Squeak-dev March 2007

squeak-dev@lists.squeakfoundation.org

143 participants
204 discussions

A few good Sophie Books needed, or 982 days of Sophie.
by John M McIntosh 22 Mar '07

22 Mar '07

Seems the Sophie Marketing Team (tm) has left the office, not that I could tell if there was one, seems to be always the case on these open source projects. Someone says why don't the developers make some books? Right... Give me a graphite pencil, I'll wonder where the wood came from, was it glued or lathed, is the yellow paint (pantone color?) eatable, and is that natural rubber or synthetic? Draw nicely shaded etching of a landscape? No way. Marker pen and stick diagrams maybe. So we need some folks who actually have artistic talents to make a few small books showing off capabilities, their cleverness, heck they'll even get free advertising, hand holding from the original developers, priceless... As a benefit you get to expose what funky bugs we have in the system after 982 calendar days (many *long* days) of work, 8,156 email messages, 7,234 mantis emails, 2,523 mantis entries (mostly all closed, that last 10% well...). Oh and countless wiki pages lurking somewhere during the design phase, and likely 10's of thousands lines of code we tossed to flash memory never to speak of again. That's my stats, other team members likely have larger numbers. But oddly enough not many nice looking books on the server, lots of lame test books lurk of course, something about marker pens and stick diagrams come to mind. After pushing I think it's about 500 downloads in the last 24 hours, number one complaint has been: Not enough examples. http://www.sophieproject.org Mantis bugs to close, ya, example books from me? Not likely... -- ======================================================================== === John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===

1 0

Closing bug 3537
by Damien Cassou 21 Mar '07

21 Mar '07

Hi, can someone close http://bugs.squeak.org/view.php?id=3537 ? Thank you -- Damien Cassou

2 1

Bug reports on collections
by Damien Cassou 21 Mar '07

21 Mar '07

Hi, while writing tests for the Collection hierarchy, I found bugs or differences with ANSI. I then wrote to the mailing list and opened the following bug reports. This bug reports contained fixes and tests to be included in future Squeak releases. Bug 2273: PositionableStream>>isEmpty is wrong Bug 6352: PositionableStream>>peekFor is not ANSI Bug 6350: Generalizes #ensureASpace with #ensureEndsWith: Bug 6351: Implements String class>>space to be coherent with other instance creation methods Bug 6070: Nicolas Cellier's implementation of RunArray>>at:put: Please include this enhancements/fixes/tests into the next release. -- Damien Cassou

1 0

Lots of tests for Collections, Numbers and Compiler
by Damien Cassou 21 Mar '07

21 Mar '07

Hi, I'm proud to announce 99 more tests for Collections (arrays, streams...), Numbers (random, integer...) and Compiler. They are all passing tests and thus can be added to 3.10 and 3.9.1. Load packages CollectionsTest, Compiler and KernelTests from: http://www.squeaksource.com/SqueakBaseClasses/ Do NOT load the package Collections which contains fixes handled by bug reports. See my following mail concerning this bug reports. Hope my work won't get lost. -- Damien Cassou

1 0

PositionnableStream>>skip: not standard (bug 6353)
by Damien Cassou 21 Mar '07

21 Mar '07

http://bugs.squeak.org/view.php?id=6353 In the standard, #skip: should explicetly not raise an error if #skip: would go after the end. It is said that "A number of objects equal to the lesser of amount and the size of the receiver's future sequence values are removed..." The current implementation raises an error. Do we choose the current implementation our the standard ? If you have other Smalltalk implementations at hand, what did they choose ? Code: (ReadStream on: 'test') skip: 6 -- Damien Cassou

2 2

Saving a headless image in linux problem
by Sebastian Sastre 21 Mar '07

21 Mar '07

Hi there, I'm using a headless squeak image for production and it happens to me that if I save it, is for sure that it will not be stable anymore. # squeak -version 3.9-8 #5 Tue Oct 10 11:56:09 PDT 2006 gcc 4.0.3 Squeak3.9alpha of 4 July 2005 [latest update: #7021] Linux ubuntu 2.6.15-27-386 #1 PREEMPT Sat Sep 16 01:51:59 UTC 2006 i686 GNU/Linux default plugin location: /usr/local/lib/squeak/3.9-8/*.so I've saved one image and then tried to telnet it (using REPLServer) and this is the result: # telnet localhost 9023 Trying 127.0.0.1... Recursive not understood error encountered -1229989588 SocketStream>flush -1229989788 ConnectingStream>flush -1229989880 ConnectingStream>flush -1230013288 REPLServer>getPasswordFrom: -1230013944 REPLServer>serve: -1230020920 [] in TcpService>value: -1230014152 BlockContext>on:do: -1230014264 BlockContext>valueWithBindingsContext: -1230014356 BlockContext>valueWithBindings: -1230015680 BindingsAccessor>clamp: -1230019688 [] in BlockContext>newProcessWithClampedBindings: -1230015864 BlockContext>on:do: -1230015988 BlockContext>valueWithBindingsContext: -1230016080 BlockContext>valueWithBindings: -1230019504 [] in BlockContext>newProcessWithBindings: -1230019320 [] in BlockContext>newProcess Connected to localhost. Escape character is '^]'. Connection closed by foreign host. By now I'll not count with squeak being trustable for headless images beign saved but I think this is a desirable feature. regards, Sebastian Sastre

5 7

Summer of Code: start submitting applications!
by Giovanni Corriga 21 Mar '07

21 Mar '07

Hi all, this is a friendly reminder to all the students interested in applying for the Google Summer of Code program. The deadline has been postponed to Monday 26th, 23:59 PST. This gives you a little less than 6 days for submitting your application. Nonetheless, you should start submitting your applications. You can always add more details and fix the errors using Google's web app. Giovanni

1 0

OSProcess: making sure that a child process always dies
by Joshua Gargus 21 Mar '07

21 Mar '07

Hi, Is there a way to use OSProcess on OS X (or unix in general) to cause forked child processes to die when the parent Squeak process quits? Thanks, Josh

4 15

beta testers needed for osx squeak browser plugin.
by John M McIntosh 20 Mar '07

20 Mar '07

Seems I need a few beta testers with dual core macintel based machines to do some testing. There are a few issues exposed by those machines, versus the solo cpu based macintel box I have. The number of folks who have doing testing in the past is rather small and they are currently constraint by their own monster work demands. Thus I need a few squeakers who would be able to do some testing in safari and firefox, report back their finding, examine console log, flip values in info.plist files etc for debugging. If you fit this picture, please email me and I'll send instructions how to install the new beta software and how to perform some testing. PS no dead products, that would be Microsoft Internet Exploder, although Opera if it works, well bonus. -- ======================================================================== === John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===

2 1

[FIX] FTPClient LoginFailedException
by Florian Minjat 20 Mar '07

20 Mar '07

Hi, Since some weeks I got a LoginFailedException very often when accessing an MCFtpRepository. My ftp server is ProFTPD 1.2.10 Server and didn't change for a long time. This bug was very annoying because I needed to save my changes many times before I could get no error. I traced the network exchange between the server and squeak and found that sometimes, the server just cut the connection after squeak sent the 'PASS' request. So I did a new method which checks when the connection is curtailed by a LoginFailedException, test if the socket was cut by the remote end and then do the login sequence again. It is pretty hackish but now it works fine. I didn't saw anything about this problem on the mailing list so I don't know if it is specific to my image, but here is the fix anyway. Florian A test to show my problem : ___________________________________________ TestFTPClient>>testFTPClientOpenOnHostNamed |client host user password nbOk nbBad| host := 'myHost'. user := 'myUser'. password := 'myPass'. nbOk := 0. nbBad := 0. 10 timesRepeat: [[[ self halt. client := FTPClient openOnHostNamed: host. client user: user. client password: password. client sendCommand: 'USER ', user. "331 Password required" client lookForCode: 331. "will ask user, if needed" client sendCommand: 'PASS ', password. "230 User logged in" client lookForCode: 230. nbOk := nbOk + 1. ] on: TelnetProtocolError do: [:ex | nbBad := nbBad +1] ] ensure: [client close] ]. self assert: (nbBad = 0) -------------------------------------------- A 'fix' in FTPClient : ____________________________________________ FTPClient class>>openOnHostNamed: host logUser: user password: password "this method comes from a recurent recent problem with ftp access with login/pass which are cut from the ftp server inexpectantly it just tests when the login sequence failed if the other end had cut the connection and retry it" |client| [ client := super openOnHostNamed: host. client loginUser: user password: password ] on: LoginFailedException do: [:ex | client stream socket isOtherEndClosed ifFalse: [ (LoginFailedException protocolInstance: client) signal: client lastResponse ] ifTrue: [ client close. client := FTPClient openOnHostNamed: host andLogUser: user password: password ] ]. ^ client ---------------------------------------------- A new test to ensure it's working well : ____________________________________________ TestFTPClient>>testFTPClientOpenOnHostNamed |client host user password nbOk nbBad| host := 'myHost'. user := 'myUser'. password := 'myPass'. nbOk := 0. nbBad := 0. 10 timesRepeat: [[[ client := FTPClient openOnHostNamed: host andLogUser: user password: password. (client getDirectory notNil) ifTrue: [ nbOk := nbOk + 1 ] ifFalse: [ nbBad := nbBad + 1 ] ] on: TelnetProtocolError do: [:ex | nbBad := nbBad +1] ] ensure: [client close] ]. self assert: (nbBad = 0) -------------------------------------------- And the fix in MCFtpRepository>>clientDo: to use my change ____________________________________________ MCFtpRepository>>clientDo: aBlock | client | client := FTPClient openOnHostNamed: host andLogUser: user password: password. directory isEmpty ifFalse: [client changeDirectoryTo: directory]. ^ [aBlock value: client] ensure: [client close] --------------------------------------------

1 0

← Newer
1
...
4
5
6
7
8
9
10
...
21
Older →

Jump to page: