December 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] IdentitySet but using #hash rather than #identityHash ?
by Henrik Sperre Johansen Dec. 16, 2011

Dec. 16, 2011

On 16.12.2011 00:40, Henrik Sperre Johansen wrote: > On 15.12.2011 23:35, Mariano Martinez Peck wrote: >> Well...it is much slower :( it seems that the cost of (aKey >> identityHash + ( aKey mareaClass identityHash bitShift: 12) + (aKey >> basicSize bitShift: 24) >> is bigger than the colisions. >> Anyway, thanks for the nice thread. I learned. >> >> Cheers >> > Well, first of all, that always creates a largeInteger, since > identityHash is already shifted by 22... You'd want to be using > basicIdentityHash. > 2nd off, basicSize is useless for non-variable classes, which I guess > is the common case. > 3rd, putting class hash in the high bits won't really help much if > you're serializing many instances of the same class, as they will all > occupy a sequential hash range. > > So if I were you, I'd try with something like > obj basicIdentityHash << 12 + (obj class basicIdentityHash) > before discarding it outright due to computation cost. > > Cheers, > Henry > Note: This is not portable to Squeak. They kept the old version of identityHash (renamed to basicIdentityHash in Pharo), and introduced scaledIdentityHash (doing the same as identityHash in Pharo), due to backwards compatability concerns for users explicitly relying on #identityHash returning a value in range [0, 4095]. Cheers, Henry

2 1

Re: [Pharo-project] IdentitySet but using #hash rather than #identityHash ?
by Henrik Sperre Johansen Dec. 16, 2011

Dec. 16, 2011

On 16.12.2011 01:06, Henrik Sperre Johansen wrote: > > Usual identitySet: > n := 1000000. > set := IdentitySet new: n. > Time millisecondsToRun: [1 to: n do: [:waste | set add: Object new.]] > 4595 > > Custom subclass with a slightly modified hash (anObject identityHash + > (anObject class basicIdentityHash bitShift: 6)): > n := 1000000. > set := PseudoIdentitySet2 new: n. > Time millisecondsToRun: [1 to: n do: [:waste | set add: Object new.]] > 3694 > Why 6 you ask? Because I'm worth it. > > Not really enough to justify the extra compexity, imho. > > Cheers, > Henry > To put that a bit more in perspective, the extra cost for not preallocating to a sane value: n := 1000000. set := IdentitySet new. Time millisecondsToRun: [1 to: n do: [:waste | set add: Object new.]] 6872 Cheers, Henry

1 0

[Pharo-project] UnixFileDirectory default directoryExists: UnixFileDirectory default asString, '/.'
by Eliot Miranda Dec. 15, 2011

Dec. 15, 2011

Hi All, just noticed that UnixFileDirectory default directoryExists: UnixFileDirectory default asString, '/.' answers false, since directoryNames excludes '.' and '..' (justifiable, and changing this might break all manner of things). i.e. FileDirectory>directoryExists: filenameOrPath "Answer true if a directory of the given name exists. The given name may be either a full path name or a local directory within this directory." "FileDirectory default directoryExists: FileDirectory default pathName" | fName dir | DirectoryClass splitName: filenameOrPath to: [:filePath :name | fName := name. filePath isEmpty ifTrue: [dir := self] ifFalse: [dir := self directoryNamed: filePath]]. ^dir exists and: [ self isCaseSensitive ifTrue:[dir directoryNames includes: fName] ifFalse:[dir directoryNames anySatisfy: [:name| name sameAs: fName]]]. So how about FileDirectory class methods for accessing currentDirectoryNickname ^'.' FileDirectory>directoryExists: filenameOrPath "Answer true if a directory of the given name exists. The given name may be either a full path name or a local directory within this directory." "FileDirectory default directoryExists: FileDirectory default pathName" | fName dir | DirectoryClass splitName: filenameOrPath to: [:filePath :name | fName := name. filePath isEmpty ifTrue: [dir := self] ifFalse: [dir := self directoryNamed: filePath]]. ^dir exists and: [fName = self class currentDirectoryNickname or: [self isCaseSensitive ifTrue:[dir directoryNames includes: fName] ifFalse:[dir directoryNames anySatisfy: [:name| name sameAs: fName]]]] -- best, Eliot

1 1

[Pharo-project] Technical "Debt" of Smalltalk with Moose? (was Re: An Improvement Over its Successors)
by Janko MivÅ¡ek Dec. 15, 2011

Dec. 15, 2011

Hi guys, I'm wondering if we can produce the same comparison of Technical Debt with Moose technology comparing Smalltalk with Java, .NET and others, as in study below? >From study: Technical debt is highest in Java-EE, .NET second. Technical Debt represents the effort required to fix problems that remain in the code when an application is released. It is shown as cost per LOC. This year is by their study 3.61 USD in average. S, Janko MivÅ¡ek piÅ¡e: > S, Casey Ransberger piÅ¡e: > >> I read recently that a study done on enterprise apps written in many different languages found the most defects per capita in Java apps. This didn't come as a surprise, but what did: the apps with the lowest defect rates were written in COBOL. > > This study can be found here (registration needed): > > The CRASH Report - 2011/12 > (CAST Report on Application Software Health) > http://www.castsoftware.com/resources/resource/whitepapers/cast-report-on-a… > > Interesting read, recommended, findings can help improve our software too! > > >> The researchers seemed to scratch their heads and proclaim "...well, I guess there've been people grinding bugs off of the COBOL code for thirty years." >> >> After the horror stories that I read about on comp.lang.cobol (I have strange hobbies? Also, some of those COBOL guys have a lot to teach) I think I nearly fell out of my chair laughing and achieved something approaching enlightenment;) >> > -- Janko MivÅ¡ek Aida/Web Smalltalk Web Application Server http://www.aidaweb.si

1 0

Re: [Pharo-project] NB Bug?
by Fernando Olivero Dec. 15, 2011

Dec. 15, 2011

Your right, this isnt a bug, NativeBoost is very well tested and works really nice! I should reformulate the comment: 'if i new user of Native Boost, and i copy and paste some Library method, to create a NB method in smalltalk, and if it happens that that particular method is written foo(void), then will i be able to know the origin of the DNU ?, that i should just put foo() ?" For me it took a while of debugging before i understood the issue, maybe others less eager to debug will think its a bug? Surely with a good documentation and examples, it shouldn't be a problem. Fernando On Thu, Dec 15, 2011 at 1:23 AM, Igor Stasenko <siguctua(a)gmail.com> wrote: > On 15 December 2011 00:12, Fernando Olivero <fernando.olivero(a)usi.ch> wrote: >> I've insist. >> >> Because maybe somebody else runs into this bug, and thinks that NB is >> broken, instead of understanding they should use new C syntax. For >> examples i was using the latest Pango code. > > but this is NOT a bug. maybe a missing feature but not a bug. > Is it so hard to put empty braces instead of void? :) And besides > empty braces it is valid C syntax. > >> >> If you want we could raise a warning, suggesting changing the code? >> (maybe too disturbing). >> > no, i just dont understand why complicating a code for nothing? > You can define a callout to functions with no arguments. > If you would not be able to do so, then this can be considered a bug. > >> Fernando >> >> On Wed, Dec 14, 2011 at 2:10 AM, Igor Stasenko <siguctua(a)gmail.com> wrote: >>> On 14 December 2011 01:17, Fernando Olivero <fernando.olivero(a)usi.ch> wrote: >>>> Igor, is this a bug? >>>> >>>> When parsing methods that send native functions withouth arguments , >>>> void, theres a DNU. >>>> For example in the following method: >>>> >>>> ^ self pangoCairoNbCall:#(PangoCairoFontMap >>>> pango_cairo_font_map_get_default ( void ) ) >>>> >>>> >>>> I fixed it like this: >>>> >>>> usesMethodArguments >>>> "first check wether my single argument is void" >>>> ^(fnSpec arguments size = 1 ==> [fnSpec arguments first class ~= >>>> NBVoid]) and:[ fnSpec arguments anySatisfy: [:type | type loader >>>> usesSTStack ] ] >>>> >>>> >>> >>> Well, the function signature syntax are not strictly following C. >>> Just use empty parens for functions with no arguments. This is quite >>> archaic C syntax, which is really not important to support. >>> >>> Or i can add this patch if you insist. :) >>> >>> -- >>> Best regards, >>> Igor Stasenko. >> > > > > -- > Best regards, > Igor Stasenko.

3 2

[Pharo-project] Magma in production ?
by Francois Stephany Dec. 15, 2011

Dec. 15, 2011

Is there anybody using the Magma OODB in production? For what kind of application? How does it behave under load? How large is your dataset? How do you perform external backup? What's the best setup? One dedicated image? Cheers, Francois -- http://tulipemoutarde.be BE: +32 (0)65 709 131 CA: +1 778 558 3225

1 0

Re: [Pharo-project] NB Bug?
by Fernando Olivero Dec. 15, 2011

Dec. 15, 2011

I've insist. Because maybe somebody else runs into this bug, and thinks that NB is broken, instead of understanding they should use new C syntax. For examples i was using the latest Pango code. If you want we could raise a warning, suggesting changing the code? (maybe too disturbing). Fernando On Wed, Dec 14, 2011 at 2:10 AM, Igor Stasenko <siguctua(a)gmail.com> wrote: > On 14 December 2011 01:17, Fernando Olivero <fernando.olivero(a)usi.ch> wrote: >> Igor, is this a bug? >> >> When parsing methods that send native functions withouth arguments , >> void, theres a DNU. >> For example in the following method: >> >> ^ self pangoCairoNbCall:#(PangoCairoFontMap >> pango_cairo_font_map_get_default ( void ) ) >> >> >> I fixed it like this: >> >> usesMethodArguments >> "first check wether my single argument is void" >> ^(fnSpec arguments size = 1 ==> [fnSpec arguments first class ~= >> NBVoid]) and:[ fnSpec arguments anySatisfy: [:type | type loader >> usesSTStack ] ] >> >> > > Well, the function signature syntax are not strictly following C. > Just use empty parens for functions with no arguments. This is quite > archaic C syntax, which is really not important to support. > > Or i can add this patch if you insist. :) > > -- > Best regards, > Igor Stasenko.

2 1

[Pharo-project] Better Gofer error messages
by Janko MivÅ¡ek Dec. 14, 2011

Dec. 14, 2011

Hi guys, Please please, let someone improve Gofer error messages, which are now totally non-understandable. Check in advance if some package is there and say "No package with that name", or even better, "no package with expected name: Blabla6.xxx" or something. What it is now is unusable, sorry. To prepare a right script is a big pain. Specially because there are nowhere rules how packages are resolved by a name. It will help if in exception at least a link to the some wiki with rules would be shown. Best regards Janko -- Janko MivÅ¡ek Svetovalec za informatiko Eranova d.o.o. Ljubljana, Slovenija www.eranova.si tel: 01 514 22 55 faks: 01 514 22 56 gsm: 031 674 565

3 7

[Pharo-project] [COTDC] 97 - CRCError
by sean＠clipperadams.com Dec. 14, 2011

Dec. 14, 2011

96 classes done 799 classes remaining Continuing in Compression... Today: CRCError Comment Of The Day Contest - One Day One Comment Rules: #1: Each day a not commented class is elected. Each day the best comment will be integrated with name of the author(s). #2: If you cannot comment it, deprecate it. Results: http://code.google.com/p/pharo/wiki/CommentOfTheDayContest p.s. this message was sent from a pharo image. Did you know you could easily do that?!

1 0

Dec. 14, 2011

14258 ----- - Issue 5101: better, cleaner, more tests for SharedPool http://code.google.com/p/pharo/issues/detail?id=5101 less pop up from script loaderâ¦. what a relief. Not as good as I wantâ¦.getting there. Stef

1 0