Vm-dev November 2018

vm-dev@lists.squeakfoundation.org

23 participants
82 discussions

[OpenSmalltalk/opensmalltalk-vm] Use Metal instead of OpenGL for the main VM Window in OS X (#306)
by Ronie Salgado 18 Feb '19

18 Feb '19

Because OpenGL is being deprecated by Apple, use Metal instead of OpenGL for main VM window in OS X. Metal only works in 64 bits mode. So, I also made the needed changes to keep using OpenGL in the 32 bits version of the OS X VM. You can view, comment on, or merge this pull request online at: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/306 -- Commit Summary -- * I ported the main OS X view from OpenGL into Metal. * Disable restoring OpenGL context hack in ioProcessEvents when using Metal for the main window. -- File Changes -- M build.macos32x86/common/Makefile.app (4) M build.macos32x86/common/Makefile.flags (2) M build.macos64x64/common/Makefile.app (8) M build.macos64x64/common/Makefile.flags (7) M build.macos64x64/common/Makefile.rules (10) M build.macos64x64/common/Makefile.vm (1) M platforms/iOS/vm/Common/Classes/sqSqueakEventsAPI.m (14) A platforms/iOS/vm/English.lproj/MainMenu-opengl.xib (1261) M platforms/iOS/vm/English.lproj/MainMenu.xib (8) M platforms/iOS/vm/OSX/SqViewBitmapConversion.m (5) M platforms/iOS/vm/OSX/SqViewClut.m (5) A platforms/iOS/vm/OSX/SqueakMainShaders.metal (73) M platforms/iOS/vm/OSX/SqueakOSXAppDelegate.m (5) A platforms/iOS/vm/OSX/sqSqueakOSXMetalView.h (93) A platforms/iOS/vm/OSX/sqSqueakOSXMetalView.m (772) -- Patch Links -- https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/306.patch https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/306.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/306

6 24

[OpenSmalltalk/opensmalltalk-vm] Pullrequest3/vmmaker (#309)
by George Ganea 18 Feb '19

18 Feb '19

Yet another pull request for merging vmmaker code into opensmalltalk-vm. Commits now have nicer looking comments as suggested in https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/308. You can view, comment on, or merge this pull request online at: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/309 -- Commit Summary -- * VMMaker.oscog-sk.2261.mcz Modify the profiling primitive cogCodeConstituents: to be able to differentiate the closedPICs from the openPICs in the profiling report * CogTools-sk.84.mcz * take into account the modified primitive collectCogCodeConstituents to display the differenciation open/closedPICs * remove unused method (VMFileSystem, which has been replaced) * VMMaker.oscog-eem.2262.mcz LargeIntegers plugin Fix the crash for 2009 nthRoot: 100000 due to digitDivLarge:with:negative: failing to check if allocations fail. The example produces 600k byet long integers and so provokes plenty of allocation failures. * VMMaker.oscog-eem.2263.mcz Fix simulation of Pharo primitiveDirectoryLookup above Squeak. * VMMaker.oscog-eem.2264.mcz Spur: The OmterpreterProxy/sqVirtualMachine.h API (e.g. in the ThreadedFFIPlugin) expects instantiateClass:indexableSize: to work with 0 indexable size for fixed classes. So remove the cppIf:... from the Spur implementation. * VMMaker.oscog-eem.2265.mcz Oops! Writing the function does nothing unless it is called. So once again, if ioScreenSize answers a zero screen extent (as is the case for headless images) then on snapshot write the sawvedWindowSize rather thna the zero extent. This should fix the experience people have with Pharo of saving headless images that then come up with a zero sized window when started headful. * CogTools-sk.85.mcz * UI for Pharo * add some comments * VMMaker.oscog-eem.2266.mcz Spur: Simplify cleverSwapHeaders:and:copyHashFlag: via xor. * VMMaker.oscog-eem.2267.mcz Slang: Allow plugins to use sends to access struct members. * VMMaker.oscog-tpr.2268.mcz First pass at removing obsolete window color setting and move to the new theme framework. * VMMaker.oscog-cb.2269.mcz Normally this commit has no impact on the production VM (some refactorings were made to share code between different part of the GC, but no new things). I tried to simulate the whole VM and generate a VM and it works, but only Travis can confirm us everything's ok. * VMMaker.oscog-cb.2270.mcz SpurSweeper is working in the simulator, effectively changing Spur's fullGC to a mark-sweep algorithm when enabled. Still need to check: * VMMaker.oscog-cb.2271.mcz Since this commit SpurSweeper seems to be production-ready. I compiled a VM with SpurSweeper and it worked just fine, handling many GCs. * VMMaker.oscog-cb.2272.mcz Fixed bug with large struct in FFI with SysV . * ImageFormat-dtl.27.mcz Sort the format numbers for more convenient display * VMMaker.oscog-cb.2273.mcz Added primitiveStringReplace in the JIT with Spur (was showing to Sophie how to write code in the JIT, so we did that together). We implemented only the quick paths for byte objects and array-like objects. * VMMaker.oscog-cb.2274.mcz Fix the growth of the remembered table during GC when there is not enough memory in old space. Allocates an extra memory segment in that case. This is very uncommon. * ImageFormat-dtl.28.mcz Fix typo, the hypothetical 68003 is a known version number, 68004 is not. * VMMaker.oscog-eem.2275.mcz Provide a logging facility that uses the Printf package. See VMClass>>log:with:* for API. See Slang changes below for support. * VMMaker.oscog-eem.2276.mcz Simplify flushLog in the logging framework by providing fflush on WriteStream. * VMMaker.oscog-eem.2277.mcz Oops. Revert an inadvertent change to flushLog. * VMMaker.oscog-eem.2278.mcz Spur: Revise scavenger logging to collect data during the scavenge and log it to the file after the scavenge, hence including an accurate elapsed time. * VMMaker.oscog-eem.2280.mcz VMMaker: Fix bad misinitialization bug where running VMMaker generation routines in 64-bits would cause 64-but sources to be generated. This caused the recent misgeneration of src/vm/cogit.c. * VMMaker.oscog- nice.2281.mcz Fix type inference of variable unskew in BitBlt >> copyLoop * VMMaker.oscog-cb.2282.mcz fix primStringReplace in the JIT. * VMMaker.oscog-nice.2283.mcz Fix incorrect size of replArray in genPrimitiveStringReplace * VMMaker.oscog-eem.2284.mcz BitBltSimulation>>copyLoop: Strength reduce the loop split for combinationRule = 3 so that less tests are done in the inner loop and so trhat when combinstionRule = 3 mergeFn is never used. Use an explicit type declaration for unskew instead of trickier relying on type inference. Eliminate bogus use of #== for numeric comparison. * VMMaker.oscog-eem.2285.mcz BitBltSimulation>>copyLoop: No changes other than reformatting indentation for readability. * VMMaker.oscog-eem.2286.mcz Simulation on 64-bit Squeak. * VMMaker.oscog-eem.2287.mcz Spur: Rewrite determining the accessorDepth for the meta primitives (primitiveDoPrimitiveWithArgs and primitiveDoNamedPrimitiveWithArgs, 118 & 218). Have primitiveDoNamedPrimitiveWithArgs fetch the accessorDepth. Discard the confused setting of newMethod to the prim index. Introduce metaAccessorDepth and eliminate tempOop3. This fixes a crash when stepping through something likely to fail such as [DoubleWordArray new: 64*1024*1024] repeat. * VMMaker.oscog-eem.2288.mcz Refactor reaping the primFailCode on method activation to pull it out of the common path and hence lay the ground for a platform error object since all failures now call a single reapAndResetErrorCodeTo:header: * VMMaker.oscog-eem.2289.mcz Add support for conveying os/library error codes through a special primitive failure code object. The error code is a signed 64-bit value (so that e.g. answering -1 is not expensive). Specific image-level clients can map that value to unsigned as appropriate. * VMMaker.oscog-eem.2290.mcz Slang: Include short int as a valid integral type. * VMMaker.oscog-eem.2291.mcz CoInterpreter: Fix bug in new ceReapAndResetErrorCodeFor:; need to write the error code to the word above stack top, stack top holding the return address. * VMMaker.oscog-eem.2292.mcz Interpreter: Eliminate some uses of #== to compare integral values. * VMMaker.oscog-eem.2293.mcz Fix vorgotten variables in the unused variable elimination scheme for inlined value:[value:*] and to:by:do: * VMMaker.oscog-eem.2294.mcz Spur Image Segments: Fix objectsReachableFromRoots: which was computing a transitive closure that could include duplicates. * VMMaker.oscog-eem.2295.mcz Spur Image Segments: The short cut in ensureNoNewObjectsIn: was wrong. Best not take any short cuts here at all. * VMMaker.oscog-eem.2296.mcz Fix bugs in DoubleWordArray>>[unsigned]long64At: which failed to exclude extraneous bytes in the upper word of unaligned accesses. These along with the fix to hackBits: in Graphics-eem.389 fix simulating 64-bit images in 64-bits, ensuring that the display is visible. * VMMaker.oscog-eem.2297.mcz Cogit: Extend Spur 64-bit primitiveAt with full support for 64-bit longs. * VMMaker.oscog-eem.2298.mcz Cogit: More work to full 64-bit access in at:[put:] primitives. Fix slip in x64 setsConditionCodesFor:. * VMMaker.oscog-eem.2299.mcz Get the StackInterpreterSimulator to a state where it can correctly simulate the DoubleByteArray, WordArray, DoubleWordArray and MemoryTests. Principally allow the SqueakFFIPrims plugin (ThreadedFFIPlugin) to load and make primitiveFFIIntegerAt[Put] function correctly in simulation. Change the two primitives to use unalignedShortAt:[put:], unalignedLong32At:[put:] & unalignedLong64At:[put:] and implement these in SpurMemoryManager (ObjectMemory can wait) and have the preambleCCode map these to the original shortAt[put], long32At[put] & long32At[put] C functions/macros. * VMMaker.oscog-eem.2300.mcz Change the contract of ThreadedFFIPlugin>>ffiAddressOf:startingAt:size:, moving the primitive fails to the callers for equivalent but more compact code. * VMMaker.oscog-eem.2301.mcz Fix a slip in CogObjectRepresentationFor64BitSpur>>genPrimitiveAt's 64-bit indexable suypport. The range check must include the sign bit since 64-bit at: answers an unsigned value. The Cogit now correctly passes the DoubleByteArrayTests, WordArrayTests, DoubleWordArrayTests and MemoryTests in simulation and actuality. * VMMaker.oscog-eem.2302.mcz Fix a slip in the new ThreadedFFIPlugin preambleCCode (the atput's need an extra parameter). * VMMaker.oscog-eem.2303.mcz Simulation/Translation tweaks. Mark some simulation-only InterpreterPlugin methods as doNotGenerate. Slow down the simulated clock on the StackInterpreter (so that in simulation fewer tests time out). Provide an optional simulation-only primTraceLog for the StackInterpreter (which was used to debug the new 64-bit at:[put:] support). * VMMaker.oscog-eem.2304.mcz Add a check for the knownClassIndex actually binding to a class in the relevant eeInstantiate routines. * VMMaker.oscog-eem.2305.mcz Slang for Plugins Make sure the comparison of types of the reference implementation and the declarative implementation in InterpreterProxy is made once types have been assigned. * VMMaker.oscog-eem.2306.mcz Slang: Make sure that numeric constant methods inlined through the tryToInlineMethodExpressionsIn: path get commented with their selector. * VMMaker.oscog-eem.2307.mcz Refactor preDeclareInterpreterProxyOn: to extract collecting the InterpreterProxy interface to its own method to simplify consistency checking. To make the checking more correct use a concrete Sour class for the referenceObjectMemoryClass. * VMMaker.oscog-eem.2309.mcz Tweak the recent inlining commenting change; don't bother to comment if what's being inlined is a named constant. * VMMaker.oscog-eem.2310.mcz Simulation: Fix charsOf: so that longPrintOop:/shortPrint: prints the full 8-bit character range, not the 7 bit character range. * VMMaker.oscog-dtl.2311.mcz Fix typo in function declaration in ImmX11Plugin, identified by Stuart Cassoff. * VMMaker.oscog-eem.2312.mcz Fix two places where cloning forgets to set the immutability bit if the input has it set (shallowCopy should /not/ copy across the immutability bit, but become: and pin: should). * BytecodeSets.spur-eem.76.mcz Save BytecodeeSets now that both Pharo and Squeak have the SistaV1 bytecode set in the base. * BytecodeSets.spur-eem.77.mcz Remove SistaV1 category (its now empty here). Better temp names in printing method. Fix typo. * VMMaker.oscog-eem.2313.mcz Cogit: Fix bad bug in ceSend:above:to:numArgs: which was written to accept an association, not a class. The rewrite agrees with the interpreter's definition. * VMMaker.oscog-eem.2314.mcz Hack fix comment generation. The API for comments in TParseNode is broken because in some subclasses it is a sequence of Strings, and elsewhere is a single String. * VMMaker.oscog-eem.2315.mcz Double-back on the rash statement that there was a bad bug in ceSend:above:to:numArgs: which was written to accept an association, not a class. This was by design. generalize the JIT support for directed super sends so that it accepts both the pushLiteral: foo; directedSuperSend: bar numArgs: baz, and * VMMaker.oscog-eem.2316.mcz Interpreter: Fix bad bug in reverseDisplayFrom:to: feedback from the leak checker. The displayBits are not uopdated soon enough after a compaction and objects may be overwritten. So refactor postGCAction: to extract postGCUpdateDisplayBits which is also used by reverseDisplayFrom:to: to obtain up-to-date bits. * VMMaker.oscog-eem.2317.mcz Simulation: Better SocketPlugin name lookup simulation. * VMMaker.oscog-eem.2318.mcz Phhhhh.... * VMMaker.oscog-eem.2319.mcz Simulation: Close files properly in the FilePliginSimulator and recycle file handle indices. * VMMaker.oscog-eem.2320.mcz Spur: Fix a bad bug in SpurPlnningCompactor. unmarkObjectsFromFirstFreeObject, used when the compactor requires more than one pass due to insufficient savedFirstFieldsSpace, expects the corpse of a moved object to be unmarked, but copyAndUnmarkObject:to:bytes:firstField: only unmarked the target. Unmarking the corpse before the copy unmarks both. This fixes a crash with ReleaseBuilder class>>saveAsNewRelease when non-use of cacheDuring: creates lots of files, enough to push the system into the multi-pass regime. * VMMaker.oscog-eem.2321.mcz Implement following of pushConstant: aBoolean; jumpTo: target; ...target: jump: target2 if: cond in genJumpTo:. * VMMaker.oscog-eem.2322.mcz RegisterAllocatingCogit: Make sure the receiverResultReg state is valid on backwards jump when the stackLimit is not exceeded. Implement branch following as per StackToRegisterMappingCogit>>#genJumpTo:. Avoid duplicating send of moveVolatileSimStackEntriesToRegisters in ensureFixupAt:. * Cog-eem.345.mcz Add the processor aliens when compiled in 64-bits for running on 64-bit Squeak * Cog-eem.346.mcz Add the missing control register accessiors for GdbARMAlien64 and delete subclass accessors that are the same as the superclass's. * VMMaker.oscog-eem.2324.mcz Fix store check call in genPrimitiveStringReplace on RISCs (i.e. on ARM save & restore LinkReg around call). * VMMaker.oscog-cb.2323.mcz Based on discussions decreased from 7 to 5 the number of instructions on Intel in primStringReplace copying loops * VMMaker.oscog-eem.2325.mcz StackToRegisterMappingCogit: Fix ancient bug in voidReceiverResultRegContainsSelf. The computed index is the right argument for ssFlushTo:; it shouldn't be subtracted from simStackPtr! * VMMaker.oscog-eem.2326.mcz StackToRegisterMappingCogit: For simplicity, especially in the RegisterAllocatingCogit subclass, eliminate simSelf and optStatus and include the receiver's simStackEntry as the 0'th element of the simStack. optStatus is therefore whether simSelf's liveRegister is ReceiverResultReg. hence loops from 0 to: methodOrBlockNumTemps - 1 become loops from 1 to: methodOrBlockNumTemps. Teh simSelf variable is replaced by self simSelf, which is short-hand for self simStackAt: 0. * VMMaker.oscog-eem.2327.mcz StackToRegisterMappingCogit: Get much stricter with simSpillBase for the benefit of RegisterAllocatingCogit (and eventually SistaCogit). * VMMaker.oscog-eem.2328.mcz StackToRegisterMappingCogit: Replace the UnknownSimStackPtrFlag hack in a fixup's simStackPtr and replace it by a proper flag. Consequently make a fixup's simStackPtr unsigned char, giving us a maximim of 255 stack slots (0 being for the receiver). * VMMaker.oscog-eem.2329.mcz RegisterAllocatingCogit: Revise merge code generation (& uke the now unneeded failed merge catcher). Have the merge update simStack reflecting reality as the merge proceeds. This update occurs in reconcilePoppingWith: & reconcilePushingWith:. N.B. reconcilePoppingWith: & reconcilePushingWith: are now essentially identical apart from whether they pusgh or pop. They will be merged and the push/pop moved to the client mergeCurrentSimStackWith: very soon. * VMMaker.oscog-eem.2330.mcz RegisterAllocatingCogit: Refactor reconcilePushingWith: & reconcilePoppingWith: into reconcileWith:spillOffset:onSpillOrUnspill:. * VMMaker.oscog-eem.2331.mcz RegisterAllocatingCogit: Refactor to allow RegisterAllocatingCogit to deal with the push initial nils hack, i.e. move the two pass loop mechanism out of compileAbstractInstructionsFrom:through: into compileMethodBody & compileBlockBodies. * VMMaker.oscog-eem.2332.mcz RegisterAllocatingCogit: Moving volatile stack contents to registers should happen once before each jump, not potentially multiple times in ensureFixupAt:. * VMMaker.oscog-eem.2333.mcz RegisterAllocatingCogit: Fix mistakes in reconcileWith:spillOffset:onSpillOrUnspill: when merging with a volatile entry that has a register. The current entry must morph into an SSRegister, /not/ an SSSpill, and it must answer that it has merged a register. mergeCurrentSimStackWith: must deassign the target's reguster, /not/ current's. And it must actually update current's spill status on (un)spill. * VMMaker.oscog-eem.2334.mcz Expand the set of names defined at compile time (e.g. for the option: pragma). Perhaps we should invert this and define the closed set of names defined at translation time. * VMMaker.oscog-eem.2335.mcz RegisterAllocatingCogit: Improve (dare I say, fix??) the register permutation engine in the control flow merge logic. Fix reconciling entries that oproduce an SSRegister entry (the liveRegister must be set to NoReg). Fux recinciling a constanty with itself (this happens at the end of ifTrue:ifFalse: given that the bytecode compiler preserves the value of ifTrue:ifFalse: until a single trailing pop). N.B. the merge engine could save effort by peeking ahead for a pop, but Scorch likely won't produyce such bad code so we shouldn't waste effort making the RegisterAllocatingCogit generate better code for non-Scorch input. * VMMaker.oscog-eem.2336.mcz RegisterAllocatingCogit: Fix the visited register filter in the register swapper in the merge engine. * VMMaker.oscog-eem.2337.mcz Rewrite the primitives in MiscPrimitivePlugin using conventional Slang, avoiding the translatedPrimitives ineffiicencies and dependence on methods in the image. Volunteers are invited to do the same for the ADPCMCodecPlugin and SoundGenerationPlugin. * VMMaker.oscog-eem.2338.mcz Fix a slip in primitiveDecompressFromByteArray caught by IncludedMethodsTests. Thank you David! * VMMaker.oscog-eem.2339.mcz Immutability: Fix primitiveSlotAtPut (& InterpreterPrimitives>>primitiveInstVarAtPut) to fail with PrimErrorNoModification for read-only objects. * VMMaker.oscog-akg.2340.mcz Extend FilePlugin to allow a file to be opened using either the file descriptor (fd) or FILE* in Pharo. * VMMaker.oscog-akg.2341.mcz Include FilePlugin>>primitiveFileOpenUseFileDescriptor & primitiveFileOpenUseFile on all platforms (not just PharoVM). * VMMaker.oscog-topa.2342.mcz The config should always come first. * VMMaker.oscog-eem.2343.mcz Better comment Tobias' inclusion of config.h in all plugins. Also ensure (almost) no duplications can occur. * VMMaker.oscog-eem.2344.mcz Cogit: Fix the macro definition for simSelf (it must take arguments). * VMMaker.oscog-eem.2345.mcz Cogit: Fix several slips. Inline a few trivial methods. * VMMaker.oscog-akg.2346.mcz FilePlugin connect to file primitives * VMMaker.oscog-eem.2347.mcz Review of Alistair's recent FilePlugin changes: * VMMaker.oscog-cb.2348.mcz This commit is partially merged, I can only merge from a more recent Squeak image to get the DoubleWordArray/DoubleByteArray things correct. Also there's a conflict in SistaV1 bytecode table in StackToRegMappingCogit I need to figure out. I'll finish merging afterwards. * VMMaker.oscog-cb.2349.mcz Merge back the DoubleWordArray code from a fresh image. Normally everything should be merged, up and running. * VMMaker.oscog-cb.2350.mcz Decreases by 1 the number of instructions for byte reads on constants. * VMMaker.oscog-cb.2351.mcz Fogrot to merge back Eliot's change in the bytecode table with mine * VMMaker.oscog-eem.2352.mcz Cogits: Fix slip in ssFlushTo:. * VMMaker.oscog-eem.2353.mcz Cogits: Fix pc mapping for genStore[AndPop]RemoteTempLongBytecode. Neither of these are isMappedIfImmutability unless they are used for inst var assignment of arbitrary objects; in which case we need to add an annotation when they're used for indirect temp access. * VMMaker.oscog-ul.2354.mcz Restored SerialPlugin's *byName* primitives on non-PharoVM VMs. * VMMaker.oscog-akg.2355.mcz FilePlugin status checks and type declaration * BytecodeSets.spur-eem.78.mcz Implement multiple bytecode set aware scanning machinery for NewsqueakV4. * VMMaker.oscog-cb.2356.mcz fixed byteAt constant removed unused annotation * VMMaker.oscog-eem.2357.mcz Fix regression introduced in VMMaker.oscog-eem.2333 or thereabouts when improving comoilation breakpoint. maybeSelectorOfMethod can answer nil so a guard is needed. * VMMaker.oscog-eem.2358.mcz Fix several (ancient) issues with the MiscPrimitivePlugin primitives, identified by Levente. Try and make all failures supply an error code. * VMMaker.oscog-eem.2359.mcz Oops! Fixed one I missed. * VMMaker.oscog-eem.2360.mcz Plugin: Deprecate the unused in-argument-order primitive argument accessors floatArg:, integerArg:, methodArg:, & objectArg:. These have never been used. Replace them with more convenient variants of methodReturnValue:, namely methodReturnBool:, methodReturnFloat:, methodReturnInteger:, and methodReturnString:. Use them in a few primitives. * VMMaker.oscog-eem.2361.mcz Spur: Make read-only (-DIMMUTABILITY) the default wth Spur VMs. * VMMaker.oscog-eem.2362.mcz For Spur MT interpreters to be generated by generateAllConfigurationsUnderVersionControl they must be in generateAllSqueakConfigurationsUnderVersionControl. * VMMaker.oscog-eem.2363.mcz Fix regression in generateInterpreterProxyFunctionDereference:on:indent: introduced in VMMaker.oscog-eem.2361. * ImageFormat-dtl.29.mcz Fix by K K Subbu: Use memcmp instead of strncmp in ckformat to compare byte arrays. * ImageFormat-dtl.30.mcz Fix declaration of main() in ckformat.c * VMMaker.oscog-CyrilFerlicot.2364.mcz Add a progress bar when generating multiple vm code with VMMaker. * VMMaker.oscog-VB.2364.mcz - Add new primitive: primitiveHostWindowIcon -> able to change the main window icon at runtime * ImageFormat-dtl.31.mcz Fix ckformat for 64 bit Spur, which saves format number in first 4 bytes of the header, versus 8 bytes for 64 bit V3. Prior version worked by accident for little endian host using strncmp() check, but failed when correct memcmp() was used without limiting check to 4 bytes for spur and 8 for V3. * VMMaker.oscog-VB.2365.mcz merge VB 2364 and CyrilFerlicot 2364 * VMMaker.oscog- nice.2366.mcz Correct a 32bit-hardcoded pointer size in FFI Correct two copy/paste typos in num32BitUnitsOf: * VMMaker.oscog-sk.2367.mcz ** new primitive to compare strings (slang + JIT) answers negative smi, 0 or positive smi (instead of 1, 2 or 3 in the MiscPlugin) * VMMaker.oscog-eem.2368.mcz Check collation order byte array size in new string compare primitive as per Levente's suggestion. * VMMaker.oscog-VB.2369.mcz add primitive primitiveStdioDescriptorIsATTY See https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/254 * VMMaker.oscog-VB.2370.mcz primitiveStdioDescriptorIsATTY will return a bool instead of an int * VMMaker.oscog-VB.2371.mcz Add arguments in primitiveIsFileDescriptorATTY * VMMaker.oscog-cb.2372.mcz Fixed a bug in inlined allocation (or so I think... Did not check the tests) * VMMaker.oscog-cb.2373.mcz Remove the APIs I added to iterate over free chunks (there was an existing API) * VMMaker.oscog-cb.2374.mcz Improved comments & rename variables for Slang to C compilation not to complain (Shadowing globals with inlining cross-classes is a little bit tricky, let's not take any risks) * VMMaker.oscog-cb.2375.mcz Made C compiler happy (addressOf:, macro & shadowing non-sense).. * VMMaker.oscog-cb.2376.mcz Finally figured out that Slang is not able to do look-ups in compactor classes for some reason, hence only planning compactor was the only one compiling correctly since it defined everything inside... * VMMaker.oscog-cb.2377.mcz Spur ancilliary classes includes now: (self compactorClass withAllSuperclasses copyUpThrough: SpurCompactor) reverse * VMMaker.oscog-cb.2378.mcz Improved compactor comments. * VMMaker.oscog-cb.2379.mcz Started implementation of Tracking. Refactor a bit more to share code between selective and tracking. * VMMaker.oscog-eem.2380.mcz Fix compiler bug with Apple LLVM version 7.0.0 (clang-700.1.76) for 64-bit Spur segment loading where compiler bug eliminated second version check in segment load when at -Os. Fix is to never inline the 32-bit word byte reversal. * VMMaker.oscog-eem.2381.mcz Fix a bug in the refactored conditional define code; a slip caused premature evaluation of the block argument. * VMMaker.oscog-VB.2382.mcz change the primitive to return the file descriptor type instead of just 0 or 1 * VMMaker.oscog-VB.2383.mcz Use methodReturn***: instead of pop: + push***: in FilePlugin * VMMaker.oscog-cb.2384.mcz Compactor clean-ups, we now only have: - Pig: first spur compactor * VMMaker.oscog-topa.2385.mcz AnObsolete slipped in, ask it to leave. * VMMaker.oscog-topa.2386.mcz Trust config.h more. It declares when to include dlfcn and takes care of _GNU_SOURCE. * VMMaker.oscog-ul.2387.mcz General: - added #methodReturnReceiver, another methodReturn* method, that pops argumentCount elements from the stack, leaving only the receiver there * VMMaker.oscog-ul.2388.mcz - fixed a bug in primitiveStringHash * VMMaker.oscog-cb.2389.mcz Fixed a bug in frameless full blocks (fetching receiver from receiver index in FullBlock and not outerContext) * VMMaker.oscog-cb.2390.mcz Fixed consistency checks in VMClass to succeed if no compactorClass is present. * VMMaker.oscog-eem.2393.mcz BitBltPlugin: Don't fail primitiveDisplayString for empty strings (still validates fully though). * VMMaker.oscog-eem.2394.mcz SpurSelectiveCompactor: Add a guard to followForwardingPointersInStackZone: to avoid scanning teh stack zone when it is empty because the system is snapshotting. * VMMaker.oscog-cb.2391.mcz Fixed a bug in Selective Compactor where remembered objects were incorrectly copied. * VMMaker.oscog-cb.2392.mcz Removed debugging code to make VM compilation work again. * VMMaker.oscog-eem.2395.mcz Fix regression in primitiveDisplayString (missed return on short-cut return for empty string). * VMMaker.oscog-eem.2396.mcz Fix some compiler warnings in the interpreter (including unused variables). Make methodReturnString: safe in the presence of allocation failures, but note that, because it may fail on returning a result (which implies the primitive has done its work) this is dubious at best. * VMMaker.oscog-eem.2397.mcz Actually fix Slang not emitting variable declarations for 'extern ...' declarations. * VMMaker.oscog-eem.2398.mcz Have Slang serve is rather than we work-around Slang, in as much as Slang now outputs any ext5ern ... declaration, so we no longer need a fake temp var with associated crap to keep the compiler happy. * VMMaker.oscog-eem.2399.mcz FFIPlugin: Fix slip in the Win64 ffiCalloutTo:SpecOnStack:in: * VMMaker.oscog-eem.2400.mcz Fix Sista build which is failing due to the lack of an export of objectBytesForSlots:. * CogTools-eem.86.mcz Fix profile buffer for 64-bits (use DoubleWordAray). So for symmetry change to using WordArray in 32-bits instead of Bitmap. Hack fix MNU cases in PICs in analyzeClosedPIC:. This is a hack because the VM is answering useless values. It simply answers #doesNotUnderstand:, whereas it could more usefully answer {theClass. #doesNotUnderstand:} or simply {theClass}. In fact, all the PIC cases should supply what the actual receiver class is, not simply the target method. We have to discuss what to do. This is in Cogit>>#createCPICData: * CogTools-eem.87.mcz Collect the new format PIC data. * VMMaker.oscog-eem.2401.mcz Cogit: Answer better closed PIC data from primitiveCollectCogCodeConstituents. i.e. scan preceding methods for references to closed PICs and store temporarily the send site's first case cache tag (mapped to a class) in the PIC's methodObject field for later harvesting. Answder am array of PIC selector followed by class, target pairs, where target is either a method or #doesNotUnderstand:. * CogTools-eem.88.mcz Eliminate obsolete primitive * VMMaker.oscog-eem.2402.mcz BitBltPlugin: The new primitiveDisplayString short cut has exposed an old bug: both loadBitBltDestForm & loadBitBltSourceForm sent byteSizeOf: to the form bits before checking if form bits isWordsOrBytes:. Also simplify loadHalftoneForm which comntained a superfluous (interpreterProxy isPointers: halftoneForm) not. * VMMaker.oscog-eem.2403.mcz BitBltPlugin: Move the checks for the dest and source forms being sufficiently large pointer objects into their respective validation routines. * VMMaker.oscog-eem.2404.mcz Spur Compaction & Slang Add support for static resolution of very limited polymorphism, specifically that required to select between the SpurSelectiveCompactor and the SpurPlanningCompactor at run-time, using SpurSelectiveCompactor for normal GC and SpurPlanningCompactor for snapshots. * VMMaker.oscog-eem.2408.mcz Initialization: Nuke the insane initializationOptions inst var in VMClass class and replace it by a simple class variable InitializationOptions. This avoids stale initializationOptions dictionaries during initialization. The exact symptom was the validation of options in VMClass class>>initializeMiscConstants; it wanted to check SpurmemoryManager's compactorClass, but since SpurMemoryManager's initialationOptions was stale it answered a bogus class. Grasshopper, enlightenment comes slowly to some. * VMMaker.oscog-eem.2409.mcz Fix static selector mapping in SpurCompactor; it needs to handle inheritance (SpurSelectiveCompactor inherits from SpurSweeper). * VMMaker.oscog-cb.2405.mcz - changed the free list representation from linked list to double linked list to unlink efficiently the free chunk in compaction phases. - Fix some simulation details * CogTools-eem.89.mcz Eliminate a deprecation warning on Squeak. * VMMaker.oscog-cb.2406.mcz Added bytesBigEnoughForPrevPointer: abstraction and patch all callers to use that. * VMMaker.oscog-cb.2407.mcz Fixing cases where double linked list prev pointer was not correctly updated. * VMMaker.oscog-cb.2410.mcz - Added assertions for free chunks like crazy. - Fixed the double linked list scheme (now everything resists the leak checker in stress GC benchs in 64 bits, previously it worked only in 32 bits) * VMMaker.oscog-cb.2411.mcz Improved assertion (we want to catch this case) * VMMaker.oscog-cb.2412.mcz - Fixed compilation errors and some warnings. * VMMaker.oscog-cb.2413.mcz Fixed a bug when new segment was allocated in C sometimes the swizzle field used by SelectiveCompactor got corrupted. Changed assertions here and there. * VMMaker.oscog-eem.2414.mcz Update some class comments. * VMMaker.oscog-eem.2415.mcz Having a VMMaker-specific decompiler test and decompiler test failure collecrtor helps checking for compiler issues (e.g. just discovered one of my images has a mis-compiled SpurCompactor hierarchy). * VMMaker.oscog-eem.2416.mcz More initializationOptions clean-ups. * VMMaker.oscog-cb.2417.mcz Reworked a bit inlining for profiling. * VMMaker.oscog-topa.2418.mcz An include got missing, put it back where it belongs * VMMaker.oscog-cb.2419.mcz Added stats for Marking time and sweep time in full GC. Make them available as vm parameters 72 and 73. * VMMaker.oscog-eem.2420.mcz Spur: Extend semantics of primitives 165 & 166 (integerAt:[put:], signed indexing of 32-bit word classes) to byte, short and double word classes; actually, all classes; this is a signedversion of 60 & 61 basicAt:[put:]. * VMMaker.oscog-eem.2421.mcz Cogits. Fix bad bug in pc mapping of methods containing primitives. methodUsesPrimitiveErrorCode:header: used the initialPC inst var, rather than deriving the initialPC of the method parameter. initialPC is stale/over ripe, being that of the last JITTED method. So many methods containing primitives get their PC mapping wrong. We haven't noticed this much because primiitves typically succeed. Nuke an obsolete version of the method. * VMMaker.oscog-eem.2422.mcz Cogits. Revamp pc mapping tests so they can apply to methods in the simulation (used to debug the issue fixed in VMMaker.oscog-eem.2421). * VMMaker.oscog-eem.2423.mcz RegisterAllocatingCogit. ensureReceiverResultRegContainsSelf should use a simple super send, since StackToRegisterMapping's version merely flushes the stack. This results in poor code when assiging a method result to an inst var; the result comes back in ReceiverResultReg, which must be copied to another register to allow ReceiverResultReg to be loaded with self for the store. * VMMaker.oscog-eem.2424.mcz SistaRegisterAllocatingCogit. Noise. Commit an updated SistaCogitClone (which is auto generated on startup). Fix the comment in syncCodeWithSistaCogit. * VMMaker.oscog-cb.2425.mcz New things can be forwarders with selective compactor. * VMMaker.oscog-AlistairGrant.2426.mcz 274: primitiveFileStdioHandles() fails to return nil if stdio file is not available * VMMaker.oscog-eem.2428.mcz Work around some Squeak specificities. Update the VMMakerDecompilerTests decompilerFailures * VMMakerCompatibilityForPharo6-eem.1.mcz Compatibility methods and classes for VMMaker when loaded into Pharo6. Right now gets around EndianDetector and the lack of MethodReference. * VMMaker.oscog-eem.2429.mcz Oops; there were two... * VMMaker.oscog-AlistairGrant.2427.mcz 274: Update primitiveFileStdioHandles error handling. * VMMakerCompatibilityForPharo6-eem.2.mcz isCompiledBlock and bytesPerElement * VMMakerCompatibilityForPharo6-EliotMiranda.3.mcz Support for embeddedBlockClosures and schematicTempNamesString because these are used currently to interface to the JIT's code decoration facilities. * VMMaker.oscog-eem.2430.mcz Slang: Send asTranslationMethodOfClass: to CompiledMethods dirctly, allowing CompiledMehtod to choose between Smalltalk-80 parse trees and RefactoringBrowser parse trees. * VMMakerCompatibilityForPharo6-eem.4.mcz Stream compatibility methods for ThreadSafeTranscript which I shouldmn;'t have to implement (it's supposed to be a WriteStream). Initial support for Slang TMehtod creation from RBProgramNodes. * VMMakerCompatibilityForPharo6-eem.5.mcz Add isNodeNil to RBProgramNode & subclass * VMMaker.oscog-eem.2431.mcz Core VM: Kernel of support for failing FFI calls that raise exceptions. This provides simulation support for catching the signal and activating the invoking method as a primitive failure. The platform exception handlers still need of course to be modified to invoke this support. * VMMaker.oscog-eem.2432.mcz Finish the VMMaker side of the failing of FFI callouts on exceptions. * VMMaker.oscog-eem.2433.mcz Fix a slip in primitiveFailForFFIException:at:. If it fails the primitive, short cutting the return side of an FFI call it is its responsibility to own the VM and clear inFFIFlags. * VMMaker.oscog-eem.2434.mcz Add the missing implementation in InterpreterProxy, and correct the comment of primitiveFailForFFIException:at:. Also activateFailingPrimitiveMethod is not an api method; primitiveFailForFFIException:at: is. * VMMaker.oscog-eem.2435.mcz Work around the hack use of (dis)ownVM: for storing the argumentCount of the current method on callback return in the non-threaded VM. The argument to disownVM: is either an immediate integer (callback) or flags which can't be immediate (FFI callout, plugin primitive callout). I regret this but it works, so keep it going for now. * VMMaker.oscog-eem.2436.mcz And bring the simulation only bridge methods up to date in NewObjectMemory, reducing duplication. * VMMaker.oscog-eem.2437.mcz StackInterpreter: Fix signature of reapAndResetErrorCodeTo:header: to leiminate warnings. * VMMaker.oscog-eem.2438.mcz General robustness, compatibility and cleanups. * VMMaker.oscog-eem.2439.mcz Slang: more Pharo/Squeak compatibiltiy refactoring. * VMMakerCompatibilityForPharo6-eem.6.mcz Add translation support to RBProgramNode such that simple methods such as primitiveMakePoint can be translated to C. Still work needed to run the workspace translation scripts, which translate everything to be able to run inlining before generating C for a single method. * VMMakerCompatibilityForPharo6-eem.7.mcz Add LongTestCase & access for the return of a quick method to RBMethodNode needed for cmacro: methods. * VMMakerCompatibilityForPharo6-eem.8.mcz Add translation support for brace constructs in case statements. And include a *Tools extension for TranscriptStream. * VMMaker.oscog-eem.2440.mcz More Pharo compatibility. * VMMaker.oscog-eem.2441.mcz Make sure is:KindOf: has a valid return type. * VMMaker.oscog-eem.2442.mcz Final piece of Pharo compatibility to allow e.g. the REPL startreader images to run in Pharo 6.1. Split the input dialog for FakeStdinStream so it works both on Pharo and Squeak. * VMMaker.oscog-eem.2443.mcz Oops; remember to use isFakeStdinStream, and avoid isKindOf: in a couple of other places. * VMMaker.oscog-eem.2444.mcz Eliminate cCode: usage in the B3DAcceleratorPlugin and HostWindowPlugin usign the new "var args" style. Rewrite mem:mo:ve: et al in the new style. * VMMakerCompatibilityForPharo6-eem.9.mcz More compatbility. InstructionPrinter et al are missing in Pharo6 and subsequent. RBProgramNode methods for Slang. GtTranscript compatibility. * VMMaker.oscog-cb.2445.mcz Change the objStack logic to mmap a new memory segment on overflow if not enough room on heap when allocating a new page. * VMMaker.oscog-cb.2446.mcz Split the VMParameter primitive in 3 methods in Slang. I did it because of jump false size overflow on V3PlusClosure BC set, but it also look nicer. * VMMaker.oscog-cb.2447.mcz Production VM: - Fixed a bug in removeSegment: where invalid memory would be accessed when the segment info array was full & a segment removed. * VMMaker.oscog-cb.2448.mcz Introduced the concept of lilliputian free chunks, which are free chunks not big enough to use the double linked list design. They do not exist in 32 bits but only in 64bits. * VMMaker.oscog-cb.2449.mcz Change SelectiveCompactor to annotate segments during sweep phase with the last lilliputian chunk they hold. When compacting specific segments, this information is re-used to jump in the lilliputian single linked list to the free chunk right before hence elements can be cheaply removed from the linked list without iterating over it. * VMMaker.oscog-cb.2450.mcz Fix a bug in lilliputian management in selectivecompactor when adding a segment to compact into while there is none available. * VMMaker.oscog-eem.2451.mcz Spur: Tweak followClassTable toi not waste effort following hiddenRootsObj. FilePlugin: avoid createDirectory: hack. * VMMaker.oscog-eem.2452.mcz General: Fix error code reporting for primitiveAdoptInstance. It uses changeClassOf:to: which expects the class to be the argument, so BadArgument/BadReceiver errors must be swapped. * VMMaker.oscog-eem.2453.mcz Cog Spur: Eliminate unnecessary branch in 32 bit version of genPrimitiveAtSigned:, to match the 64 bit version. * VMMaker.oscog-eem.2454.mcz General: Correct a slip in primitiveVMParameter; statMaxAllocSegmentTime is Spur only. * VMMaker.oscog-AlistairGrant.2455.mcz VMClass strlen, strncpy and getenv * VMMaker.oscog-AlistairGrant.2456.mcz FakeStdinStream and FilePluginSimulator do double duty with the #atEnd flag to allow #sqFile:Read:Into:At: to break out of its loop. This is brittle as a additional calls to #atEnd breaks the simulation - which is what Pharo does. * VMMaker.oscog-eem.2457.mcz BitBltPlugin/BitBltSimulation * VMMaker.oscog-eem.2458.mcz InterpreterPrimitives Rewrite the getenv: simulation so that the original primitiveGetenv can be used. * VMMaker.oscog-AlistairGrant.2459.mcz VMClass>>strncpy:_:_: refactor * VMMaker.oscog-eem.2460.mcz Plugins: Use more Beck formatting (I find it very hard to read anything else; sorry). * VMMaker.oscog-eem.2461.mcz Plugins: BitBltSimulation * VMMaker.oscog-eem.2462.mcz JPEGReadWriter2Plugin: Fix a regression in the last commit. * VMMaker.oscog-eem.2463.mcz Oops! Fix the major regressions in the last but one commit. * VMMaker.oscog-eem.2464.mcz Cogit Slang Reflection: Fix nasty bug with CogAbstractInstruction computing opcodes via reflection caused by full blocks. We must use method allLiterals to include literals in blocks now. SysV is referenced only from blocks in CogX64Compiler class>>#initialize and so with full blocks SysV was being moved to Undeclared, causing invalid source generation for the X64 cogits. * VMMaker.oscog-AlistairGrant.2465.mcz strncpy & primitiveFailForOSError * VMMaker.oscog-akg.2466.mcz VMMaker Unicode strings * VMMaker.oscog-eem.2467.mcz Plugins: Fix a bug in FloatArrayPlugin>>primitiveDivFloatArray; the old code a) didn't check for -0.0 and b) didn't simulate. Eliminate several unnecessary stackObjectValue:'s in Matrix2x3Plugin; isWords: checkxs for immediates anyway. Fix several comments in FloatArrayPlugin primitives. * VMMaker.oscog-eem.2468.mcz Plugins: BalloonEngine & simulation. Clean up primitiveInitializeBuffer. Make some initial fixes to BalloonArray (the simulation of the engine's work buffer). I have other fish to fry right now but this is a stab in the right direction. * VMMaker.oscog-eem.2469.mcz :blush: * VMMaker.oscog-eem.2470.mcz Plugins: Fix a regression in VMMaker.oscog-eem.2467. * VMMaker.oscog-eem.2471.mcz Simulation: Fix a regression in memcpy:_:_: in VMMaker.oscog-eem.2467. Make malloc: simuulate a la alloca:. * VMMaker.oscog-eem.2472.mcz Simplify two uses of malloc: now that we use the right simulation form for VMClass>>malloc:. * VMMaker.oscog-akg.2473.mcz VM simulation fixes: * VMMaker.oscog-akg.2474.mcz VMClass>>memcpy:_:_: handle CugMethodSurrogates * VMMaker.oscog-akg.2475.mcz Add LocalePluginSimulator * VMMaker.oscog-eem.2476.mcz Slang: Make directory names in VMMaker work on FileSystem as well as FileDirectory; a first step to being able to generate sources on Pharo. Eliminate the obsolete machinesDirName. Remove unnecessary overrides of shouldBeTranslated (the default is ^true). Make baseTypeForType: answer e.g. 'int *' for the type 'int*', to comnform to the style I use in coerceTo:sim: and Slang type inferrence. * VMMaker.oscog-eem.2477.mcz Plugins: Fix primDigitCompare for SmallIntegers. The old code compared the values of the receiver and argument, not their magnitudes, if both were SmallIntegers. * VMMaker.oscog-eem.2478.mcz Fix a couple of tests * VMMaker.oscog-eem.2479.mcz Plugins: Fix slip in primitiveDirectoryDelimitor. * VMMaker.oscog-eem.2480.mcz Simulator: Provide simulation stubs for the B3DAcceleratorPlugin C api. * VMMaker.oscog-eem.2481.mcz ThreadedFFIPlugin: Fix a bug passing floats on the stack on X64. The original code passed stacked floats as doubles. * VMMaker.oscog-eem.2482.mcz Slang Fix mis-handling of trailing boolean on inlining in conditionals when trailing boolean would cause the conditional to be taken. e.g. in * VMMaker.oscog-eem.2483.mcz Add support for a sendWheelEvents flag that persists in the image header and is settable via vmParameterAt: 48 put: ...32... * VMMaker.oscog-eem.2484.mcz And sendWheelEvents must be global (not in VM's vars struct). * merge vmmaker into opensmalltalk-vm * Added baseline for loading vm maker from the github repo.mcz -- File Changes -- A .project (3) A smalltalksrc/.properties (3) A smalltalksrc/Android-Base/Android.class.st (144) A smalltalksrc/Android-Base/package.st (1) A smalltalksrc/Balloon-Engine-Pools/BalloonEngineConstants.class.st (558) A smalltalksrc/Balloon-Engine-Pools/package.st (1) A smalltalksrc/BaselineOfVMMaker/BaselineOfVMMaker.class.st (62) A smalltalksrc/BaselineOfVMMaker/package.st (1) A smalltalksrc/BytecodeSets-NewsqueakV3/BlockLocalTempCounter.extension.st (18) A smalltalksrc/BytecodeSets-NewsqueakV3/BytecodeEncoder.extension.st (53) A smalltalksrc/BytecodeSets-NewsqueakV3/ClosureExtractor.extension.st (6) A smalltalksrc/BytecodeSets-NewsqueakV3/ContextPart.extension.st (20) A smalltalksrc/BytecodeSets-NewsqueakV3/Decompiler.extension.st (7) A smalltalksrc/BytecodeSets-NewsqueakV3/EncoderForNewsqueakV3.class.st (33) A smalltalksrc/BytecodeSets-NewsqueakV3/EncoderForNewsqueakV4.class.st (1263) A smalltalksrc/BytecodeSets-NewsqueakV3/EncoderForSistaV1.class.st (1478) A smalltalksrc/BytecodeSets-NewsqueakV3/EncoderForSistaV1.extension.st (11) A smalltalksrc/BytecodeSets-NewsqueakV3/EncoderForV3.extension.st (9) A smalltalksrc/BytecodeSets-NewsqueakV3/FullBlockClosure.class.st (241) A smalltalksrc/BytecodeSets-NewsqueakV3/InstructionClient.extension.st (57) A smalltalksrc/BytecodeSets-NewsqueakV3/InstructionPrinter.extension.st (106) A smalltalksrc/BytecodeSets-NewsqueakV3/InstructionStream.extension.st (727) A smalltalksrc/BytecodeSets-NewsqueakV3/MethodContext.extension.st (9) A smalltalksrc/BytecodeSets-NewsqueakV3/MethodNode.extension.st (6) A smalltalksrc/BytecodeSets-NewsqueakV3/StackDepthFinder.extension.st (6) A smalltalksrc/BytecodeSets-NewsqueakV3/package.st (1) A smalltalksrc/BytecodeSets-NewsqueakV4/BlockLocalTempCounter.extension.st (18) A smalltalksrc/BytecodeSets-NewsqueakV4/BytecodeEncoder.extension.st (61) A smalltalksrc/BytecodeSets-NewsqueakV4/ClosureExtractor.extension.st (6) A smalltalksrc/BytecodeSets-NewsqueakV4/ContextPart.extension.st (20) A smalltalksrc/BytecodeSets-NewsqueakV4/Decompiler.extension.st (7) A smalltalksrc/BytecodeSets-NewsqueakV4/EncoderForNewsqueakV4.class.st (1288) A smalltalksrc/BytecodeSets-NewsqueakV4/EncoderForSistaV1.class.st (1530) A smalltalksrc/BytecodeSets-NewsqueakV4/EncoderForV3.extension.st (9) A smalltalksrc/BytecodeSets-NewsqueakV4/FullBlockClosure.class.st (241) A smalltalksrc/BytecodeSets-NewsqueakV4/InstructionClient.extension.st (48) A smalltalksrc/BytecodeSets-NewsqueakV4/InstructionPrinter.extension.st (110) A smalltalksrc/BytecodeSets-NewsqueakV4/InstructionStream.extension.st (443) A smalltalksrc/BytecodeSets-NewsqueakV4/MethodContext.extension.st (9) A smalltalksrc/BytecodeSets-NewsqueakV4/MethodNode.extension.st (6) A smalltalksrc/BytecodeSets-NewsqueakV4/StackDepthFinder.extension.st (7) A smalltalksrc/BytecodeSets-NewsqueakV4/package.st (1) A smalltalksrc/CMakeVMMaker-Android/CMCairo.class.st (133) A smalltalksrc/CMakeVMMaker-Android/CMCairoBundle.class.st (29) A smalltalksrc/CMakeVMMaker-Android/CMFreetype2.class.st (124) A smalltalksrc/CMakeVMMaker-Android/CMIOSFreetype2.class.st (66) A smalltalksrc/CMakeVMMaker-Android/CMLibPng.class.st (109) A smalltalksrc/CMakeVMMaker-Android/CMOSXFreetype2.class.st (18) A smalltalksrc/CMakeVMMaker-Android/CMOpenSSL.class.st (10) A smalltalksrc/CMakeVMMaker-Android/CMPixman.class.st (108) A smalltalksrc/CMakeVMMaker-Android/CMPkgConfig.class.st (83) A smalltalksrc/CMakeVMMaker-Android/CMThirdpartyLibrary.class.st (327) A smalltalksrc/CMakeVMMaker-Android/CMWin32Cairo.class.st (53) A smalltalksrc/CMakeVMMaker-Android/CMWin32Freetype2.class.st (94) A smalltalksrc/CMakeVMMaker-Android/CMWin32LibPng.class.st (67) A smalltalksrc/CMakeVMMaker-Android/CMWin32OpenSSL.class.st (113) A smalltalksrc/CMakeVMMaker-Android/CMWin32Pixman.class.st (45) A smalltalksrc/CMakeVMMaker-Android/CMWin32PkgConfig.class.st (67) A smalltalksrc/CMakeVMMaker-Android/CMWin32ZLib.class.st (96) A smalltalksrc/CMakeVMMaker-Android/CMakeAndroidGenerator.class.st (151) A smalltalksrc/CMakeVMMaker-Android/CMakeAndroidPluginGenerator.class.st (172) A smalltalksrc/CMakeVMMaker-Android/CMakeGenScripts.class.st (249) A smalltalksrc/CMakeVMMaker-Android/CMakeGenerator.class.st (253) A smalltalksrc/CMakeVMMaker-Android/CMakePluginGenerator.class.st (196) A smalltalksrc/CMakeVMMaker-Android/CMakeVMGenerator.class.st (376) A smalltalksrc/CMakeVMMaker-Android/CPlatformConfig.class.st (971) A smalltalksrc/CMakeVMMaker-Android/CocoaIOSConfig.class.st (1145) A smalltalksrc/CMakeVMMaker-Android/CogCocoaIOSConfig.class.st (66) A smalltalksrc/CMakeVMMaker-Android/CogFamilyCocoaIOSConfig.class.st (360) A smalltalksrc/CMakeVMMaker-Android/CogFamilyUnixConfig.class.st (397) A smalltalksrc/CMakeVMMaker-Android/CogFamilyWindowsConfig.class.st (378) A smalltalksrc/CMakeVMMaker-Android/CogFreeBSDConfig.class.st (301) A smalltalksrc/CMakeVMMaker-Android/CogMTCocoaIOSConfig.class.st (43) A smalltalksrc/CMakeVMMaker-Android/CogMTCocoaIOSDebugConfig.class.st (22) A smalltalksrc/CMakeVMMaker-Android/CogMTFreeBSDConfig.class.st (63) A smalltalksrc/CMakeVMMaker-Android/CogMTMacOSConfig.class.st (60) A smalltalksrc/CMakeVMMaker-Android/CogMTUnixConfig.class.st (63) A smalltalksrc/CMakeVMMaker-Android/CogMTWindowsConfig.class.st (54) A smalltalksrc/CMakeVMMaker-Android/CogMacOSConfig.class.st (911) A smalltalksrc/CMakeVMMaker-Android/CogMacOSDebugConfig.class.st (17) A smalltalksrc/CMakeVMMaker-Android/CogOnDebian64Config.class.st (31) A smalltalksrc/CMakeVMMaker-Android/CogUnixConfig.class.st (378) A smalltalksrc/CMakeVMMaker-Android/CogUnixDebugConfig.class.st (18) A smalltalksrc/CMakeVMMaker-Android/CogUnixNoGLConfig.class.st (201) A smalltalksrc/CMakeVMMaker-Android/CogWindowsConfig.class.st (229) A smalltalksrc/CMakeVMMaker-Android/CogWindowsDebugConfig.class.st (20) A smalltalksrc/CMakeVMMaker-Android/InterpreterPlugin.extension.st (22) A smalltalksrc/CMakeVMMaker-Android/MacOSConfig.class.st (415) A smalltalksrc/CMakeVMMaker-Android/StackCocoaIOSARMConfig.class.st (399) A smalltalksrc/CMakeVMMaker-Android/StackCocoaIOSCLANGConfig.class.st (28) A smalltalksrc/CMakeVMMaker-Android/StackCocoaIOSConfig.class.st (58) A smalltalksrc/CMakeVMMaker-Android/StackEvtAndroidConfig.class.st (416) A smalltalksrc/CMakeVMMaker-Android/StackEvtUnixConfig.class.st (71) A smalltalksrc/CMakeVMMaker-Android/StackEvtUnixDebugConfig.class.st (11) A smalltalksrc/CMakeVMMaker-Android/StackFreeBSDConfig.class.st (31) A smalltalksrc/CMakeVMMaker-Android/StackIPhoneConfig.class.st (674) A smalltalksrc/CMakeVMMaker-Android/StackMacOSConfig.class.st (39) A smalltalksrc/CMakeVMMaker-Android/StackMacOSDebugConfig.class.st (17) A smalltalksrc/CMakeVMMaker-Android/StackSimulatorConfig.class.st (32) A smalltalksrc/CMakeVMMaker-Android/StackUnixConfig.class.st (40) A smalltalksrc/CMakeVMMaker-Android/StackUnixDebugConfig.class.st (18) A smalltalksrc/CMakeVMMaker-Android/StackUnixDebugFixedVerSIConfig.class.st (38) A smalltalksrc/CMakeVMMaker-Android/StackWindowsConfig.class.st (37) A smalltalksrc/CMakeVMMaker-Android/String.extension.st (29) A smalltalksrc/CMakeVMMaker-Android/package.st (1) A smalltalksrc/CMakeVMMaker/CMCairo.class.st (133) A smalltalksrc/CMakeVMMaker/CMCairoBundle.class.st (29) A smalltalksrc/CMakeVMMaker/CMFreetype2.class.st (0) A smalltalksrc/CMakeVMMaker/CMIOSFreetype2.class.st (0) A smalltalksrc/CMakeVMMaker/CMLibPng.class.st (0) A smalltalksrc/CMakeVMMaker/CMOSXFreetype2.class.st (0) A smalltalksrc/CMakeVMMaker/CMOpenSSL.class.st (0) A smalltalksrc/CMakeVMMaker/CMPixman.class.st (0) A smalltalksrc/CMakeVMMaker/CMPkgConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CMThirdpartyLibrary.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32Cairo.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32Freetype2.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32LibPng.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32OpenSSL.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32Pixman.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32PkgConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CMWin32ZLib.class.st (0) A smalltalksrc/CMakeVMMaker/CMakeAndroidGenerator.class.st (0) A smalltalksrc/CMakeVMMaker/CMakeAndroidPluginGenerator.class.st (0) A smalltalksrc/CMakeVMMaker/CMakeGenScripts.class.st (0) A smalltalksrc/CMakeVMMaker/CMakeGenerator.class.st (0) A smalltalksrc/CMakeVMMaker/CMakePluginGenerator.class.st (0) A smalltalksrc/CMakeVMMaker/CMakeVMGenerator.class.st (0) A smalltalksrc/CMakeVMMaker/CPlatformConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CocoaIOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogCocoaIOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogFamilyCocoaIOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogFamilyUnixConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogFamilyWindowsConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogFreeBSDConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTBuilder.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTCocoaIOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTFreeBSDConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTMacOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTUnixConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMTWindowsConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMacOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogMacOSDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogOnDebian64Config.class.st (0) A smalltalksrc/CMakeVMMaker/CogUnixConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogUnixDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogUnixNoGLConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogWindowsConfig.class.st (0) A smalltalksrc/CMakeVMMaker/CogWindowsDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/InterpreterPlugin.extension.st (0) A smalltalksrc/CMakeVMMaker/MacOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackCocoaIOSCLANGConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackCocoaIOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackCrossRaspbianConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackCrossRaspbianFastBltConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackEvtAndroidConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackEvtUnixConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackEvtUnixDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackFreeBSDConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackIPhoneConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackMacOSConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackMacOSDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackRaspbianConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackRaspbianFastBltConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackSimulatorConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackUnixConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackUnixDebugConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackUnixDebugFixedVerSIConfig.class.st (0) A smalltalksrc/CMakeVMMaker/StackWindowsConfig.class.st (0) A smalltalksrc/CMakeVMMaker/package.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddCompileOptions.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddCustomCommandOutput.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddDefinitions.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddExecutableNameOptionSource.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddLibrary.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAddSubDirectory.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAppendConfigHIn.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeAppendConfigStatus.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckAlloca.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckFunctionAtExitOnExit.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckFunctionExists.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckIncludeFile.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckLibDL.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckLibraryExists.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckStructHasMember.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckTryCompileHaveLangInfoCodeset.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckTryRunDoubleWordAlignmentOrder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckTypeSize.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCheckVariableExists.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCommand.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeCompilerIsGNUCC.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeConfigDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeConfigUnDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeDebug.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeDirDefs.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeFindPackage.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeFindPackageOpenGL.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeGeneratorForSqueak.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeHeader.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIOSInstallCode.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIfAddDefinitionsElseAddDefinitions.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIfDefinedConfigDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIfNotFlagSetConfigDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIfNotSetConfigDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIfPluginDependencies.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeInclude.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeIncludeDirectories.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeLinkDirectories.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeListAppend.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeMessage.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeMinimumRequired.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakePluginExternal.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakePluginGeneratorForSqueak.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakePluginInternal.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakePluginVm.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeProject.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSet.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSetConfigDefine.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSetProperty.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSetSourceFilesProperties.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSetTargetProperties.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeSqConfigH.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeTargetLinkLibraries.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeTemplate.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeTestBigEndian.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMDisplayCustom.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMDisplayFbdev.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMDisplayNull.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMDisplayQuartz.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMDisplayX11.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMGeneratorForSqueak.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakBuildersHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakBuildersTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakCMakeTemplatesTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakCommonConfigTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakConfigurationsHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakDesignPatternsHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakDeveloperHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakDistroGeneratorHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakGeneratorsHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakHistoryHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakHistoryIanPiumartaHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakHistoryIgorStasenkoHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakMacintoshConfigTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakOverviewHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakPluginGeneratorHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakRedirectMethodsTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakRedirectMethodsWithArgTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakStartHereHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakTemplatesHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakTutorialEndUserHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakTutorialNewBuilderHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakTutorialNewConfigurationHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakTutorialsHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakUnixConfigTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakVMGeneratorHelp.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMMakerSqueakWindowsConfigTest.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMPlugin.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundALSA.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundCustom.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundMacOSX.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundNAS.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundNull.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundOSS.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundPulse.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMSoundSun.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVMakerConfigurationInfo.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CMakeVersion.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/CPlatformConfigForSqueak.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/InterpreterPlugin.extension.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux32ARMv6Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux32x86Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux32x86SqueakCogV3Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x64Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x64SqueakCogSpurConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x86w32BitConfigUsrLib.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x86w32BitConfigUsrLib32.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x86w32BitSqueakCogSpurConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/Linux64x86w32BitSqueakCogV3Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakBSD32x86Builder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakBSD32x86Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakCMakeSourceDistroBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakCMakeVMMakerAbstractBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsLinuxConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsMacOSConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsMacOSXConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIA32BochsWin32Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIOSBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakIOSConfig.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakLinux32ARMv6Builder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakLinux32x86Builder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakLinux64x64Builder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakLinux64x86w32CompatBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakMacOSPowerPCBuilder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakMacOSX32x86Builder.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakMacOSX32x86Config.class.st (0) A smalltalksrc/CMakeVMMakerSqueak/SqueakMacOSXPowerPCConfig.class.st (0) -- Patch Links -- https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/309.patch https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/309.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/309

2 2

[OpenSmalltalk/opensmalltalk-vm] Second attempt to improve updateSCCSVersions (#227)
by Jakob Reschke 29 Dec '18

29 Dec '18

I finally got around to solve the "not a git repository" issue with my previous pull request. For the analysis see below. Please review and try it out before merging. When the script is run as a hook, the environment variable `GIT_DIR` is set to `.git` (path relative to the working tree) for regular repositories. This affects the behavior of `rev-parse`. When the script is run from the working tree and the checkout commands towards the bottom are executed, the script actually invokes itself as the post-checkout script. During that inner invocation, the script would `cd` to `.git/hooks`, and because of the GIT_DIR variable, `rev-parse` would then look for the git repository in `.git/hooks/.git`. This failed, of course. The second error message "unrecognized input" came from `git apply` when it was fed with empty input (because there were no uncommitted changes to either *Version.h file). Since GIT_DIR caused the trouble, why not use it to detect whether the script is run as a hook... You can view, comment on, or merge this pull request online at: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/227 -- Commit Summary -- * Second attempt to improve updateSCCSVersions * Eschew --git-path because it was only added in Git 2.5 * Properly handle staged changes to the version headers * whitespace -- File Changes -- M scripts/updateSCCSVersions (40) -- Patch Links -- https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/227.patch https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/227.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/227

5 9

[OpenSmalltalk/opensmalltalk-vm] Make updateSCCSVersions work in more situations (#213)
by Jakob Reschke 29 Dec '18

29 Dec '18

- allow spaces in paths (added quotes) - support cases where the repository is not in a .git directory at the top of the working tree - do not hardcode the directory level of this script - avoid stashing the whole working tree when only two well-known files need to be updated - touch should suffice to make a file different from its index version (i. e., eligible for checkout) Please review for compatibility and pick those parts of the changes that you like. You can view, comment on, or merge this pull request online at: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/213 -- Commit Summary -- * make updateSCCSVersions work in more situations -- File Changes -- M scripts/updateSCCSVersions (26) -- Patch Links -- https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/213.patch https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/213.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/213

5 9

Minheadless trial
by Ben Coman 05 Dec '18

05 Dec '18

wow time flies, its been 18 months since Ronie announced his Minheadless VM branch... [1] http://forum.world.st/Minheadless-VM-flavour-status-update-td4928091.html I've been meaning this whole time to try it out. Recently in my meandering thoughts on FreeRTOS port, I guessed Minheadless might make a good starting basis to experiment with. So a quick test on Ubuntu16.04.3(Windows10-WSL)... Long ago I'd already done... git clone git@github.com:OpenSmalltalk/opensmalltalk-vm.git so I now did... cd opensmalltalk-vm git remote add ronie git@github.com:ronsaldo/opensmalltalk-vm.git git fetch ronie git checkout MinimalisticHeadless ./scripts/updateSCCSVersions cd build.minheadless.cmake/x64/pharo.stack.spur/ ./mvm VM=`pwd`/release/dist/pharo ls $VM ==> confirmed it exists cd ~/pharo wget http://files.pharo.org/image/70/latest-minimal-64.zip unzip latest-minimal-64.zip sudo $VM Pharo7.0-metacello-64bit-65cff7b.image eval "19 + 23" ==> 42 yay Particularly interesting from [1] is "using the same CMake scripts for building on the three platforms: Windows, Linux and OS X." So I thought I'd also give it a try under Windows10-VisualStudio2017. I hardly used VS before, so I'll detail what I did as a reminder for myself (and also in case someone might be able to help if I get stuck). Under a windows cmd.exe prompt (having installed git and cloned opensmalltalk-vm a long time ago) cd opensmalltalk-vm git remote add ronie git@github.com:ronsaldo/opensmalltalk-vm.git git fetch ronie git checkout -b MinimalisticHeadless ronie/MinimalisticHeadless I watched video [2] about CMake integration with Visual Studio 10. and checked the VS workloads configuration specified [2@1:13] using [3]. There is a CMakeLists.txt file in the git root folder, so.. started VS > File > Open > Folder... C:\Repos\OpenSmalltalk\opensmalltalk-vm The CMake command run ([2]@2:13) was... C:\...\cmake.exe -G "Ninja" -DCMAKE_INSTALL_PREFIX:PATH="C:\Users\Ben\CMakeBuilds\...\x86-Debug" -DCMAKE_CXX_COMPILER="C:/.../MSVC/14.14.26428/bin/HostX86/x86/cl.exe" -DCMAKE_C_COMPILER="C:/.../MSVC/14.14.26428/bin/HostX86/x86/cl.exe" -DCMAKE_BUILD_TYPE="Debug" -DCMAKE_MAKE_PROGRAM="C:\...\CMAKE\Ninja\ninja.exe" "C:\Repos\OpenSmalltalk\opensmalltalk-vm" I changed the cmake configuration [2@2:27] to "x64-Debug". For debug target [2@3:00] I chose "pharo.exe" Hey, thats pretty cool [2@4:05] that switching between targets changes the highlighting of #ifdef code. I had a moment when I was missing the "Debug and Launch Settings" menu item [2@4:40-4:55] and mentioned at [4], but maybe there was some clash with previously experimenting with git cloning opensmalltalk-vm from inside VS. After closing VS and reopening my cmd.exe git'd opensmalltalk-vm it showed up, but I didn't change any settings there. Then (with x64-Debug & pharo.exe selected) I went Cmake > Build All ==> 10 Errors, 9 Warnings, which I'll discuss in a followup post. cheers -ben [2] https://blogs.msdn.microsoft.com/vcblog/2016/10/05/cmake-support-in-visual-… [3] https://docs.microsoft.com/en-us/visualstudio/install/modify-visual-studio [4] https://blogs.msdn.microsoft.com/vcblog/2016/12/20/cmake-support-in-visual-… [5] https://developercommunity.visualstudio.com/content/problem/234808/cmake-de…

3 10

[OpenSmalltalk/opensmalltalk-vm] Fix the build of the minheadless VM for Pharo in Win32. (#311)
by Ronie Salgado 01 Dec '18

01 Dec '18

Quick fixup of the minheadless VM for windows using the CMake building scripts. The VM builds and works with Cygwin for 32 bits. It builds with MSVC, but fails with several errors on runtime. You can view, comment on, or merge this pull request online at: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/311 -- Commit Summary -- * Fix the build of the minheadless VM for Pharo in Win32. -- File Changes -- M CMakeLists.txt (23) M build.minheadless.cmake/x64/common/configure_variant.sh (1) M cmake/Plugins.cmake (9) M cmake/PluginsPharo.cmake (19) M platforms/Cross/plugins/IA32ABI/ia32abicc.c (4) M platforms/Cross/plugins/IA32ABI/x64win64abicc.c (8) M platforms/Cross/plugins/IA32ABI/xabicc.c (6) M platforms/Cross/plugins/Mpeg3Plugin/libmpeg/changesForSqueak.c (3) M platforms/Cross/plugins/Mpeg3Plugin/libmpeg/mpeg3private.h (3) M platforms/minheadless/windows/sqPlatformSpecific-Win32.c (123) M platforms/minheadless/windows/sqWin32Directory.c (269) A platforms/minheadless/windows/sqWin32Stubs.c (19) -- Patch Links -- https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/311.patch https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/311.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/311

3 2

The insanity of using make for development builds
by Eliot Miranda 01 Dec '18

01 Dec '18

Hi All, I've just attempted to build minheadless on win32 in build.minheadless.cmake. It failed, but it took a long time to do so, while I waited. So I cleaned and repeated to measure just how long I waited. So this time is a best case; all read files are in cache etc. I'm running a Windows VM on a top of the line MacBook Pro and yet it takes 3 minutes and 20 seconds to configure and then start to build and fail because it can't find sqplatfozrmspecifc-win32.h. I have to wait 3 minutes and 20 seconds before I can find out a missing file. This is *wrong*. Using make on a build slave is fine, if you insist, but clearly not necessary (our current builds do not use cmake, and even then on ARM build slaves they sometimes timeout). But for development this is madness. We should have a build system which is reactive, which gives feedback too the developer quickly, not after 3 minutes and 20 seconds on state of the art hardware. I find it absurd that a Smalltalk community, which well understands the value of incrementally and reactivity, can be satisfied with a VM build system that takes 3 minutes and 20 seconds before it does anything useful. It feels like being back at York University in the early '80's, waiting for the PDP 10. No, that was faster. On reflection, it reminds me of having to post coding sheets to be typed in by entry clerks through the mail while I was at school in the '70's. Back to the future indeed. _,,,^..^,,,_ best, Eliot

9 14

#isContext: is sent to self instead of objectMemory
by Alistair Grant 29 Nov '18

29 Nov '18

Hi Eliot, I think I've found a couple of typos: - CoInterpreter>>ceNonLocalReturn: - StackInterpreter>>commonReturn both send #isContext: to self, while I think it should be sent to objectMemory. Can you take a look at: http://source.squeak.org/VMMakerInbox/VMMaker.oscog-akg.2487.mcz and let me know what you think? Thanks, Alistair

2 2

VM Maker: VMMaker.oscog-akg.2487.mcz
by commits＠source.squeak.org 29 Nov '18

29 Nov '18

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-akg.2487.mcz ==================== Summary ==================== Name: VMMaker.oscog-akg.2487 Author: akg Time: 29 November 2018, 6:40:57.726983 pm UUID: 81e92544-31de-4944-a742-11a331113606 Ancestors: VMMaker.oscog-cb.2486 #isContext: is sent to self instead of objectMemory Fix a couple of typo's in: - CoInterpreter>>ceNonLocalReturn: - StackInterpreter>>commonReturn where #isContext: is sent to self instead of objectMemory. =============== Diff against VMMaker.oscog-cb.2486 =============== Item was changed: ----- Method: CoInterpreter>>ceNonLocalReturn: (in category 'trampolines') ----- ceNonLocalReturn: returnValue <api> | closure home unwindContextOrNilOrZero ourContext frameToReturnTo contextToReturnTo theFP callerFP newPage | <var: #frameToReturnTo type: #'char *'> <var: #theFP type: #'char *'> <var: #callerFP type: #'char *'> <var: #newPage type: #'StackPage *'> <var: #thePage type: #'StackPage *'> "self shortPrintFrameAndCallers: framePointer. self printOop: returnValue. self halt." self assert: (self isMachineCodeFrame: framePointer). self assert: (self frameIsBlockActivation: framePointer). "Update the current page's headFrame pointers to enable the search for unwind protects below to identify widowed contexts correctly." self externalWriteBackHeadFramePointers. "Since this is a block activation the closure is on the stack above any args and the frame." closure := self pushedReceiverOrClosureOfFrame: framePointer. home := nil. "avoid compiler warning" "Walk the closure's lexical chain to find the context or frame to return from (home)." [closure ~= objectMemory nilObject] whileTrue: [home := objectMemory followField: ClosureOuterContextIndex ofObject: closure. + (objectMemory isContext: home) ifFalse: - (self isContext: home) ifFalse: ["error: can't find home on chain; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]. closure := objectMemory followField: ClosureIndex ofObject: home]. "home is to be returned from provided there is no unwind-protect activation between this frame and home's sender. Search for an unwind. findUnwindThroughContext: will answer either the context for an unwind-protect activation or nilObj if the sender cannot be found or 0 if no unwind is found but the sender is." unwindContextOrNilOrZero := self findUnwindThroughContext: home. unwindContextOrNilOrZero = objectMemory nilObject ifTrue: ["error: can't find home on chain; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]. unwindContextOrNilOrZero ~= 0 ifTrue: [^self externalAboutToReturn: returnValue through: unwindContextOrNilOrZero]. "Now we know home is on the sender chain. We could be returning to either a context or a frame. Find out which." contextToReturnTo := nil. (self isMarriedOrWidowedContext: home) ifTrue: [self assert: (self checkIsStillMarriedContext: home currentFP: framePointer). theFP := self frameOfMarriedContext: home. (self isBaseFrame: theFP) ifTrue: [contextToReturnTo := self frameCallerContext: theFP] ifFalse: [frameToReturnTo := self frameCallerFP: theFP]] ifFalse: [contextToReturnTo := objectMemory fetchPointer: SenderIndex ofObject: home. ((objectMemory isContext: contextToReturnTo) and: [self isMarriedOrWidowedContext: contextToReturnTo]) ifTrue: [self assert: (self checkIsStillMarriedContext: contextToReturnTo currentFP: framePointer). frameToReturnTo := self frameOfMarriedContext: contextToReturnTo. contextToReturnTo := nil]]. "If returning to a context we must make a frame for it unless it is dead." contextToReturnTo ~= nil ifTrue: [frameToReturnTo := self establishFrameForContextToReturnTo: contextToReturnTo. frameToReturnTo = 0 ifTrue: ["error: home's sender is dead; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]]. "Now we have a frame to return to. If it is on a different page we must free intervening pages and nil out intervening contexts. We must free intervening stack pages because if we leave the pages to be divorced then their contexts will be divorced with intact senders and instruction pointers. This code is similar to primitiveTerminateTo." self assert: stackPages pageListIsWellFormed. newPage := stackPages stackPageFor: frameToReturnTo. newPage ~~ stackPage ifTrue: [| currentCtx thePage nextCntx | currentCtx := self frameCallerContext: stackPage baseFP. self assert: (objectMemory isContext: currentCtx). stackPages freeStackPage: stackPage. [self assert: (objectMemory isContext: currentCtx). (self isMarriedOrWidowedContext: currentCtx) and: [(stackPages stackPageFor: (theFP := self frameOfMarriedContext: currentCtx)) = newPage]] whileFalse: [(self isMarriedOrWidowedContext: currentCtx) ifTrue: [thePage := stackPages stackPageFor: theFP. currentCtx := self frameCallerContext: thePage baseFP. stackPages freeStackPage: thePage] ifFalse: [nextCntx := objectMemory fetchPointer: SenderIndex ofObject: currentCtx. self markContextAsDead: currentCtx. currentCtx := nextCntx]]. self setStackPageAndLimit: newPage. self setStackPointersFromPage: newPage]. "Two cases. Returning to the top frame or an interior frame. The top frame has its instruction pointer on top of stack. An interior frame has its instruction pointer in the caller frame. We need to peel back any frames on the page until we get to the correct frame." framePointer = frameToReturnTo ifTrue: [instructionPointer := self popStack] ifFalse: [[callerFP := framePointer. framePointer := self frameCallerFP: framePointer. framePointer ~~ frameToReturnTo] whileTrue. instructionPointer := (self frameCallerSavedIP: callerFP) asUnsignedInteger. stackPointer := (self frameCallerSP: callerFP)]. ^self return: returnValue toExecutive: false! Item was changed: ----- Method: StackInterpreter>>commonReturn (in category 'return bytecodes') ----- commonReturn "Do an ^-return (return from method), checking for unwinds if this is a block activation. Note: Assumed to be inlined into the dispatch loop." <sharedCodeInCase: #returnReceiver> | closure home unwindContextOrNilOrZero frameToReturnTo contextToReturnTo theFP callerFP newPage | <var: #frameToReturnTo type: #'char *'> <var: #theFP type: #'char *'> <var: #callerFP type: #'char *'> <var: #newPage type: #'StackPage *'> <var: #thePage type: #'StackPage *'> "If this is a method simply return to the sender/caller." (self iframeIsBlockActivation: localFP) ifFalse: [^self commonCallerReturn]. "Update the current page's headFrame pointers to enable the search for unwind protects below to identify widowed contexts correctly." self writeBackHeadFramePointers. self externalizeIPandSP. "Since this is a block activation the closure is on the stack above any args and the frame." closure := self pushedReceiverOrClosureOfFrame: localFP. home := nil. "avoid compiler warning" "Walk the closure's lexical chain to find the context or frame to return from (home). If home is missing (Sista closures) then throw cannotReturn rather than crash." [closure ~= objectMemory nilObject] whileTrue: [home := objectMemory followField: ClosureOuterContextIndex ofObject: closure. + (objectMemory isContext: home) ifFalse: - (self isContext: home) ifFalse: [^self internalCannotReturn: localReturnValue]. closure := objectMemory followField: ClosureIndex ofObject: home]. "home is to be returned from provided there is no unwind-protect activation between this frame and home's sender. Search for an unwind. findUnwindThroughContext: will answer either the context for an unwind-protect activation or nilObj if the sender cannot be found or 0 if no unwind is found but the sender is." unwindContextOrNilOrZero := self findUnwindThroughContext: home. unwindContextOrNilOrZero = objectMemory nilObject ifTrue: ["error: can't find home on chain; cannot return" ^self internalCannotReturn: localReturnValue]. unwindContextOrNilOrZero ~= 0 ifTrue: [^self internalAboutToReturn: localReturnValue through: unwindContextOrNilOrZero]. "Now we know home is on the sender chain. We could be returning to either a context or a frame. Find out which." contextToReturnTo := nil. (self isMarriedOrWidowedContext: home) ifTrue: [self assert: (self checkIsStillMarriedContext: home currentFP: localFP). theFP := self frameOfMarriedContext: home. (self isBaseFrame: theFP) ifTrue: [contextToReturnTo := self frameCallerContext: theFP] ifFalse: [frameToReturnTo := self frameCallerFP: theFP]] ifFalse: [contextToReturnTo := objectMemory fetchPointer: SenderIndex ofObject: home. ((objectMemory isContext: contextToReturnTo) and: [self isMarriedOrWidowedContext: contextToReturnTo]) ifTrue: [self assert: (self checkIsStillMarriedContext: contextToReturnTo currentFP: localFP). frameToReturnTo := self frameOfMarriedContext: contextToReturnTo. contextToReturnTo := nil]]. "If returning to a context we must make a frame for it unless it is dead." contextToReturnTo ~= nil ifTrue: [frameToReturnTo := self establishFrameForContextToReturnTo: contextToReturnTo. frameToReturnTo = 0 ifTrue: "error: home's sender is dead; cannot return" [^self internalCannotReturn: localReturnValue]]. "Now we have a frame to return to. If it is on a different page we must free intervening pages and nil out intervening contexts. We must free intervening stack pages because if we leave the pages to be divorced then their contexts will be divorced with intact senders and instruction pointers. This code is similar to primitiveTerminateTo. We must move any frames on intervening pages above the frame linked to because these may be in use, e.g. via co-routining (see baseFrameReturn)." self assert: stackPages pageListIsWellFormed. newPage := stackPages stackPageFor: frameToReturnTo. newPage ~~ stackPage ifTrue: [| currentCtx thePage nextCntx | currentCtx := self frameCallerContext: stackPage baseFP. stackPages freeStackPage: stackPage. [self assert: (objectMemory isContext: currentCtx). (self isMarriedOrWidowedContext: currentCtx) and: [(stackPages stackPageFor: (theFP := self frameOfMarriedContext: currentCtx)) = newPage]] whileFalse: [(self isMarriedOrWidowedContext: currentCtx) ifTrue: [thePage := stackPages stackPageFor: theFP. theFP ~= thePage headFP ifTrue: ["Since we've just deallocated a page we know that newStackPage won't deallocate an existing one." self moveFramesIn: thePage through: (self findFrameAbove: theFP inPage: thePage) toPage: stackPages newStackPage]. currentCtx := self frameCallerContext: thePage baseFP. stackPages freeStackPage: thePage] ifFalse: [nextCntx := objectMemory fetchPointer: SenderIndex ofObject: currentCtx. self markContextAsDead: currentCtx. currentCtx := nextCntx]]. self setStackPageAndLimit: newPage. localSP := stackPage headSP. localFP := stackPage headFP]. "Two cases. Returning to the top frame on a new page or an interior frame on the current page. The top frame has its instruction pointer on top of stack. An interior frame has its instruction pointer in the caller frame. We need to peel back any frames on the page until we get to the correct frame." localFP = frameToReturnTo ifTrue: "pop the saved IP, push the return value and continue." [localIP := self pointerForOop: self internalStackTop] ifFalse: [[callerFP := localFP. localFP := self frameCallerFP: localFP. localFP ~~ frameToReturnTo] whileTrue. localIP := self frameCallerSavedIP: callerFP. localSP := (self frameCallerSP: callerFP) - objectMemory wordSize]. self maybeReturnToMachineCodeFrame. self setMethod: (self frameMethod: localFP). self fetchNextBytecode. self internalStackTopPut: localReturnValue!

1 0

VM Maker Inbox: VMMaker.oscog-akg.2487.mcz
by commits＠source.squeak.org 29 Nov '18

29 Nov '18

Alistair Grant uploaded a new version of VMMaker to project VM Maker Inbox: http://source.squeak.org/VMMakerInbox/VMMaker.oscog-akg.2487.mcz ==================== Summary ==================== Name: VMMaker.oscog-akg.2487 Author: akg Time: 29 November 2018, 6:40:57.726983 pm UUID: 81e92544-31de-4944-a742-11a331113606 Ancestors: VMMaker.oscog-cb.2486 #isContext: is sent to self instead of objectMemory Fix a couple of typo's in: - CoInterpreter>>ceNonLocalReturn: - StackInterpreter>>commonReturn where #isContext: is sent to self instead of objectMemory. =============== Diff against VMMaker.oscog-cb.2486 =============== Item was changed: ----- Method: CoInterpreter>>ceNonLocalReturn: (in category 'trampolines') ----- ceNonLocalReturn: returnValue <api> | closure home unwindContextOrNilOrZero ourContext frameToReturnTo contextToReturnTo theFP callerFP newPage | <var: #frameToReturnTo type: #'char *'> <var: #theFP type: #'char *'> <var: #callerFP type: #'char *'> <var: #newPage type: #'StackPage *'> <var: #thePage type: #'StackPage *'> "self shortPrintFrameAndCallers: framePointer. self printOop: returnValue. self halt." self assert: (self isMachineCodeFrame: framePointer). self assert: (self frameIsBlockActivation: framePointer). "Update the current page's headFrame pointers to enable the search for unwind protects below to identify widowed contexts correctly." self externalWriteBackHeadFramePointers. "Since this is a block activation the closure is on the stack above any args and the frame." closure := self pushedReceiverOrClosureOfFrame: framePointer. home := nil. "avoid compiler warning" "Walk the closure's lexical chain to find the context or frame to return from (home)." [closure ~= objectMemory nilObject] whileTrue: [home := objectMemory followField: ClosureOuterContextIndex ofObject: closure. + (objectMemory isContext: home) ifFalse: - (self isContext: home) ifFalse: ["error: can't find home on chain; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]. closure := objectMemory followField: ClosureIndex ofObject: home]. "home is to be returned from provided there is no unwind-protect activation between this frame and home's sender. Search for an unwind. findUnwindThroughContext: will answer either the context for an unwind-protect activation or nilObj if the sender cannot be found or 0 if no unwind is found but the sender is." unwindContextOrNilOrZero := self findUnwindThroughContext: home. unwindContextOrNilOrZero = objectMemory nilObject ifTrue: ["error: can't find home on chain; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]. unwindContextOrNilOrZero ~= 0 ifTrue: [^self externalAboutToReturn: returnValue through: unwindContextOrNilOrZero]. "Now we know home is on the sender chain. We could be returning to either a context or a frame. Find out which." contextToReturnTo := nil. (self isMarriedOrWidowedContext: home) ifTrue: [self assert: (self checkIsStillMarriedContext: home currentFP: framePointer). theFP := self frameOfMarriedContext: home. (self isBaseFrame: theFP) ifTrue: [contextToReturnTo := self frameCallerContext: theFP] ifFalse: [frameToReturnTo := self frameCallerFP: theFP]] ifFalse: [contextToReturnTo := objectMemory fetchPointer: SenderIndex ofObject: home. ((objectMemory isContext: contextToReturnTo) and: [self isMarriedOrWidowedContext: contextToReturnTo]) ifTrue: [self assert: (self checkIsStillMarriedContext: contextToReturnTo currentFP: framePointer). frameToReturnTo := self frameOfMarriedContext: contextToReturnTo. contextToReturnTo := nil]]. "If returning to a context we must make a frame for it unless it is dead." contextToReturnTo ~= nil ifTrue: [frameToReturnTo := self establishFrameForContextToReturnTo: contextToReturnTo. frameToReturnTo = 0 ifTrue: ["error: home's sender is dead; cannot return" ourContext := self ensureFrameIsMarried: framePointer SP: stackPointer. ^self externalCannotReturn: returnValue from: ourContext]]. "Now we have a frame to return to. If it is on a different page we must free intervening pages and nil out intervening contexts. We must free intervening stack pages because if we leave the pages to be divorced then their contexts will be divorced with intact senders and instruction pointers. This code is similar to primitiveTerminateTo." self assert: stackPages pageListIsWellFormed. newPage := stackPages stackPageFor: frameToReturnTo. newPage ~~ stackPage ifTrue: [| currentCtx thePage nextCntx | currentCtx := self frameCallerContext: stackPage baseFP. self assert: (objectMemory isContext: currentCtx). stackPages freeStackPage: stackPage. [self assert: (objectMemory isContext: currentCtx). (self isMarriedOrWidowedContext: currentCtx) and: [(stackPages stackPageFor: (theFP := self frameOfMarriedContext: currentCtx)) = newPage]] whileFalse: [(self isMarriedOrWidowedContext: currentCtx) ifTrue: [thePage := stackPages stackPageFor: theFP. currentCtx := self frameCallerContext: thePage baseFP. stackPages freeStackPage: thePage] ifFalse: [nextCntx := objectMemory fetchPointer: SenderIndex ofObject: currentCtx. self markContextAsDead: currentCtx. currentCtx := nextCntx]]. self setStackPageAndLimit: newPage. self setStackPointersFromPage: newPage]. "Two cases. Returning to the top frame or an interior frame. The top frame has its instruction pointer on top of stack. An interior frame has its instruction pointer in the caller frame. We need to peel back any frames on the page until we get to the correct frame." framePointer = frameToReturnTo ifTrue: [instructionPointer := self popStack] ifFalse: [[callerFP := framePointer. framePointer := self frameCallerFP: framePointer. framePointer ~~ frameToReturnTo] whileTrue. instructionPointer := (self frameCallerSavedIP: callerFP) asUnsignedInteger. stackPointer := (self frameCallerSP: callerFP)]. ^self return: returnValue toExecutive: false! Item was changed: ----- Method: StackInterpreter>>commonReturn (in category 'return bytecodes') ----- commonReturn "Do an ^-return (return from method), checking for unwinds if this is a block activation. Note: Assumed to be inlined into the dispatch loop." <sharedCodeInCase: #returnReceiver> | closure home unwindContextOrNilOrZero frameToReturnTo contextToReturnTo theFP callerFP newPage | <var: #frameToReturnTo type: #'char *'> <var: #theFP type: #'char *'> <var: #callerFP type: #'char *'> <var: #newPage type: #'StackPage *'> <var: #thePage type: #'StackPage *'> "If this is a method simply return to the sender/caller." (self iframeIsBlockActivation: localFP) ifFalse: [^self commonCallerReturn]. "Update the current page's headFrame pointers to enable the search for unwind protects below to identify widowed contexts correctly." self writeBackHeadFramePointers. self externalizeIPandSP. "Since this is a block activation the closure is on the stack above any args and the frame." closure := self pushedReceiverOrClosureOfFrame: localFP. home := nil. "avoid compiler warning" "Walk the closure's lexical chain to find the context or frame to return from (home). If home is missing (Sista closures) then throw cannotReturn rather than crash." [closure ~= objectMemory nilObject] whileTrue: [home := objectMemory followField: ClosureOuterContextIndex ofObject: closure. + (objectMemory isContext: home) ifFalse: - (self isContext: home) ifFalse: [^self internalCannotReturn: localReturnValue]. closure := objectMemory followField: ClosureIndex ofObject: home]. "home is to be returned from provided there is no unwind-protect activation between this frame and home's sender. Search for an unwind. findUnwindThroughContext: will answer either the context for an unwind-protect activation or nilObj if the sender cannot be found or 0 if no unwind is found but the sender is." unwindContextOrNilOrZero := self findUnwindThroughContext: home. unwindContextOrNilOrZero = objectMemory nilObject ifTrue: ["error: can't find home on chain; cannot return" ^self internalCannotReturn: localReturnValue]. unwindContextOrNilOrZero ~= 0 ifTrue: [^self internalAboutToReturn: localReturnValue through: unwindContextOrNilOrZero]. "Now we know home is on the sender chain. We could be returning to either a context or a frame. Find out which." contextToReturnTo := nil. (self isMarriedOrWidowedContext: home) ifTrue: [self assert: (self checkIsStillMarriedContext: home currentFP: localFP). theFP := self frameOfMarriedContext: home. (self isBaseFrame: theFP) ifTrue: [contextToReturnTo := self frameCallerContext: theFP] ifFalse: [frameToReturnTo := self frameCallerFP: theFP]] ifFalse: [contextToReturnTo := objectMemory fetchPointer: SenderIndex ofObject: home. ((objectMemory isContext: contextToReturnTo) and: [self isMarriedOrWidowedContext: contextToReturnTo]) ifTrue: [self assert: (self checkIsStillMarriedContext: contextToReturnTo currentFP: localFP). frameToReturnTo := self frameOfMarriedContext: contextToReturnTo. contextToReturnTo := nil]]. "If returning to a context we must make a frame for it unless it is dead." contextToReturnTo ~= nil ifTrue: [frameToReturnTo := self establishFrameForContextToReturnTo: contextToReturnTo. frameToReturnTo = 0 ifTrue: "error: home's sender is dead; cannot return" [^self internalCannotReturn: localReturnValue]]. "Now we have a frame to return to. If it is on a different page we must free intervening pages and nil out intervening contexts. We must free intervening stack pages because if we leave the pages to be divorced then their contexts will be divorced with intact senders and instruction pointers. This code is similar to primitiveTerminateTo. We must move any frames on intervening pages above the frame linked to because these may be in use, e.g. via co-routining (see baseFrameReturn)." self assert: stackPages pageListIsWellFormed. newPage := stackPages stackPageFor: frameToReturnTo. newPage ~~ stackPage ifTrue: [| currentCtx thePage nextCntx | currentCtx := self frameCallerContext: stackPage baseFP. stackPages freeStackPage: stackPage. [self assert: (objectMemory isContext: currentCtx). (self isMarriedOrWidowedContext: currentCtx) and: [(stackPages stackPageFor: (theFP := self frameOfMarriedContext: currentCtx)) = newPage]] whileFalse: [(self isMarriedOrWidowedContext: currentCtx) ifTrue: [thePage := stackPages stackPageFor: theFP. theFP ~= thePage headFP ifTrue: ["Since we've just deallocated a page we know that newStackPage won't deallocate an existing one." self moveFramesIn: thePage through: (self findFrameAbove: theFP inPage: thePage) toPage: stackPages newStackPage]. currentCtx := self frameCallerContext: thePage baseFP. stackPages freeStackPage: thePage] ifFalse: [nextCntx := objectMemory fetchPointer: SenderIndex ofObject: currentCtx. self markContextAsDead: currentCtx. currentCtx := nextCntx]]. self setStackPageAndLimit: newPage. localSP := stackPage headSP. localFP := stackPage headFP]. "Two cases. Returning to the top frame on a new page or an interior frame on the current page. The top frame has its instruction pointer on top of stack. An interior frame has its instruction pointer in the caller frame. We need to peel back any frames on the page until we get to the correct frame." localFP = frameToReturnTo ifTrue: "pop the saved IP, push the return value and continue." [localIP := self pointerForOop: self internalStackTop] ifFalse: [[callerFP := localFP. localFP := self frameCallerFP: localFP. localFP ~~ frameToReturnTo] whileTrue. localIP := self frameCallerSavedIP: callerFP. localSP := (self frameCallerSP: callerFP) - objectMemory wordSize]. self maybeReturnToMachineCodeFrame. self setMethod: (self frameMethod: localFP). self fetchNextBytecode. self internalStackTopPut: localReturnValue!

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev November 2018