Squeak-dev June 2014

squeak-dev@lists.squeakfoundation.org

41 participants
191 discussions

85 10 195 197 down?
by gettimothy 18 Jun '14

18 Jun '14

Hi folks, Saving a monticello package fails do to timeout openConnectionToHost: hostIP port: portNumber | socket | socket := Socket new. socket connectTo: hostIP port: portNumber. ^self on: socket host is 85.10.195.197 port: 80 pinging the i.p did not work. thx tty

2 2

Exhuberant use of massive Traits on many sublcasses has an adverse effect on Monticello Performance
by gettimothy 17 Jun '14

17 Jun '14

Hi all. At least I think its the Trait. In CMakeVMakerSqueak, I am (ab)using a Trait to enforce pseudo-abstract base class behavior. Everything is fine in my development environment. There is some sluggishness when adding a new method or updating a Trait method as that change has to 'percolate' to all the classes (that would be most of them) in the CMakeVMMakerSqueak package. For a coding idiom, it works very nicely as it enforces implementation discipline. Today, I updated the Cog SVN tree locally and retested the install scripts I use and the CMakeVMMakerSqueak-tty.63 appeared to hang on installation into a fresh 4.5 image. I thought at first that I broke something, so, I started importing from -tty.20 forward and I noticed that at -tty32, loading became very slow. I am currently loaded -tty.52 and performance is un-acceptable. I think it is the Trait! I know that Traits are to be used as "little plugins" to attach to "the side" of a thing--I call it "sideways abstract base class" stuff, and that is what I have done. Only, I have done it on a massive scale. How massive? Well, here are the methods I am setting as must be overridden in ~150 subclasses. ('accessing' executableName executableType frameworks generate generateLicense generatePluginsList linkFlags version) ('cmake configuration' availableBuilds buildType: configureBuild configureBuildAssert configureBuildAssertITimerHeartbeat configureBuildDebug configureBuildDebugITimerHeartbeat configureBuildDebugMultiThreaded configureBuildIHeartbeatTimer configureBuildMultiThreaded configureBuildMultiThreadedAssert configureBuildMultiThreadedDebug configureForBuildType: configureNoBuildType excludeFromBuild) ('cmake config build type ' compilerFlagsAssert compilerFlagsAssertITimerHeartbeat compilerFlagsBuild compilerFlagsDebug compilerFlagsDebugITimerHeartbeat compilerFlagsIHeartbeatTimer compilerFlagsMultiThreaded compilerFlagsMultiThreadedAssert compilerFlagsMultiThreadedDebug compilerFlagsNoBuildType externalPluginsBuild externalPluginsBuildAssert externalPluginsBuildAssertITimerHeartbeat externalPluginsBuildDebug externalPluginsBuildDebugITimerHeartbeat externalPluginsBuildITimerHeartbeat externalPluginsBuildMultiThreaded externalPluginsBuildMultiThreadedAssert externalPluginsBuildMultiThreadedDebug externalPluginsBuildNone internalPluginsBuild internalPluginsBuildAssert internalPluginsBuildAssertITimerHeartbeat internalPluginsBuildDebug internalPluginsBuildDebugITimerHeartbeat internalPluginsBuildITimerHeartbeat internalPluginsBuildMultiThreaded internalPluginsBuildMultiThreadedAssert internalPluginsBuildMultiThreadedDebug internalPluginsBuildNone preferredIncludesBuild preferredIncludesBuildAssert preferredIncludesBuildAssertITimerHeartbeat preferredIncludesBuildDebug preferredIncludesBuildDebugITimerHeartbeat preferredIncludesBuildITimerHeartbeat preferredIncludesBuildMultiThreaded preferredIncludesBuildMultiThreadedAssert preferredIncludesBuildMultiThreadedDebug preferredIncludesBuildNone setExtraTargetPropertiesBuild: setExtraTargetPropertiesBuildAssert: setExtraTargetPropertiesBuildAssertITimerHeartbeat: setExtraTargetPropertiesBuildDebug: setExtraTargetPropertiesBuildDebugITimerHeartbeat: setExtraTargetPropertiesBuildITimerHeartbeat: setExtraTargetPropertiesBuildMultiThreaded: setExtraTargetPropertiesBuildMultiThreadedAssert: setExtraTargetPropertiesBuildMultiThreadedDebug: setExtraTargetPropertiesBuildNone: setGlobalOptionsAfterDetermineSystemBuild: setGlobalOptionsAfterDetermineSystemBuildAssert: setGlobalOptionsAfterDetermineSystemBuildAssertITimerHeartbeat: setGlobalOptionsAfterDetermineSystemBuildDebug: setGlobalOptionsAfterDetermineSystemBuildDebugITimerHeartbeat: setGlobalOptionsAfterDetermineSystemBuildITimerHeartbeat: setGlobalOptionsAfterDetermineSystemBuildMultiThreaded: setGlobalOptionsAfterDetermineSystemBuildMultiThreadedAssert: setGlobalOptionsAfterDetermineSystemBuildMultiThreadedDebug: setGlobalOptionsAfterDetermineSystemBuildNone setGlobalOptionsBuild: setGlobalOptionsBuildAssert: setGlobalOptionsBuildAssertITimerHeartbeat: setGlobalOptionsBuildDebug: setGlobalOptionsBuildDebugITimerHeartbeat: setGlobalOptionsBuildITimerHeartbeat: setGlobalOptionsBuildMultiThreaded: setGlobalOptionsBuildMultiThreadedAssert: setGlobalOptionsBuildMultiThreadedDebug: setGlobalOptionsBuildNone: standardIncludesBuild standardIncludesBuildAssert standardIncludesBuildAssertITimerHeartbeat standardIncludesBuildDebug standardIncludesBuildDebugITimerHeartbeat standardIncludesBuildITimerHeartbeat standardIncludesBuildMultiThreaded standardIncludesBuildMultiThreadedAssert standardIncludesBuildMultiThreadedDebug standardIncludesBuildNone thirdpartyLibsBuild thirdpartyLibsBuildAssert thirdpartyLibsBuildAssertITimerHeartbeat thirdpartyLibsBuildDebug thirdpartyLibsBuildDebugITimerHeartbeat thirdpartyLibsBuildITimerHeartbeat thirdpartyLibsBuildMultiThreaded thirdpartyLibsBuildMultiThreadedAssert thirdpartyLibsBuildMultiThreadedDebug thirdpartyLibsBuildNone) ('cmake directory ' dirARMv6 dirAndroid dirBSD32x86 dirBuild dirBuildAssert dirBuildAssertITimerHeartbeat dirBuildDebug dirBuildDebugITimerHeartbeat dirBuildDebugMultiThreaded dirBuildITimerHeartbeat dirBuildLanguageVMMM dirBuildMultiThreaded dirBuildMultiThreadedAssert dirBuildMultiThreadedDebug dirBuildPlatform dirIA32Bochs dirIOS dirInstall dirLinux32Armv6 dirLinux32x86 dirLinux32x8664 dirMacOS dirMacOSPowerPC dirMacOSX32x86 dirOutput dirSource dirSunOS32x86 dirWin32x86 newspeakCogSpur newspeakCogV3 newspeakSistaSpur newspeakSistaV3 newspeakStackSpur newspeakStackV3 oscogvm outputDirName squeakCogSpur squeakCogV3 squeakSistaSpur squeakSistaV3 squeakStackSpur squeakStackV3) ('cmake executable names' eventStackV3 vmCogExecutableName vmSistaExecutableName vmStackExecutableName) ('compiler flags' commonCompilerFlags compilerFlags) ('headers generation' configH configHBuild configHBuildAssert configHBuildAssertITimerHeartbeat configHBuildDebug configHBuildDebugITimerHeartbeat configHBuildITimerHeartbeat configHBuildMultiThreaded configHBuildMultiThreadedAssert configHBuildMultiThreadedDebug configHNoBuildType generateConfigH generateConfigH:) ('plugin extra rules' extraPluginSettings:) ('settings' extraVMSettings:) ('source files' preferredIncludes) ('source generation' cogitClass interpreterClass prepareForGeneration) ('squeak compatibility' buildDir buildDirName dirFrom: outputDir setupDirectories: srcDir topDir validateSourcesPresent write:toFile:) ('utils' setExtraTargetProperties: setGlobalOptions: setGlobalOptions:buildType: setGlobalOptionsAfterDetermineSystem:buildType:) ('cmake configurations') So, I have a big-ass Trait, used by ~150 subclasses and I "think" that is what is taking the long install time. Now, you are thinking to yourself, 'tty! what the heck where you thinking!" Good question! This CMakeVMMaker port to squeak is NOT to be a fork. The existing CMakeVMaker inheritence tree precluded me from interjecting a thin SqueakCompatibility class under CPlatformConfig. So, I used a Trait to to get the same effect. As it became apparent that the Squeak implementation required more functionality than the Pharo implementation has, the Trait grew and grew.. Anyway, the point of the email is to let people know that Massive Traits used with Wild Abandon have and adverse effect on Monticello Performance cheers, tty

1 0

Re: [Pharo-dev] Spur Squeak Trunk Image Available
by Eliot Miranda 17 Jun '14

17 Jun '14

Hi Philippe, On Tue, Jun 17, 2014 at 1:03 AM, Philippe Marschall < philippe.marschall(a)netcetera.ch> wrote: > On 13.06.14 01:41, Eliot Miranda wrote: > >> >> >> >> >> Hi All, >> >> it gives me great pleasure to let you know that a spur-format trunk >> Squeak image is finally available at >> http://www.mirandabanda.org/files/Cog/SpurImages/. Spur VMs are >> available at http://www.mirandabanda.org/files/Cog/VM/VM.r2987/. >> >> … >> > > I'm seeing a Seaside request handling benchmark going from 10k req/s to > 11k req/s. > Don't be too quick to dismiss this as being IO-bound (being IO-bound is > actually quite hard on Squeak/Pharo). During the benchmark Squeak fully > saturates one core. It is hard to tell what the limiting factor for this > benchmark actually is. But removing one or two String allocations from the > request handling loop usually yields about 100 to 200 additional req/s. > That's good news at least. One thing you can try is the VMProfiler. It's an interactive Morphic application so it will try and open and may crash Load the VMProfiler and use it via VMProfiler openInstance spyOn: [...]; report: aStream. e.g. VMProfiler openInstance spyOn: [1 tinyBenchmarks]; report: (Transcript cr; yourself). Transcript flush. produces: /Users/eliot/Cog/oscogvm/build.macos32x86/squeak.cog.spur/Fast.app/Contents/MacOS/Squeak 6/17/2014 eden size: 2,603,344 stack pages: 160 code size: 1,048,576 1 tinyBenchmarks gc prior. clear prior. 6.298 seconds; sampling frequency 1473 hz 9231 samples in the VM (9275 samples in the entire program) 99.53% of total 8798 samples in generated vm code 95.31% of entire vm (94.86% of total) 433 samples in vanilla vm code 4.69% of entire vm ( 4.67% of total) % of generated vm code (% of total) (samples) (cumulative) 61.29% (58.13%) Integer>>benchFib (5392) (61.29%) 13.47% (12.78%) Integer>>benchmark (1185) (74.76%) 11.48% (10.89%) Object>>at:put: (1010) (86.24%) 9.60% ( 9.11%) SmallInteger>>+ (845) (95.84%) 3.99% ( 3.78%) Object>>at: (351) (99.83%) 0.07% ( 0.06%) ceBaseFram...Trampoline(6) (99.90%) 0.03% ( 0.03%) ceMethodAbort0Args (3) (99.93%) 0.03% ( 0.03%) Sequenceabl...om:to:put: (3) (99.97%) 0.02% ( 0.02%) Magnitude>>min: (2) (99.99%) 0.01% ( 0.01%) ceEnterCog...ceiverReg (1) (100.0%) % of vanilla vm code (% of total) (samples) (cumulative) 32.33% ( 1.51%) primitiveStringReplace (140) (32.33%) 24.25% ( 1.13%) copyAndForward (105) (56.58%) 13.63% ( 0.64%) marryFrameSP (59) (70.21%) 10.85% ( 0.51%) instantiateClassindexableSize (47) (81.06%) 3.46% ( 0.16%) ceBaseFrameReturn (15) (84.53%) 2.08% ( 0.10%) scavengeLoop (9) (86.61%) 1.62% ( 0.08%) checkForEvent...ontextSwitch (7) (88.22%) 1.39% ( 0.06%) externalEnsureIsBaseFrame (6) (89.61%) 1.39% ( 0.06%) voidVMStateForSn...nalPrimitivesIf(6) (90.99%) 1.15% ( 0.05%) mapInterpreterOops (5) (92.15%) 1.15% ( 0.05%) returnToExecuti...tContextSwitch(5) (93.30%) 1.15% ( 0.05%) scavengeReferentsOf (5) (94.46%) 0.92% ( 0.04%) handleStackOverflow (4) (95.38%) 0.92% ( 0.04%) signed64BitIntegerFor (4) (96.30%) 0.46% ( 0.02%) primitiveShortAt (2) (96.77%) 0.46% ( 0.02%) ceNonLocalReturn (2) (97.23%) 0.23% ( 0.01%) gMoveCwR (1) (97.46%) 0.23% ( 0.01%) genInnerPrimit...CharacterinReg (1) (97.69%) 0.23% ( 0.01%) genPrimitiveMultiply (1) (97.92%) 0.23% ( 0.01%) primitiveInflateDecompressBlock(1) (98.15%) 0.23% ( 0.01%) primitiveNewWithArg (1) (98.38%) 0.23% ( 0.01%) primitiveTestAnd...fCriticalSection(1) (98.61%) 0.23% ( 0.01%) processWeaklings (1) (98.85%) 0.23% ( 0.01%) scavengingGCTenuringIf (1) (99.08%) 0.23% ( 0.01%) updateMaybeObjRefAt (1) (99.31%) 0.23% ( 0.01%) wakeHighestPriority (1) (99.54%) 0.23% ( 0.01%) ceCannotResume (1) (99.77%) 0.23% ( 0.01%) ioPositionOfWindowSetxy (1) (100.0%) **Memory** old +574,744 bytes **GCs** full 0 totalling 0ms (0% elapsed time) scavenges 330 totalling 93ms (1.477% elapsed time), avg 0.282ms tenures 0 root table 0 overflows **Compiled Code Compactions** 0 totalling 0ms (0% elapsed time) **Events** Process switches 66 (10 per second) ioProcessEvents calls 314 (50 per second) Interrupt checks 3462 (550 per second) Event checks 3585 (569 per second) Stack overflows 421897 (66989 per second) Stack page divorces 0 (0 per second) While this profile is flat, it does show where the VM is spending its time, and since the VM is not something with deep call chains, this view is the most useful I have. The VMProfiler is part of the CogTools package at http://source.squeak.org/VMMaker. P.S. Extending the size of the DefaultTabsArray will result in a less cramped report. HTH Eliot

1 0

Re: [squeak-dev] Building SqueakVM on Samsung ARM7, Chromebook????(ChrUbuntu)
by Ken.Dickey 17 Jun '14

17 Jun '14

> What are you trying to accomplish? Do you already run Squeak and want to compile your own VM? or are you just getting started with Squeak? I am running Cuis on the RasPi VM (ARM6). I would like to build the VM for the ARM7 and eventually have access to quicker VMs (Cog) and the Samsung Exynos GPU. -KenD

1 0

Building SqueakVM on Samsung ARM7 Chromebook (ChrUbuntu)
by Ken.Dickey 17 Jun '14

17 Jun '14

Greetings, I am trying to build a SqueakVM. I have tried both the current subversion branch and the 4.10.2-2614-src tarball. Both fail consistently. (sq-make.out attached) I have not done makefiles or C in many years. Can some kind person help me out? Thanks much! -KenD

3 3

Cog 3000, libcrypto, and Altitude
by Chris Cunnington 17 Jun '14

17 Jun '14

I downloaded 3000 versions of coglinux, coglinuxht, cogmtlinuxht to my Ubuntu laptop. [1] I wanted to see if I could use the Altitude web framework without creating a symlink to the libcrypto library [2]. I think the 3000 series looks for libcrypto.so.1.0.0, which I think I have installed. [3]. The coglinux image starts, but I get the standard Module Not Found error. [5]. This is what the symlink solves. The other two VMs, coglinuxht and cogmtlinuxht, don't start. [4] Cheers, Chris [1] Ubuntu 12.04 LTS Intel Core i7-3687u CPU @ 2.10 GHzx4 [2] chris@ooplu:~/Desktop/Cog/lib/squeak/4.0-2761$ file libcrypto.so libcrypto.so: symbolic link to `/usr/lib/i386-linux-gnu/libcrypto.so.0.9.8' [3] chris@ooplu:~/Desktop/Cog/lib/squeak/4.0-2761$ locate libcrypto /home/chris/.local/share/Trash/files/libcrypto.so.0.9.8 /home/chris/.local/share/Trash/info/libcrypto.so.0.9.8.trashinfo /lib/i386-linux-gnu/libcrypto.so.0.9.8 /lib/i386-linux-gnu/libcrypto.so.1.0.0 /lib/x86_64-linux-gnu/libcrypto.so.1.0.0 /usr/lib/i386-linux-gnu/libcrypto.so.0.9.8 /usr/lib/x86_64-linux-gnu/libcrypto.a /usr/lib/x86_64-linux-gnu/libcrypto.so /usr/lib/x86_64-linux-gnu/pkgconfig/libcrypto.pc [4] chris@ooplu:~/Desktop/coglinux$ ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' & [1] 5932 chris@ooplu:~/Desktop/coglinux$ cd ../coglinuxht [1]+ Done ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' (wd: ~/Desktop/coglinux) (wd now: ~/Desktop/coglinuxht) chris@ooplu:~/Desktop/coglinuxht$ ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' & [1] 5949 chris@ooplu:~/Desktop/coglinuxht$ pthread_setschedparam failed; consider using ITIMER_HEARTBEAT: Operation not permitted ^C [1]+ Exit 1 ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' chris@ooplu:~/Desktop/coglinuxht$ cd ../cogmtlinuxht chris@ooplu:~/Desktop/cogmtlinuxht$ ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' & [1] 5962 chris@ooplu:~/Desktop/cogmtlinuxht$ warning, processHasThreadId flag is unset; cannot function as a threaded VM if so. pthread_setschedparam failed; consider using ITIMER_HEARTBEAT: Operation not permitted ^C [1]+ Exit 1 ./squeak '/home/chris/Desktop/ squeak.org/Squeak4.4-12327.image' [5] Error: External module not found 15 June 2014 12:15:47.612 pm VM: unix - Smalltalk Image: Squeak4.4 [latest update: #12327] SecurityManager state: Restricted: false FileAccess: true SocketAccess: true Working Dir /home/chris/Desktop/squeak.org Trusted Dir /home/chris/Desktop/squeak.org/secure Untrusted Dir /home/chris/Desktop/squeak.org/My Squeak XTLibCryptoHash(Object)>>error: Receiver: a XTLibCryptoHash Arguments and temporary variables: aString: 'External module not found' Receiver's instance variables: library: nil context: a XTEVPMDCTX algorithm: nil XTLibCryptoHash(Object)>>externalCallFailed Receiver: a XTLibCryptoHash Arguments and temporary variables: errCode: 15 Receiver's instance variables: library: nil context: a XTEVPMDCTX algorithm: nil XTLibCryptoHash>>EVP_MD_CTX_init: Receiver: a XTLibCryptoHash Arguments and temporary variables: <<error during printing> Receiver's instance variables: library: nil context: a XTEVPMDCTX algorithm: nil

2 1

ObjectExplorer and Text instances
by karl ramberg 16 Jun '14

16 Jun '14

Hi, While inspecting and exploring at various Text instances I noticed that the inspector shows the instance variables to Text instances (string and runs) but explorer just show the string directly, not as a instance variable, and runs is not shown at all. Anybody know why that is ? Cheers, Karl

2 2

New Cog & Spur VMs available
by Eliot Miranda 16 Jun '14

16 Jun '14

in http://www.mirandabanda.org/files/Cog/VM/VM.r3003/. Apologies for the non-Spur VMs in r3000 which should have had the "relocating call to invalid address" bug fixed but I made a mistake. It is fixed in 3003. Further, those who got crashes running the test suite due to the Aio tests should now find the VM does not crash. CogVM source as per VMMaker.oscog-eem.774/r3003 Remember to regenerate the cogit.c's from the non-Spur configurations to fix the "relocating call to invalid address" bug. Add a hack to link the XDisplayControlPlugin against vm-display-X11 (a hack because it assumes paths etc). But at least the plugin works. Add an initial revision of platforms/unix/plugins/UUIDPlugin/acinclude.m4 to search for uuid.h in either /usr/include/uuid or /usr/include and use either uuidgenerate or uuidgen. -- best, Eliot

2 1

a Cog BASH install script, please (as opposed to using 'make install')
by Chris Cunnington 16 Jun '14

16 Jun '14

The short answer (could be wrong): Almost nobody can compile Cog from sources. That means that almost nobody can install Cog in the appropriate way on Linux. This means a BASH script that does what "make install" does would be a good thing. The long answer (still could be wrong): I ranted recently about installer scripts for VMs. Afterward I wasn't even sure what I was talking about. I was sure I'd seen two expert programmers have a hellish time trying to install Cog on Linux servers. I don't have that problem, because I don't install. I just put the vm on the command line and use. When I suggested to these programmers that they should do the same, if they could have reached through the screen, they would have slapped me. But what was going on? They both tried to compile Cog from sources, not because it's a fun thing to do, but in hindsight, I think they were doing it, because they thought it was the only way to install it on Linux. You need to execute: make before you can install: make install If they could have just done the second part, I think they would have. But I think it necessitated the first part. Perhaps pre-compiled Cog binaries should come with a Makefile for "make install"? I remember watching one programmer do it and talk on Squeak IM at the same time. The consensus with another expert level programmer was that Eliot's way ( Hi, Eliot !) of building VMs was particular to him and not really a process for wide consumption. What does that mean? I have no idea. If you want to comb through the Squeak IM transcripts of the last four-five months, you'll find it. Eliot releases his VMs, which are wonderful, with a script called squeak. I think that's just for starting images on the command line. I went back to r2151 and it comes with two scripts - squeak and squeak.sh. They appear to do the same thing. Then I went back to r1596 to see how I successfully installed a VM on a Linux box in 2006. I think anybody could use "make" and then "make install" back then. I don't think it came with a script for starting on the command line. On Friday I made the squeak.org homepage run on CogSpur. I did it by doing horrible things to daemontools. *shiver* But I don't think I can install any kind of Cog linux VM on a server. I've seen two experts try and it was an ugly sight. Perhaps this explanation just shows my ignorance and somebody will point out what I don't understand. I'm ready for that. But if not, if the only way to install Cog is to compile Cog, then Houston, we have a problem. Chris

3 3

Daily Commit Log
by commits＠source.squeak.org 16 Jun '14

16 Jun '14

Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours: http://lists.squeakfoundation.org/pipermail/packages/2014-June/007073.html Name: Kernel-dtl.856 Ancestors: Kernel-cmm.855 Fix inconsistent image state as reported on squeak-dev: Date: Wed, 11 Jun 2014 23:44:42 -0400 From: Florin Mateoc To: The general-purpose Squeak developers list Subject: [squeak-dev] Class subclasses I think this is an old bug, presumably since ProtoObject was introduced (and presumably because ObjectTracer was used in the process). If you ask Class for its subclasses (or if you inspect it) you see that ObjectTracer class is listed along with ProtoObject class, although ObjectTracer's superclass is ProtoObject. This can lead to some funny bugs if you try to write some hierarchy traversing code. =============================================

1 0

← Newer
1
...
11
12
13
14
15
16
17
...
20
Older →

Jump to page: