haltIf: (in category 'debugging') ----- + haltIf: aCondition + "This is the typical message to use for inserting breakpoints during debugging. Param can be a block or expression, halt if true. - haltIf: condition - "This is the typical message to use for inserting breakpoints during - debugging. Param can be a block or expression, halt if true. If the Block has one arg, the receiver is bound to that. + If the condition is a selector, we look up in the callchain. Halt if any method's selector equals selector." - If the condition is a selector, we look up in the callchain. Halt if - any method's selector equals selector." - | cntxt | + ^ (self meetsHaltCondition: aCondition) + ifTrue: [self halt]! - condition isSymbol ifTrue:[ - "only halt if a method with selector symbol is in callchain" - cntxt := thisContext. - [cntxt sender isNil] whileFalse: [ - cntxt := cntxt sender. - (cntxt selector = condition) ifTrue: [Halt signal]. - ]. - ^self. - ]. - (condition isBlock - ifTrue: [condition cull: self] - ifFalse: [condition] - ) ifTrue: [ - Halt signal - ].! Item was added: + ----- Method: Object>>haltOnceIf: (in category 'debugging') ----- + haltOnceIf: aCondition + "Check aCondition and halt, unless we have already done it once. See #meetsHaltCondition:." + + self haltOnceEnabled ifFalse: [^ self]. + ^ (self meetsHaltCondition: aCondition) + ifTrue: [ + self clearHaltOnce; halt].! Item was added: + ----- Method: Object>>meetsHaltCondition: (in category 'debugging') ----- + meetsHaltCondition: aCondition + "If the Block has one arg, the receiver is bound to that. + If the condition is a selector, we look up in the callchain. Halt if any method's selector equals selector." + + aCondition isSymbol ifTrue: [ | ctxt | + "only halt if a method with selector symbol is in callchain" + ctxt := thisContext. + [(ctxt := ctxt sender) isNil] whileFalse: [ + (ctxt selector = aCondition) ifTrue: [^ true] ]. + ^ false ]. + ^ aCondition isBlock + ifTrue: [aCondition cull: self] + ifFalse: [aCondition value]!

2 1

Jump to page: