- Pharo-dev - lists.pharo.org

Re: [Pharo-project] SIF working - needs a small fix
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

Bryce, It's every line of (eventually, if all goes well) several years worth of work. YIKES!! :) Speaking of the mac, does Wikipedia have it right? Did they change from CR to LF in OS-X? Do we plan a similar move? Regardless, I should be able to drop extra control characters either on the way out of Dolphin or into Pharo. Understood about inside strings; escaping them might work. RE reformatting via the RB, I realize you mean well, and therefore, please do not read even a hint of hostility into this. That said, the answer is NO WAY. Far too many times, I have been saved from myself thanks to comments dated 3-6 months apart, suggesting that something had simply not settled down and required attention. If the RB ever begins to respect pre-existing formatting, I will use it more aggressively than I do now. As it is, I use it freely for up/down moves and perhpas other miscellaneous features that do not format code. I sometimes use it on new code that has not had time to accumulate meaningful comments. Once I used Dolphin's scoped refactorings (not really refactorings at all, but useful just the same) to make sweeping changes through a couple of packages, and then manually repaired the damage; it was worth it in that case. For the above reasons, I have come to disagree with Damien about using the RB to rewrite "ClassName on:" constructs. IMHO, the resulting scattering of comments would be very poor engineering practice, especially since the class-side methods can be made to answer appropriate types w/o destroying or degrading any knowledge captured in comments. There are mentions of a magic fix for this, but every time I test the RB, it continues to mishandle comments. Thanks!! Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> bryce(a)kampjes.demon.co.uk 10/1/2008 3:50 PM >>> Bill Schwab writes: > Hello, > > I would like to tell you all the incredible things I did to get SIF > going - truth is, it loaded and just worked. The only gripe I have is > that the Pharo browsers are displaying a large number of LF or other > non-printable characters as boxes in the Pharo browsers. I have seen > Squeak do this many times, and either just ignored it or deleted them, > as my need to view the code and tolerance level allowed. It will get > old with the amount of code I plan to move. > > How would YOU fix this? Is it a SIF problem (dumping out things it > should not from Dolphin), a SIF problem (not supressing characters as > they are loaded in Pharo), or something missing in the browsers that is > expected to be fixed in Pharo? It's a line ending problem most likely. Squeak uses Mac style line endings. If there's only a few I'd delete the characters, if there's more I might try to automate it. In general automation would be a little tricky because you'd want to ignore any carriage returns in strings that you want to keep. If you know the source well enough to know that there aren't any then just write a script possibly in Squeak to remove them. It might be worthwhile trying to automatically reformat your code. That's likely to get rid of them as the RB formatter runs off the parse tree and ignores whitespace. Bryce _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

[Pharo-project] Image tidyup convention
by Igor Stasenko Oct. 1, 2008

Oct. 1, 2008

I would really like if Pharo will introduce a convention for tidying up the image, proposed by Keith. It could save a lot of time and nerves for developers/maintainers :) http://www.nabble.com/Squeak-low-on-space...-td13243777.html#a13244357 -- Best regards, Igor Stasenko AKA sig.

3 4

[Pharo-project] SIF working - needs a small fix
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

Hello, I would like to tell you all the incredible things I did to get SIF going - truth is, it loaded and just worked. The only gripe I have is that the Pharo browsers are displaying a large number of LF or other non-printable characters as boxes in the Pharo browsers. I have seen Squeak do this many times, and either just ignored it or deleted them, as my need to view the code and tolerance level allowed. It will get old with the amount of code I plan to move. How would YOU fix this? Is it a SIF problem (dumping out things it should not from Dolphin), a SIF problem (not supressing characters as they are loaded in Pharo), or something missing in the browsers that is expected to be fixed in Pharo? Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029

2 1

[Pharo-project] [build broken] 10083/10084
by Marcus Denker Oct. 1, 2008

Oct. 1, 2008

Please dont' update, it's broken right now. (And I don't have time to fix it, I have to admit) Marcus -- Marcus Denker -- denker(a)iam.unibe.ch http://www.iam.unibe.ch/~denker

2 2

[Pharo-project] [updates] 10083
by Marcus Denker Oct. 1, 2008

Oct. 1, 2008

Issue 191: freeze in World/Configuration menu Issue 134: ReadOnlyVariableBinding should be removed (unused experiment) Issue 194: some random cleanups from 23.09 Issue 193: Squeakmap fails to open if it is purged and no sm/ dir present Issue 205: [ENH] >10% faster senders of -> removed "CurrentProjectRefectoring". --> all references replaced by "Project current". -> ReadOnlyVariableBinding removed -> cleanGenie: removes genie related methods from HandMorph -> DNDAnimationKill: there is an option nobody ever uses to draw a line on dragndrop. -> cleanSmallLandTheme: removes methods from Preferences. More needed later. -> md-random-cleans.1: remove KidNavigationMorph and some never called methods. -> Fullscreen of/on isrendered like soundof/on, that is just one entry that changes according to state of display. -- Marcus Denker -- denker(a)iam.unibe.ch http://www.iam.unibe.ch/~denker

6 14

Re: [Pharo-project] OSProcess under Pharo
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

Bryce, Thanks for your reply. Actually, I mentioned FFI etc. because that appeared to be a way to get what OSProcess wanted in terms of dependencies; using Universes should handle that part for me. FFI would be of interest on its own, but appears to make people nervous. Is that for good reason? Put another way, if I want to make some calls into a dll/so, do I really need to mess around with creating/compiling/distributing a plugin? It seems like a lot of overhead, especially since Dolphin does such a nice job based on a mechanism very similar (in apperance at least) to FFI. That said, I am interested in candid opinions; maybe things really are different in Squeak/Pharo. Thanks! Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> bryce(a)kampjes.demon.co.uk 9/30/2008 2:12 PM >>> Bill Schwab writes: > Before doing something I might regret, I wanted to check with you about > installing OSProcess, VMMaker(?), FFI, etc. Do you have anything to say > "try it and report any problems?" I load OSProcess and FFI from universes, it works loading into a recent pharo-dev image. There's an issue with CommandShell as it subclasses off some MVC classes, I think it's safe to ignore. I'd load VMMaker from SqueakSource directly, but it depends on what you want to do. I've merged the FreeType changes into the Exupery VM's and will add closure support in the next few days. Loading Exupery-dev from the universes will load all the packages you listed. (Though, there's an incompatibility between Exupery and the stricter instance variable checks in Pharo, will be fixed in the next release.) Bryce _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

2 1

Re: [Pharo-project] strange problem with 10083
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

Stef, To reproduce, one would start with 10083 and invoke Installer loadOB. Is that correct? It will be a couple of hours before I would have time to try it, and my hunch is probably wrong anyway, but here goes: could it be a nil stream (file or URL that is incorrect or otherwise non-openable) or a read off the end of a stream that defaults to nil? Among other things, I have been working on File and FileStream classes (mostly facades around Pharo's streams, with lots of error checking and signalling) to avoid silent failures that merely postpone the inevitable until down stream with a nil receiver and NO clue as to what really happened. My primary goal with File and FileStream is to get a large amount of my Dolphin code loaded into Pharo and working without a huge fight. A close second is to provide the alternate selectors (#nextOne, etc.) I described to avoid silent failures in stream I/O. Whether or not it is at the root of your current problem (my guess is that something else changed, and a failure to raise an error is masking it), I strongly recommend deprecating (but keeping) the offending selectors. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> stephane.ducasse(a)inria.fr 10/01/08 1:34 PM >>> Hi I tried to load the following code into 10082 and 10083 in 10082 it works like a charm, in 10083 loadOB "self loadOB" | instClass | instClass := self environment at: #Installer ifAbsent: [self new installingInstaller]. instClass ss project: 'AST'; install: 'AST-lr.158'. instClass ss project: 'RefactoringEngine'; install: 'RefactoringEngine-ab.51'. instClass wiresong project: 'ob'; install: 'OmniBrowser-dr.418'; install: 'OB-Standard-dr.353'; install: 'OB-Morphic-dr.56'; install: 'OB-Enhancements-dr.216'; install: 'OB-Refactory-lr.100'; install: 'OB-SUnitIntegration-dc.9'. I get a extremely strange bug UndefinedObject DNU value: !!!!! when Instance := nil is selected and RefactoryChangeManager class>>nuke Instance notNil ifTrue: [Instance release]. Instance := nil RefactoryChangeManager class >>initialize self nuke. UndoSize := 5 _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

3 3

Oct. 1, 2008

Hi I tried to load the following code into 10082 and 10083 in 10082 it works like a charm, in 10083 loadOB "self loadOB" | instClass | instClass := self environment at: #Installer ifAbsent: [self new installingInstaller]. instClass ss project: 'AST'; install: 'AST-lr.158'. instClass ss project: 'RefactoringEngine'; install: 'RefactoringEngine-ab.51'. instClass wiresong project: 'ob'; install: 'OmniBrowser-dr.418'; install: 'OB-Standard-dr.353'; install: 'OB-Morphic-dr.56'; install: 'OB-Enhancements-dr.216'; install: 'OB-Refactory-lr.100'; install: 'OB-SUnitIntegration-dc.9'. I get a extremely strange bug UndefinedObject DNU value: !!!!! when Instance := nil is selected and RefactoryChangeManager class>>nuke Instance notNil ifTrue: [Instance release]. Instance := nil RefactoryChangeManager class >>initialize self nuke. UndoSize := 5

2 3

Re: [Pharo-project] [updates] 10083
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

I will confess to being in a similar situation. I have been grabbing new images and looking at them, but am hoping to wait for the integration before moving my code from my "very old" (in Pharo time) image to a more recent one. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> simon(a)pinesoft.co.uk 10/01/08 5:27 AM >>> Not entirely fair I know, but I'm using the demo TT + PolyMorph image which was based on 10074. If I try to update from the server on this image it throws up an emergency debugger, which is unsavable. This may well be expected, given that the TT and PolyMorph integration isn't a standard Pharo thing yet. Are you planning to release another TT and PolyMorph image from 10083? Cheers, Simon On 1 Oct 2008, at 10:15, Adrian Lienhard wrote: > > On Oct 1, 2008, at 11:03 , Damien Cassou wrote: > >> On Wed, Oct 1, 2008 at 11:01 AM, Marcus Denker >> <denker(a)iam.unibe.ch> wrote: >>>> May be time to release? How do we proceed for the dev packages? >>>> >>> >>> not yet. in the changesets I added there of course was a >>> conflict.. and some >>> more cleanup is needed (remove preferences not needed, and so on) >> >> I may not have been clear enough. I meant a downloadable 10083 core >> image. > > Actually, this would be a good task to automate (i.e., a script that > checks once a day for new updates, loads them and then saves the > image). Shouldn't be hard... Concerning terminology, I'd rather talk > about "nightly builds" than releases (I think the latter is a bit > misleading). > > Adrian > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project ****************************************************************************************************************************************** This email is from Pinesoft Limited. Its contents are confidential to the intended recipient(s) at the email address(es) to which it has been addressed. It may not be disclosed to or used by anyone other than the addressee(s), nor may it be copied in anyway. If received in error, please contact the sender, then delete it from your system. Although this email and attachments are believed to be free of virus, or any other defect which might affect any computer or IT system into which they are received and opened, it is the responsibility of the recipient to ensure that they are virus free and no responsibility is accepted by Pinesoft for any loss or damage arising in any way from receipt or use thereof. ******************************************************************************************************************************************* Pinesoft Limited are registered in England, Registered number: 2914825. Registered office: 266-268 High Street, Waltham Cross, Herts, EN8 7EA _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0

Re: [Pharo-project] On Win32 "metrics"
by Bill Schwab Oct. 1, 2008

Oct. 1, 2008

Michael, I grabbed Sophie 1..0.3 (linux) and do not see FileLocations. Is there another version that I need to get? There is a Sophie-FFI-Win32 category, and probably other Win32 specifics. I need to see what you have in mind, but if it is in the spirit of other "look in the usual places" methods I have seen, it would not be enough IMHO, at least not to get it right. Windows can be installed in non-default locations, and I (and perhaps others) confound things by creating "missing" directories to make things work. As an example, long ago, I had code that expected c:\windows\desktop to exist - it was as good a place as any to dump the result; when win2k came along, I simply created the directory and got back to work. I have never "fixed" that because it is not really a problem. Many changes in Windows require more attention. If we are to claim to support Windows, the system directory is a minimal thing to get right. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> m.rueger(a)acm.org 10/01/08 12:39 AM >>> Bill Schwab wrote: > Does Pharo already expose the needed information? All I found was 'NT' The cleanest way do to it is to use the FileLocations package that we use in Sophie. It makes system calls to find out all the interesting locations. Tweak uses a class called CPlatform and subclasses to host calls that return platform specific information. In one image I worked on I generalized it, called it simply Platform (I think). Comes in very handy :-) Michael _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

1 0