- Pharo-dev - lists.pharo.org

Re: [Pharo-project] Initial experience coming from Squeak

by Sean P. DeNigris

I filed bugs for these: Issue 2288: http://code.google.com/p/pharo/issues/detail?id=2288 : Sean P. DeNigris wrote: > > * what is with the workflow of the default browser? When you click on a > protocol, the method template appears in the bottom pane, all highlighted > and ready to be replaced. But because hovering over the pane doesn't give > it the keyboard focus (like Squeak), you have to click in it, which > unselects the template, which you then have to reselect manuallyâ¦ > Certainly not a big deal, but weird. I though of a few possible fixes: > - automatically give the bottom pane the keyboard focus when working in > the protocol column (which doesn't seem to need it (I like this one - > fewest steps for the user) > - don't have the template pre-selected, so that clicking in the pane > below the template selects it (easiest fix, requires one more user click) > Issue 2289: http://code.google.com/p/pharo/issues/detail?id=2289 : Sean P. DeNigris wrote: > > * bizarre window changing - the (on Mac) Cmd+(hold arrow) to rotate > through the open windows is pretty cool, but when you tap the arrow key > (instead of holding it), it just goes back and forth between two windows, > which I can't see a use for. It seems it would be more useful to have the > behavior to be: if you hold down the Cmd key and repeatedly tap the arrow > hit, the windows rotate one after the other until you get through them all > and recycle and v.v. for the opposite direction arrow. > Issue 2290 : http://code.google.com/p/pharo/issues/detail?id=2290 : Sean P. DeNigris wrote: > > * new windows e.g. workspaces appear wherever they feel like (is it a > standard location?). It seems to flow better if they appeared where the > user is already looking i.e. where the active hand is. > Sean P. DeNigris -- View this message in context: http://n4.nabble.com/Initial-experience-coming-from-Squeak-tp1799044p181979… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

April 9, 2010

[Pharo-project] Communication with VM guys (part three)

by StÃ©phane Ducasse

Begin forwarded message: > From: "David T. Lewis" <lewis(a)mail.msen.com> > Date: April 9, 2010 2:52:27 PM GMT+02:00 > To: Squeak Virtual Machine Development Discussion <vm-dev(a)lists.squeakfoundation.org> > Subject: Re: [Vm-dev] Questions, communication and process > Reply-To: Squeak Virtual Machine Development Discussion <vm-dev(a)lists.squeakfoundation.org> > > > On Fri, Apr 09, 2010 at 11:23:18AM +0200, stephane ducasse wrote: >> >> Hi vm guys, >> >> I would like to know the process to report problems, issues, enh to the vm because people asked me >> and I have no idea how this is done. I read the web site but I could not find a bugtracker. > > Hi Stef, > > In addition to Ian's explanations, I'll add couple of points. > >> >> Several questions: >> - How do people report problems? Just sending an email in the mailing-list is enough? > > The bugtracker is Mantis at bugs.squeak.org. Use category "VM" when > reporting or searching for problems. For any problems or enhancement > requests that may time time to resolve, it is very helpful to use > the bugtracker because it provides a process for ensuring that issues > are addressed or at least not forgotten, and also a reference for > people who run into an issue that someone else has addressed. > > I try to catch issues that I see on the Pharo and squeak-dev lists > and enter them on Mantis if they will need follow up, but I'm sure > that I miss things from time to time. > > Follow Ian's instructions for reporting problems to vm-dev and the > platform developer. That is the way that problems get noticed, > discussed, and addressed. But to ensure follow-up, please also > use the bugtracker. > >> - How a fix in one OS is propagated to the others? >> - How the user can know that? > > For issues that affect the VM overall, the change history of the > VMMaker package on squeaksource is helpful. You can use Monticello > of course, but I like to point a web browser at www.squeaksource.com, > project VMMaker, and browse through the entries under the "News" tab. > > A remaining problem is that that the VM user cannot always tell if > their VM actually contains a particular fix or feature. To some extent > you can figure it out from the name of the VM file, but we need a better > run time check. We (or I) will probably add a way to check this through > a primitive, but for now it remains confusing. > > Dave >

April 9, 2010

[Pharo-project] bizarre window cycling behavior

by Sean P. DeNigris

On the (on Mac) Cmd+(hold arrow) to rotate through the open windows is pretty cool, but when you tap the arrow key (instead of holding it), it just goes back and forth between two windows, which I can't see a use for. It seems it would be more useful to have the behavior to be: if you hold down the Cmd key and repeatedly tap the arrow hit, the windows rotate one after the other until you get through them all and recycle and v.v. for the opposite direction arrow. Issue 2289: http://code.google.com/p/pharo/issues/detail?id=2289 -- View this message in context: http://n4.nabble.com/bizarre-window-cycling-behavior-tp1819783p1819783.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

April 9, 2010

[Pharo-project] new windows (e.g. workspaces) appear wherever they feel like

by Sean P. DeNigris

It seems to flow better if they appeared where the user is already looking i.e. where the active hand is. Issue 2290 : http://code.google.com/p/pharo/issues/detail?id=2290. Sean -- View this message in context: http://n4.nabble.com/new-windows-e-g-workspaces-appear-wherever-they-feel-l… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

April 9, 2010

[Pharo-project] Communication with VM guys (part two)

by StÃ©phane Ducasse

Begin forwarded message: > From: Ian Piumarta <piumarta(a)speakeasy.net> > Date: April 9, 2010 12:35:16 PM GMT+02:00 > To: stephane ducasse <stephane.ducasse(a)gmail.com> > Cc: Squeak Virtual Machine Development Discussion <vm-dev(a)lists.squeakfoundation.org> > Subject: Re: [Vm-dev] Questions, communication and process > > Hi Stef, > > On Apr 9, 2010, at 2:23 AM, stephane ducasse wrote: > >> - How do people report problems? Just sending an email in the mailing-list is enough? > > Here are the options, necessarily biased towards my preference which is to try to avoid having to poll for outstanding issues. > > If it's a general problem of installation or bizarre behaviour that differs between platforms, ask for help on the Squeak VM mailing list. > > http://lists.squeakfoundation.org/mailman/listinfo/vm-dev > > Pharo and Squeak users are EQUALLY WELCOME to report VM issues on that list. Squeak has several people who regularly take the time to forward things from squeak-dev to vm-dev when appropriate. You might want to encourage the same thing on the Pharo list. I know for a fact that there are several vm-dev subscribers that monitor the Pharo list. (You might want to make it clear that Pharo uses the Squeak VM too. This does not necessarily appear to be universally understood.) > > If it's a Unix-specific problem that you think might be a bug, report it to the above mailing list AND send me email. Fixes are best submitted as complete modified files (not 'diffs'). (There might well be a way to tell ediff to use diff output directly, but that does not guarantee that your 'original' and my 'original' are in any way related. Sending whole files is MUCH more robust.) > > If it's Windows-specific, try the above list and Andreas Raab. For Mac-specific, the above list and John McIntosh. If it's common code shared by all platforms, especially code generated by VMMaker, send to Dave Lewis and the above list. > > VM bugs can fed to mantis here: bugs.squeak.org > > The people mentioned above probably look at mantis on a regular basis. Anything with more than four legs gives me the heebie-jeebies, which isn't why I don't regularly trawl mantis but it'll do. > >> - How a fix in one OS is propagated to the others? > > Fixes that affect more than one OS are very often in the common code that is generated automatically and shared between all the platforms. There is no explicit manual propagation required. (That being said, the people mentioned above are often in contact on a daily basis. When co-ordination is required, it will happen.) > > The exception is that some code is shared between Unix and Mac. John is good at spotting commits from me that affect him, and is good at bugging me when necessary into fixing anything Unixy that might affect him. > >> - How the user can know that? > > Watch vm-dev for discussions or commit reports that say the issue was fixed. If it is in mantis, it might even get closed too. > > The above reflects how I like to work, and I can only speak for myself. MPG from the others mentioned above might not be comparable. > > Cheers, > Ian > > (Feel free to forward, re-post, etc., the above as you see fit.) >

April 9, 2010

[Pharo-project] Communication with VM guys (part one)

by stephane ducasse

Hi guys I will send a following mail. We should probably synthetize and put them some summary on the web. Stef Begin forwarded message: > From: Ian Piumarta <ian(a)vpri.org> > Date: April 9, 2010 11:43:33 AM GMT+02:00 > To: laurent laffont <laurent.laffont(a)gmail.com> > Cc: stephane ducasse <stephane.ducasse(a)free.fr>, stephane ducasse <stephane.ducasse(a)gmail.com> > Subject: Re: [Vm-dev] Fwd: loading library with FFI fails (was [Pharo-project] Linux vm: assertions and strategy) > > Hi Laurent, > >> I've just seen you reply me in private. Do you mind if I forward your mail to Pharo & VM mailing lists, to see how we can improve communication and collaboration between projects ? I don't want to start a "counterproductive" flame war, I hate this, so maybe we can write another mail which doesn't offense people ? Maybe you have ideas on this ? > > How about we just say this... > >> It seems there is a complete lack of communication within the Pharo community regarding the virtual machine and its development/support. VM development and support (including bug reports and fixes) are discussed on the vm-dev mailing list. Both Pharo users and Squeak users are equally welcome there (and both communities should avoid discussing VM problems on their respective developer lists, since the people who should see those VM bug reports are more likely reading vm-dev than the other lists). Problems that appear to be platform-independent should always be reported to the vm-dev list. Problems specific to a particular platform should be sent to the maintainer of the platform whose e-mail address is included in the release notes or in the download area on squeakvm.org. > > Nobody could be offended by that? :) > > Cheers, > Ian >

April 9, 2010

Re: [Pharo-project] [Vm-dev] Fwd: loading library with FFI fails (was Linux vm: assertions and strategy)

by Schwab,Wilhelm K

Ian, I will try what you want. -----Original Message----- From: Ian Piumarta [mailto:ian@vpri.org] Sent: Friday, April 09, 2010 5:42 AM To: Schwab,Wilhelm K Cc: laurent laffont; Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Vm-dev] Fwd: loading library with FFI fails (was [Pharo-project] Linux vm: assertions and strategy) Hi Bill, On Apr 9, 2010, at 3:24 AM, Schwab,Wilhelm K wrote: > I cannot ignore what was never delivered to me - you might want to > reconsider your logic in this connection. My initial assumption appears to be incorrect and so the suggestion I sent not to be relevant to debugging your problem. But the original is attached in case you care. It seems something between you and me is unreliable in delivery or in spam detection. > You might also note that #moduleName can check the OS, so "it will > never work on Windows" is just plain wrong. My mistake. The Unix moduleName should be 'odbc' unless there's a good reason for it not to be. > Later in the day I will revert to your vm and try 'odb' Recompile from source. My best attempt at a fix for your issue is in the repo. > - I will admit to being less than optimistic, especially given what I > "ignored" - make that never received. Well now that your mailer likes me, why not try to be happy to work with me to resolve this rather than assuming my attempts are futile and in vain? Cheers, Ian From: Ian Piumarta Date: April 7, 2010 11:56:14 AM PDT To: "Schwab,Wilhelm K" <bschwab(a)anest.ufl.edu> Subject: Re: [squeak-dev] RE: Linux 3.11-3 vm and shared libraries Received: (qmail 20182 invoked from network); 7 Apr 2010 18:56:15 -0000 Received: from dsl092-038-093.lax1.dsl.speakeasy.net (HELO [10.0.1.2]) by mail7.sea5.speakeasy.net (qmail-ldap-1.03) with SMTP for <bschwab(a)anest.ufl.edu >; 7 Apr 2010 18:56:15 -0000 Message-Id: <70021DC8-9A49-48C1-B26F-6E75820EDC38(a)speakeasy.net> In-Reply-To: <A93938CC72687245835AC0BE10199F7C7B73C7B4(a)HSC-CMS01.ad.ufl.edu > Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v936) References: <A93938CC72687245835AC0BE10199F7C7B73C767(a)HSC-CMS01.ad.ufl.edu > <A93938CC72687245835AC0BE10199F7C7B73C7B4(a)HSC-CMS01.ad.ufl.edu> Bill, On Apr 7, 2010, at 10:34 AM, Schwab,Wilhelm K wrote: It is one thing for my hacked together .so to avoid detection due to some mistake on my part If your .so is in the library search path then it should be found. If your .so contains unresolved references then they should be reported on the console when it fails to load. has anyone tried ODBC on Linux with the new vm? It's sitting in /usr/ lib like it has always been; the vm should find it, right? Yes it should, but try setting your LD_LIBRARY_PATH=/lib:/usr/lib and running the VM. I think there is something wrong with the library searches in 3.11-3. Just different, and a lot more predictable: From: Ian Piumarta <piumarta(a)gmail.com> Date: September 16, 2009 7:36:01 PM PDT To: The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org > Subject: 3.11.3.2135 for Unix [...] NOTE: FFI libraries are no longer searched for in every single nook and cranny, which means you might have to set LD_LIBRARY_PATH (or your local equivalent) for FFI-based programs to work properly. Cheers, Ian

April 9, 2010

Re: [Pharo-project] Canvas transform bug when rendering text?

by Henrik Johansen

On Apr 9, 2010, at 11:40 38AM, StÃ©phane Ducasse wrote: > > On Apr 9, 2010, at 11:28 AM, Lukas Renggli wrote: > >> I don't write code like this. > > It was a joke :) > but I really like the visual effect > >> Rotating a window is just a good test to >> see if works, because it calls all kind of different code. >> >> If you rotate a more complex window than a workspace, my system hangs >> btw. So there must be other stuff broken there. > > oops > with the fix too? Yes. The fix was for which draw mode is used, and nothing else. The transformBy: method still creates a new Form of extent and depth same as the Tranformation each draw call, so it gives the GC a real workout. Sometimes this manifests as system hangs, others as a constantly higher CPU usage. (depending on how large your transformed morph is) Cheers, Henry

April 9, 2010

Re: [Pharo-project] Canvas transform bug when rendering text?

by Lukas Renggli

>> If you rotate a more complex window than a workspace, my system hangs >> btw. So there must be other stuff broken there. > > oops > with the fix too? Yes, with and without the change. Lukas -- Lukas Renggli www.lukas-renggli.ch

April 9, 2010

Re: [Pharo-project] [Vm-dev] Fwd: loading library with FFI fails (was Linux vm: assertions and strategy)

by Schwab,Wilhelm K

Ian, I cannot ignore what was never delivered to me - you might want to reconsider your logic in this connection. You might also note that #moduleName can check the OS, so "it will never work on Windows" is just plain wrong. Later in the day I will revert to your vm and try 'odb' - I will admit to being less than optimistic, especially given what I "ignored" - make that never received. Bill -----Original Message----- From: Ian Piumarta [mailto:ian@vpri.org] Sent: Friday, April 09, 2010 2:25 AM To: Schwab,Wilhelm K Cc: laurent laffont Subject: Re: [Vm-dev] Fwd: loading library with FFI fails (was [Pharo-project] Linux vm: assertions and strategy) On Apr 8, 2010, at 11:40 PM, Ian Piumarta wrote: > I have no idea. I sent email to Bill with a suggestion but he ignored > me. I therefore have no information about whether there is something > funny in his Linux installation or whether there is a real problem in > the VM (which appears to work fine for me). I could be completely wrong about this. If so, please disregard. But... From the patch I am inferring that the odbc library specifies "module: '/usr/lib/libodbc.so'". This would indeed be a problem for three reasons. 1. Mine might be in /usr/local/lib, or might be a personally-compiled version to fix some issue with the system version. In either case I'm screwed. 2. I might be using Darwin, in which case '/usr/lib/libodbc.dylib' might be the only thing that would work -- or even '/System/Library/ Frameworks/ODBCFramework/Libraries/libodbc.dylib'. Same comment as before. 3. It will never work on Windows, even if some day Windows supplies an 'odbc.dll' that is 100% compatible with the Linux one. The usual way for the plugin to specify this library would be "module: 'odbc'" which should work perfectly in all the above cases using the latest VM (and explains why I could not reproduce the problem using a simpler test case). I'm happy to modify the VM to cater for this specific situation of an over-specified library name, but the real solution would be fix the primitive declarations in the ODBC plugin. Cheers, Ian

April 9, 2010