- Pharo-dev - lists.pharo.org

[Pharo-project] Pharo 1.4 Tests-Unix - Build # 294 - Still Unstable!
by jenkins-pharo.ci.inria.fr＠jenkins-server.lille.inria.fr March 19, 2012

March 19, 2012

1 0

Re: [Pharo-project] Zodiac is cool!
by Marcus Denker March 19, 2012

March 19, 2012

On Mar 19, 2012, at 11:26 AM, Norbert Hartl wrote: > I thought using zodiac might be a fragile thing to do. Expecting library or other version trouble. But I just copied the plugin in my vm directory, loaded the code and it worked instantaneously. That just made pharo feel quite a bit more modern :) I would like to see Zodiac staying that way or being part of a pharo distribution whatever it might be. > We want to ship Pharo and the VMs with Zodiac by default... it's on the TODO. > Great work, Indeed! Marcus -- Marcus Denker -- http://marcusdenker.de

1 0

[Pharo-project] [update 1.4] #14400
by Marcus Denker March 19, 2012

March 19, 2012

14400 ----- Issue 5513: remove InputEventSensor>>testJoystick: http://code.google.com/p/pharo/issues/detail?id=5513 Issue 5510: Methods in category objects from disk in SmalltalkImage should be re - categorized http://code.google.com/p/pharo/issues/detail?id=5510 Issue 5507: test RGClassDefinitionTest >> testWithComment failure / TimeStamp comparison http://code.google.com/p/pharo/issues/detail?id=5507 Issue 5504: To have better common interface between TranscriptStream and ThreadSafeTranscript http://code.google.com/p/pharo/issues/detail?id=5504 -- Marcus Denker -- http://marcusdenker.de

1 0

Re: [Pharo-project] What fixed the tests? :-)
by Marcus Denker March 19, 2012

March 19, 2012

On Mar 18, 2012, at 11:17 PM, Pavel Krivanek wrote: > Hi, > > what fixed the tests on Linux CI in 14287 so instead of 39 failures we > had only 4? Fixes in CogVM? > I changed the test that creates the file with Ã¤Ã¼ in the name and put in on this tracker entry: http://code.google.com/p/pharo/issues/detail?id=2438 (the report is from 2010...) So the problem is not fixed... but the idea is that failing test are in the tracker close to the report, while the tests in the release are green. Marcus -- Marcus Denker -- http://marcusdenker.de

1 0

[Pharo-project] ConnectionQueue broken?
by Schwab,Wilhelm K March 19, 2012

March 19, 2012

Hello all, I did some long overdue networking, and tried ConnectionQueue in 1.3. My changed version, that appears to work, is below. The key to the potential fix is in adding the #yourself to this line: newConnection := socket waitForConnectionFor: 10; yourself. #waitForConnectionFor: returns true or signals an error, the true ends up replacing the socket with a boolean and badness results. Can anyone confirm this? Bill listenLoop "Private! This loop is run in a separate process. It will establish up to maxQueueLength connections on the given port." "Details: When out of sockets or queue is full, retry more frequently, since a socket may become available, space may open in the queue, or a previously queued connection may be aborted by the client, making it available for a fresh connection." "Note: If the machine is disconnected from the network while the server is running, the currently waiting socket will go from 'isWaitingForConnection' to 'unconnected', and attempts to create new sockets will fail. When this happens, delete the broken socket and keep trying to create a socket in case the network connection is re-established. Connecting and disconnecting was tested under PPP on Mac system 8.1. It is not if this will work on other platforms. Fixed to not accept the connection if the queue is full (gvc)." | newConnection | socket := Socket newTCP. "We'll accept four simultanous connections at the same time" socket listenOn: portNumber backlogSize: 4. "If the listener is not valid then the we cannot use the BSD style accept() mechanism." socket isValid ifFalse: [^self oldStyleListenLoop]. [true] whileTrue: [ socket isValid ifFalse: [ "socket has stopped listening for some reason" socket destroy. (Delay forMilliseconds: 10) wait. ^self listenLoop ]. [newConnection := socket waitForConnectionFor: 10; yourself. ] on: ConnectionTimedOut do: [:ex | newConnection := nil]. (newConnection notNil and: [newConnection isConnected]) ifTrue: [(accessSema critical: [connections size < maxQueueLength]) ifTrue: [newConnection := socket accept] ifFalse: [socket closeAndDestroy: 0]] ifFalse: [newConnection := nil]. (newConnection notNil and: [newConnection isConnected]) ifTrue: [ accessSema critical: [connections addLast: newConnection]. newConnection := nil. self changed]. self pruneStaleConnections].

2 1

[Pharo-project] Understanding RunArray>>scanFrom:
by Guillermo Polito March 18, 2012

March 18, 2012

This method parses a stream from a fileout with styling and generates the text style specific objects. Now, 1) are we writing that style section in our fileouts? 2) Is this really necessary in RunArray? It's class comment does not say anything to me about Text nor TextStyle or stuff: "My instances provide space-efficient storage of data which tends to be constant over long runs of the possible indices. Essentially repeated values are stored singly and then associated with a "run" length that denotes the number of consecutive occurrences of the value." The method is the following: scanFrom: strm "Read the style section of a fileOut or sources file. nextChunk has already been done. We need to return a RunArray of TextAttributes of various kinds. These are written by the implementors of writeScanOn:" | runs values attrList char | (strm peekFor: $( ) ifFalse: [^ nil]. runs := OrderedCollection new. [strm skipSeparators. strm peekFor: $)] whileFalse: [runs add: (Number readFrom: strm)]. values := OrderedCollection new. "Value array" attrList := OrderedCollection new. "Attributes list" [(char := strm next) == nil] whileFalse: [ char == $, ifTrue: [values add: attrList asArray. attrList := OrderedCollection new]. char == $a ifTrue: [attrList add: (TextAlignment new alignment: (Integer readFrom: strm ifFail: [0]))]. char == $f ifTrue: [attrList add: (TextFontChange new fontNumber: (Integer readFrom: strm ifFail: [0]))]. char == $F ifTrue: [attrList add: (TextFontReference toFont: (Smalltalk at: #StrikeFont ifPresent: [:sf | sf familyName: (strm upTo: $#) size: (Integer readFrom: strm ifFail: [0])]))]. char == $b ifTrue: [attrList add: (TextEmphasis bold)]. char == $i ifTrue: [attrList add: (TextEmphasis italic)]. char == $u ifTrue: [attrList add: (TextEmphasis underlined)]. char == $= ifTrue: [attrList add: (TextEmphasis struckOut)]. char == $n ifTrue: [attrList add: (TextEmphasis normal)]. char == $- ifTrue: [attrList add: (TextKern kern: -1)]. char == $+ ifTrue: [attrList add: (TextKern kern: 1)]. char == $c ifTrue: [attrList add: (TextColor scanFrom: strm)]. "color" char == $L ifTrue: [attrList add: (TextLink scanFrom: strm)]. "L not look like 1" char == $R ifTrue: [attrList add: (TextURL scanFrom: strm)]. "R capitalized so it can follow a number" char == $P ifTrue: [attrList add: (TextPrintIt scanFrom: strm)]. char == $d ifTrue: [attrList add: (TextDoIt scanFrom: strm)]. "space, cr do nothing" ]. values add: attrList asArray. ^ self runs: runs asArray values: (values copyFrom: 1 to: runs size) asArray " RunArray scanFrom: (ReadStream on: '(14 50 312)f1,f1b,f1LInteger +;i') " Guille

1 1

[Pharo-project] longer builds history
by Pavel Krivanek March 18, 2012

March 18, 2012

Hi, I extended the history of CI job Pharo 1.4 Tests-Unix to 60 builds/days. I wanted to find when testStWriter tests started to fail and the history was too short (7). This job takes 2MB so I hope it is not a problem. The two tests failures Graphics.Tests.Primitives.BitmapTest.testCompressToByteArrayIntegrity Graphics.Tests.Primitives.BitmapTest.testCompressToByteArraySize seem to be only accidental The tests Ring.Tests.Kernel.RGClassDefinitionTest.testWithComment Ring.Tests.Kernel.RGCommentDefinitionTest.testNewComment have a fix here: http://code.google.com/p/pharo/issues/detail?id=5507 Cheers, -- Pavel

1 1

[Pharo-project] What fixed the tests? :-)
by Pavel Krivanek March 18, 2012

March 18, 2012

Hi, what fixed the tests on Linux CI in 14287 so instead of 39 failures we had only 4? Fixes in CogVM? -- Pavel

1 1

[Pharo-project] squeak tutorials metahowto, parts 1 and 2, find all comments in all classes in squeak:
by Lawson English March 18, 2012

March 18, 2012

metahowto, parts 1 and 2, find all comments in all classes in squeak: http://www.youtube.com/watch?v=rADmEwHcMDI http://www.youtube.com/watch?v=VeSnbA_9Rds assumes you have already watched the rest of the series. Lawson

1 0

[Pharo-project] Display extent during startUp sequence
by Hilaire Fernandes March 18, 2012

March 18, 2012

Hello, I need to know the Display extent during the startUp sequence. Sadly under CogDroid I don't get it right at this time. I am suspect the underneath system is sendeing a fullscreen order to the VM window only propagated after the startUp sequence. I tried "Display forceDisplayUpdate". I got a workaround with a forked delay in my start up sequence but unclean. Thanks Hilaire -- Dr. Geo -- http://www.drgeo.eu

1 0