Vm-dev September 2017

vm-dev@lists.squeakfoundation.org

27 participants
18 discussions

win32 SetEvent, ResetEvent and MsgWaitForMultipleObjects
by Holger Freyther 15 Sep '17

15 Sep '17

Hi Eliot, Bert, as part of reading the "goToSleep" primitive I had a look at Windows and the last time I dealt with win32 API was quite some time ago so maybe I am wrong but doesn't the time slice below look like a race condition? If not what am I missing? VM Thread Socket Read Thread ioRelinquishProcess.. SIGNAL(pss->readSema) -> SetEvent(vmWakeUpEvent) ResetEvent(vmWakeUpEvent) MsgWaitForMultipleObjects... Couldn't we have lost the vmWakeUpEvent for the socket and then wait the full timeout? The history in Opensmalltalk-vm doesn't go far but couldn't we create the vmWakeUpEvent with the autoreset flag? thank you holger Code snippets: sqInt ioRelinquishProcessorForMicroseconds(sqInt microSeconds) { /* wake us up if something happens */ ResetEvent(vmWakeUpEvent); if (WAIT_TIMEOUT == MsgWaitForMultipleObjects(1, &vmWakeUpEvent, FALSE, microSeconds / 1000, QS_ALLINPUT)) int synchronizedSignalSemaphoreWithIndex(int semaIndex) { int result; /* Do our job - this is now synchronized in signalSemaphoreWithIndex */ result = signalSemaphoreWithIndex(semaIndex); /* wake up interpret() if sleeping */ SetEvent(vmWakeUpEvent); static DWORD WINAPI readWatcherThread(privateSocketStruct *pss) { struct timeval tv= { 1000, 0 }; /* Timeout value == 1000 sec */ fd_set fds; int n, doWait; while(1) { ... switch(pss->readWatcherOp) { case WatchData: /* Data may be available */ pss->sockState |= SOCK_DATA_READABLE; SIGNAL(pss->readSema);

1 1

Use less CPU (improve battery life or reduce cost in the cloud)
by Holger Freyther 08 Sep '17

08 Sep '17

Hi, I have done some early prototype for the Unix VM end of 2015(?) and I have improved and repeated these for MacOS now and the thread based heartbeat (now that it is the universal default). I won't make it to ESUG this year but this might be something to play with? The motivations is simple: Polling increases the CPU usage which will reduce your battery life, takes away resources from other processes (e.g. more Pharo images) or these days increases your cloud computing bill. On top of that it might increase network latency (time from socket becoming readable to the time the semaphore is signaled). To complete the work we have work inside the Image and the VM and some of it is on the way and others might need more discussion. The idle process: ProcessorScheduler>># idleProcess "A default background process which is invisible." [true] whileTrue: [self relinquishProcessorForMicroseconds: 1000] Let's please yield the CPU for more than a 1ms. Unless I am missing something an expired Delay or network IO would make us wake up earlier anyway? The delay scheduler: The VM supports that when the next wake-up time is set to 0, the VM can sleep indefinitely. There is a pending patch to sleep "0" in our Delay scheduler. Currently we force a wake-up earlier than that. I think we should trust the VM to do wake us up even if it is a second away. Morphic UI: I don't understand the WorldState>>#interCyclePause but then I never looked at Morphic. Do we really need to poll like that? Under which circumstances does the world update? We get an event (where we have the event semaphore), we get some I/O (where we have a semaphore) or we have a timeout (where we sleep on a semaphore). Did anyone ever look at removing the tick? VM I/O: Currently we receive a SIGIO but from what I can see (and I still need to write a benchmark) the processing might be delayed 20ms? My hack removes the usage of nextPollUsecs and instead checks a variable that is set by the SIGIO handler. Besides missing memory barriers this should work(tm). The biggest issue seems that for macos/ios the input is driven by polling. E.g. some wheel events seem to require to pump the event queue. Is this something we could trigger from the image in the future? I had hoped to get a fd to a machport we could get SIGIO for.. but that doesn't seem to exist. I have hacked out the honoring of the relinquish delay, added the polling into a iOS specific routine and thanks to the Morphic Delay we bump the event loop frequently enough. VM heartbeat thread: The process keeps ticking even if the VM doesn't run. E.g. sleeps and waits for an event. There is a cost in deciding when to halt the thread so there must be a cut-off for which delays we bother to disable the heartbeat thread. I think the current code would allow the hearbeat to drift so the new code might just make it a bit worse. Where are we now? I have pushed my changes to https://github.com/zecke/opensmalltalk-vm/tree/mac-use-less-cpu and would be happy to have people look at it, look at the memory synchronization, maybe run to see if they notice extra delays or such? I started the same image with the plain-vm and my hacked one and let it run for about 20min. The output is coming from top. COMMAND %CPU TIME Pharo 4.3 00:48.49 Pharo 0.8 00:10.20 Looking for comments and feedback. regards holger

8 8

[OpenSmalltalk/opensmalltalk-vm] 7c1485: try without openssl
by GitHub 07 Sep '17

07 Sep '17

Branch: refs/heads/krono/winssltest Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 7c14856285a2fd7622fb79022e73adf27ef28824 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/7c14856285a2fd7622… Author: Tobias Pape <tobias(a)netshed.de> Date: 2017-09-07 (Thu, 07 Sep 2017) Changed paths: M .appveyor.yml M platforms/win32/plugins/SqueakSSL/Makefile.plugin Log Message: ----------- try without openssl

1 0

Custom module primitive error codes
by monty 06 Sep '17

06 Sep '17

#primitiveFailFor: can be used to signal a primitive failure with an integer error code. The VM can map a small set of these codes to symbols like #'bad argument' that can be checked for in <primitive:...error:> methods. How should a module primitive use #primitiveFailFor: with its own custom error codes, so that they don't collide with the VM's current and future set of PrimErr* codes? This RSqueak issue says to use negative integers for application-specific errors: https://github.com/HPI-SWA-Lab/RSqueak/issues/130 Is that preferred for modules?

1 0

Re: [Vm-dev] [squeak-dev] Inbox/Files-monty.172
by Eliot Miranda 06 Sep '17

06 Sep '17

Hi Nicolas, I'm happy to integrate the VMMaker and open Smalltalk-vm changes if you have them. _,,,^..^,,,_ (phone) > On Aug 30, 2017, at 1:13 PM, Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> wrote: > > > > 2017-08-29 8:25 GMT+02:00 Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com>: >> Not only my eyes are old, my brain too: here is the original intention of Monty: >> >> https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/140/commits/c1db1eb4… >> >> So it avoids file existence test in pre-condition, not in post-condition... >> I swear I had seen other changes about passing failure status back to image for avoiding post-condition tests, but maybe it was not in this batch of changes. > > Ah, it seems I'm senile since I can't remember it's me who integrated those changes, > but not completely senile since at least I remember that I saw such changes ;) > > ------------------------------- > > Name: VMMaker.oscog-nice.2233 > Author: nice > Time: 5 June 2017, 6:13:56.418116 pm > UUID: ff8286e4-b96f-4ce6-9cb5-e7b286679e03 > Ancestors: VMMaker.oscog-eem.2232 > > Integrate changes from monty to sqFileOpenNew. > > There must be a way to check if failure is due to existence of file (implemented via PrimErrInappropriate). > If we would test existence afterward, the file could have been deleted in interim leading to another race condition. > > This VMMaker change requires a coordinated platform source change (adding a 4th parameter to sqFileOpenNew). > > >> >> Forget about all I wrote before, the changes definitely deserve a proper review. >> Maybe check what happens in case someone did not update the VM properly before applying the image changes... >> >> 2017-08-29 1:36 GMT+02:00 tim Rowledge <tim(a)rowledge.org>: >>> >>> > On 28-08-2017, at 11:47 AM, Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> wrote: >>> > Main request was to avoid case of race condition that can happen when separating the action (like opening a file) from the query in case of failure (like was file unwritable?), because another concurrent program might have changed the file system status in between (think of concurrent squeak images creating concurrent /tmp/foobar.baz for example). >>> >>> Oh, *that* problem. I hate that. It’s such a stupid bit of (mis)design. I think I’ve been complaining about that for around 30 years… >>> >>> > >>> > I have already integrated the necessary VM changes, but it remains to integrate image side changes. >>> > Last time I checked there were unecessary changes mixed that I'd preferred to see discussed here, and the fallback code might not work in older VM, then I came short of spare time... >>> > But in any cases, there are too many .mcz roting in inbox, so I'm glad another pair of eyes analyze the changes. >>> >>> It’s all very well Torvalds declaring that all bugs are shallow with enough pairs of eyes, *if* you have enough pairs of eyes. We need more. >>> >>> >>> tim >>> -- >>> tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim >>> Strange OpCodes: RSC: Rewind System Clock >>> >>> >>> >> > >

2 1

Current VM build problems. (was: Fwd: Errored: OpenSmalltalk/opensmalltalk-vm#835 (Cog - 7dff211))
by Tobias Pape 05 Sep '17

05 Sep '17

Hi all, currently the vm fails to build properly due to two main problems: a) #testBecomeIdentityHash fails on Squeak5.1 for linux32x86 squeak.cog.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505448#L3172 linux32x86 squeak.sista.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505458#L3208 linux32x86 squeak.stack.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505461#L1842 macos32x86 squeak.cog.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505499#L8455 macos32x86 squeak.sista.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505503#L8517 macos32x86 squeak.stack.spur https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505505#L7383 (Or said differently: on 32x86 for squeak.*.spur) I recall we had problems there but I thought those were fixed 2..3 years ago? Someone can comment? b) the pharo deployment seems to not work correctly: https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505445#L1900 https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505471#L8223 https://travis-ci.org/OpenSmalltalk/opensmalltalk-vm/jobs/269505444#L1900 for 64x64 pharo.cog.spur. If it is correct that those 3 vms cannot be deployed, please fix the deploy script to not try to. IF incorrect, pleas fix on your side/site. Best regards -Tobias > Begin forwarded message: > > From: Travis CI <builds(a)travis-ci.org> > Subject: Errored: OpenSmalltalk/opensmalltalk-vm#835 (Cog - 7dff211) > Date: 29. August 2017 um 13:02:09 MESZ > > OpenSmalltalk / opensmalltalk-vm (Cog) > > Build #835 has errored. 48 minutes and 25 seconds > Tobias Pape 7dff211 Changeset → > [squeakssl] follow suit with the unix hostname validation things > Want to know about upcoming build environment updates? > Would you like to stay up-to-date with the upcoming Travis CI build environment updates? We set up a mailing list for you! Sign up here. > Documentation about Travis CI > Need help? Mail support! > Choose who receives these build notification emails in your configuration file. > > Would you like to test your private code? > Travis CI for Private Projects could be your new best friend!

3 10

[OpenSmalltalk/opensmalltalk-vm] e839f1: Add 64bit newspeak.(cog|stack).spur build targets
by GitHub 05 Sep '17

05 Sep '17

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: e839f105c4aacf9d0ef459c3c8cea0cd5683e747 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/e839f105c4aacf9d0e… Author: Fabio Niephaus <code(a)fniephaus.com> Date: 2017-09-04 (Mon, 04 Sep 2017) Changed paths: M .appveyor.yml Log Message: ----------- Add 64bit newspeak.(cog|stack).spur build targets

1 0

VM Maker: VMMaker.oscog-eem.2267.mcz
by commits＠source.squeak.org 02 Sep '17

02 Sep '17

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.2267.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.2267 Author: eem Time: 1 September 2017, 4:16:52.098824 pm UUID: 9a398445-4694-49b8-a1d1-77e877f29295 Ancestors: VMMaker.oscog-eem.2266 Slang: Allow plugins to use sends to access struct members. To this end replace all sends of TSendNode>>isStructSendIn: with CCodeGenerator>>isStructSend:. Plugins wanting to use this facility must override InterpreterPlugin>>#isStructType: to answer true for types denoting structs. There's probably more work to be done (e.g. to TSendNode>>structTargetKindIn: which sends to VMStructType directly). Biut this is enough for the FileAttributesPlugin. =============== Diff against VMMaker.oscog-eem.2266 =============== Item was changed: ----- Method: CCodeGenerator>>generateIfNilAsArgument:on:indent: (in category 'C translation') ----- generateIfNilAsArgument: msgNode on: aStream indent: level "Generate the C code for this message onto the given stream." (self isNilConstantReceiverOf: msgNode) ifFalse: [aStream nextPutAll: '(!!('. msgNode receiver emitCCodeAsArgumentOn: aStream level: 0 generator: self. aStream nextPut: $); crtab: level + 1; nextPut: $?; space. msgNode args last emitCCodeAsArgumentOn: aStream level: level + 2 generator: self. aStream crtab: level + 1; nextPut: $:; space. msgNode receiver emitCCodeAsArgumentOn: aStream level: 0 generator: self. (msgNode receiver isLeaf or: [msgNode receiver isSend + and: [(self isStructSend: msgNode receiver) - and: [(msgNode receiver isStructSendIn: self) and: [msgNode receiver receiver isLeaf]]]) ifFalse: [logger cr; nextPutAll: 'sending ifNil: to non-leaf in '; nextPutAll: currentMethod selector]. aStream nextPut: $)] ifTrue: [msgNode args first emitCCodeAsArgumentOn: aStream level: level generator: self]! Item was added: + ----- Method: CCodeGenerator>>isStructSend: (in category 'utilities') ----- + isStructSend: aTSendNode + "Answer if the argument aTSendNode is a send of a structure accessor. + This is tricky. We want + foo bar => foo->bar + foo bar => foo.bar + foo bar: expr => foo->bar = expr + foo bar: expr => foo.bar = expr + depending on whether foo is a struct or a pointer to a struct, + but only if both foo is a struct type and bar is a field accessor. + The tricky cases are self-sends within struct class methods. Here we need to + distinguish between self-sends of ordinary methods from self sends of accessors." + ^aTSendNode numArgs <= 1 + and: [(aTSendNode receiver structTargetKindIn: self) notNil + and: [(self methodNamed: aTSendNode selector) + ifNil: [false] + ifNotNil: [:method| method isStructAccessor]]]! Item was added: + ----- Method: CCodeGenerator>>structTargetKindForDeclaration: (in category 'C code generator') ----- + structTargetKindForDeclaration: typeName "<String>" + ^VMStructType structTargetKindForDeclaration: typeName! Item was changed: ----- Method: CCodeGenerator>>structTargetKindForVariableName: (in category 'C code generator') ----- structTargetKindForVariableName: varName "<String>" + ^(self typeOfVariable: varName) ifNotNil: + [:declaration| + self structTargetKindForDeclaration: declaration]! - ^VMStructType structTargetKindForDeclaration: (self typeOfVariable: varName)! Item was added: + ----- Method: InterpreterPlugin class>>isStructType: (in category 'translation') ----- + isStructType: typeName "<String>" + "Subclasses should override to answer trye for struct types they use." + ^false! Item was changed: ----- Method: TMethod>>computePossibleSideEffectsInto:visited:in: (in category 'inlining support') ----- computePossibleSideEffectsInto: writtenToVars visited: visitedSelectors in: aCodeGen "Add all variables written to by this method and its callees to writtenToVars. Avoid circularity via visitedSelectors" (visitedSelectors includes: selector) ifTrue: [^self]. visitedSelectors add: selector. writtenToGlobalVarsCache ifNotNil: [writtenToVars addAll: writtenToGlobalVarsCache. ^self]. parseTree nodesDo: [ :node | (node isAssignment and: [(locals includes: node variable name) not]) ifTrue: [writtenToVars add: node variable name]. (node isSend and: [node isBuiltinOperator not + and: [(aCodeGen isStructSend: node) not]]) ifTrue: - and: [(node isStructSendIn: aCodeGen) not]]) ifTrue: [(aCodeGen methodNamed: node selector) ifNotNil: [:method| method computePossibleSideEffectsInto: writtenToVars visited: visitedSelectors in: aCodeGen]]]. writtenToGlobalVarsCache := writtenToVars copy! Item was changed: ----- Method: TMethod>>ensureToByDoLoopLimitIsSafeAndEfficient:in: (in category 'transformations') ----- ensureToByDoLoopLimitIsSafeAndEfficient: node in: aCodeGen "For both safety and efficiency, make sure that to:[by:]do: loops with complex limits have a variable to hold the limit expression. In C the limit expression is evaluated each time round the loop so if the loop has side-effects (which it usually will), the C compiler may not be able to optimize the limit expression itself." | limitExpr hasSideEffects | limitExpr := node args first. hasSideEffects := limitExpr anySatisfy: [:subNode| subNode isSend and: [(aCodeGen isBuiltinSelector: subNode selector) not + and: [(aCodeGen isStructSend: subNode) not]]]. - and: [(subNode isStructSendIn: aCodeGen) not]]]. node args size = 4 ifTrue: [hasSideEffects ifTrue: [locals add: node args last name] ifFalse: [node arguments: node args allButLast]] ifFalse: "If the expression is complex but as yet there is no limit variable, add it" [hasSideEffects ifTrue: [| var | var := self unusedNamePrefixedBy: 'toDoLimit' avoiding: locals. "N.B. adds it to locals!!!!" node arguments: node args, {TVariableNode new setName: var; yourself}. declarations at: node args third args first ifPresent: [:decl| self declarationAt: var put: (self typeFor: node args third args first in: aCodeGen), ' ', var]]]! Item was changed: ----- Method: TMethod>>isSubstitutableNode:intoMethod:in: (in category 'inlining') ----- isSubstitutableNode: aNode intoMethod: targetMeth in: aCodeGen "Answer true if the given parameter node is either a constant, a local variable, or a formal parameter of the receiver. Such parameter nodes may be substituted directly into the body of the method during inlining. Note that global variables cannot be subsituted into methods with possible side effects (i.e., methods that may assign to global variables) because the inlined method might depend on having the value of the global variable captured when it is passed in as an argument." | var | aNode isConstant ifTrue: [ ^ true ]. aNode isVariable ifTrue: [ var := aNode name. ((locals includes: var) or: [args includes: var]) ifTrue: [ ^ true ]. (#(self true false nil) includes: var) ifTrue: [ ^ true ]. (targetMeth maySubstituteGlobal: var in: aCodeGen) ifTrue: [ ^ true ]. ]. "For now allow literal blocks to be substituted. They better be accessed only with value[:value:*] messages though!!" aNode isStmtList ifTrue: [^true]. (aNode isSend and: [aNode numArgs = 0 + and: [aCodeGen isStructSend: aNode]]) ifTrue: - and: [aNode isStructSendIn: aCodeGen]]) ifTrue: [^true]. "scan expression tree; must contain only constants, builtin ops, and inlineable vars" aNode nodesDo: [ :node | node isSend ifTrue: [ (node isBuiltinOperator or: [node numArgs = 0 + and: [aCodeGen isStructSend: node]]) ifFalse: [ ^false ]. - and: [node isStructSendIn: aCodeGen]]) ifFalse: [ ^false ]. ]. node isVariable ifTrue: [ var := node name. ((locals includes: var) or: [(args includes: var) or: [(#(self true false nil) includes: var) or: [targetMeth maySubstituteGlobal: var in: aCodeGen]]]) ifFalse: [ ^ false ]. ]. (node isConstant or: [node isVariable or: [node isSend]]) ifFalse: [ ^false ]. ]. ^ true! Item was changed: ----- Method: TMethod>>tryToInlineMethodExpressionsIn: (in category 'inlining') ----- tryToInlineMethodExpressionsIn: aCodeGen "Expand any (complete) inline methods sent by this method as receivers or parameters. Answer if anything was inlined." | sendsToInline | sendsToInline := Dictionary new: 100. + aCodeGen + pushScope: declarations + while: [parseTree + nodesDo: + [:node| + (self inlineableFunctionCall: node in: aCodeGen) ifTrue: + [(self inlineFunctionCall: node in: aCodeGen) ifNotNil: + [:replacement| + sendsToInline at: node put: replacement]]] + unless: "Don't inline the arguments to asserts to keep the asserts readable" + [:node| + node isSend + and: [node selector == #cCode:inSmalltalk: + or: [aCodeGen isAssertSelector: node selector]]]]. - parseTree - nodesDo: - [:node| - (self inlineableFunctionCall: node in: aCodeGen) ifTrue: - [(self inlineFunctionCall: node in: aCodeGen) ifNotNil: - [:replacement| - sendsToInline at: node put: replacement]]] - unless: "Don't inline the arguments to asserts to keep the asserts readable" - [:node| - node isSend - and: [node selector == #cCode:inSmalltalk: - or: [aCodeGen isAssertSelector: node selector]]]. sendsToInline isEmpty ifTrue: [^false]. self replaceNodesIn: sendsToInline. ^true! Item was changed: ----- Method: TSendNode>>emitCCodeAsFieldReferenceOn:level:generator: (in category 'C code generation') ----- emitCCodeAsFieldReferenceOn: aStream level: level generator: aCodeGen "If appropriate, translate this message send as a pointer dereference" | parenCount | + (aCodeGen isStructSend: self) ifFalse: - (self isStructSendIn: aCodeGen) ifFalse: [^false]. parenCount := receiver isSend ifTrue: [2] ifFalse: [1]. aStream next: parenCount put: $(. receiver emitCCodeAsExpressionOn: aStream level: 0 generator: aCodeGen. parenCount > 1 ifTrue: [aStream nextPut: $)]. (receiver structTargetKindIn: aCodeGen) caseOf: { [#pointer] -> [aStream nextPut: $-; nextPut: $>]. [#struct] -> [aStream nextPut: $.] }. aStream nextPutAll: (aCodeGen cFunctionNameFor: selector). arguments isEmpty ifFalse: [self assert: arguments size = 1. aStream nextPutAll: ' = '. arguments first emitCCodeAsExpressionOn: aStream level: level generator: aCodeGen]. aStream nextPut: $). ^true! Item was removed: - ----- Method: TSendNode>>isStructSendIn: (in category 'testing') ----- - isStructSendIn: aCodeGen - "Answer if the recever is a send of a structure accessor. - This is tricky. We want - foo bar => foo->bar - foo bar => foo.bar - foo bar: expr => foo->bar = expr - foo bar: expr => foo.bar = expr - depending on whether foo is a struct or a pointer to a struct, - but only if both foo is a struct type and bar is a field accessor. - The tricky cases are self-sends within struct class methods. Here we need to - distinguish between self-sends of ordinary methods from self sends of accessors." - ^arguments size <= 1 - and: [(receiver structTargetKindIn: aCodeGen) notNil - and: [(aCodeGen methodNamed: selector) - ifNil: [false] - ifNotNil: [:method| method isStructAccessor]]]! Item was changed: ----- Method: VMBasicConstants class>>namesDefinedAtCompileTime (in category 'C translation') ----- namesDefinedAtCompileTime "Answer the set of names for variables that should be defined at compile time. Some of these get default values during simulation, and hence get defaulted in the various initializeMiscConstants methods. But that they have values should /not/ cause the code generator to do dead code elimination based on their default values." ^#( VMBIGENDIAN IMMUTABILITY STACKVM COGVM COGMTVM SPURVM PharoVM "Pharo vs Squeak" TerfVM "Terf vs Squeak" EnforceAccessControl "Newspeak" CheckRememberedInTrampoline "IMMUTABILITY" + LLDB "As of lldb-370.0.42 Swift-3.1, passing funciton parameters to printOopsSuchThat fails with Internal error [IRForTarget]: Couldn't rewrite one of the arguments of a function call. Turning off link time optimization with -fno-lto has no effect. hence we define some debugging functions as being <option: LLDB>" + WIN32)! - LLDB) "As of lldb-370.0.42 Swift-3.1, passing funciton parameters to printOopsSuchThat fails with Internal error [IRForTarget]: Couldn't rewrite one of the arguments of a function call. Turning off link time optimization with -fno-lto has no effect. hence we define some debugging functions as being <option: LLDB>"! Item was added: + ----- Method: VMPluginCodeGenerator>>isStructSend: (in category 'utilities') ----- + isStructSend: aTSendNode + "Answer if the argument aTSendNode is a send of a structure accessor. + This is tricky. We want + foo bar => foo->bar + foo bar => foo.bar + foo bar: expr => foo->bar = expr + foo bar: expr => foo.bar = expr + depending on whether foo is a struct or a pointer to a struct, + but only if both foo is a struct type and bar is a field accessor. + The tricky cases are self-sends within struct class methods. Here we need to + distinguish between self-sends of ordinary methods from self sends of accessors. + + Override to avoid requiring that there be a struct accessor method for the selector." + ^aTSendNode numArgs <= 1 + and: [(aTSendNode receiver structTargetKindIn: self) notNil]! Item was added: + ----- Method: VMPluginCodeGenerator>>structTargetKindForDeclaration: (in category 'C code generator') ----- + structTargetKindForDeclaration: decl "<String>" + ^(super structTargetKindForDeclaration: decl) ifNil: + [pluginClass ifNotNil: + [(pluginClass isStructType: (decl last = $* + ifTrue: [decl allButLast] + ifFalse: [decl]) withBlanksTrimmed) ifTrue: + [(decl indexOf: $*) > 0 + ifTrue: [#pointer] + ifFalse: [#struct]]]]!

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev September 2017