- Pharo-dev - lists.pharo.org

[Pharo-project] SqueakSource is unusable
by jaayer April 10, 2010

April 10, 2010

or if not that, then nearly so. The biggest problem is the overall smallness of each page. The site is far too narrow and apparently was designed for monitors with 800x600 screen resolutions. This was unacceptable five years ago and is even more unacceptable now. Further, the fonts are tiny, and worse, they and almost every other sized property are specified using hard-coded pixel widths. For scalable, cross-platform design, percentages and em-widths should be preferred. Links should display some sort of text-decoration when moused-over so it is always clear which link will be selected when the user clicks, even when two or more links are especially close together. SqueakSource is, after the primary home pages, the public face of the Squeak and Pharo projects. It is crucial then that it look--if not great--at least decent. Enclosed is prettified CSS that fixes some of these defects. (Note that Image url() references need to be filled in, and it still probably needs some tweaking.) body { border:0; background:gray; text-align:center; margin:0; padding:10px; font-size:0.9em; } body,div,table,tr,td,th,form input,form submit { font-family:Geneva, Arial, Helvetica, sans-serif; } table,tr,td,th,form { border:0; margin:0; padding:0; } th { text-align:left; } a,a:visited { color:#360484; text-decoration:none; } a:hover,a:active { text-decoration:underline; } #container { width:90%; background:white url(/inserturl) repeat-y 4%; border:1px solid black; border-top:5px solid black; border-bottom:5px solid black; text-align:left; margin:auto; } #top { width:100%; height:50px; background:url(/inserturl) no-repeat; border-bottom:1px solid black; font-size:0.7em; text-align:right; } #sstoolbar { height:2em; border-bottom:1px solid black; background:white; } #sstoolbar a { display:block; float:left; margin-left:-1px; border-left:1px solid black; border-right:1px solid black; line-height:2em; height:2em; letter-spacing:1px; text-decoration:none; padding:0px 10px; } #sstoolbar a:hover,#sstoolbar a:active { background-color:#360484; color:white; text-decoration: none; } #page { padding:0; } #page td,#page th { padding-right:5px; vertical-align:top; } #navigation { float:left; width:12%; padding:0 10px 10px 10px; } #navigation .title { margin-top:11px; font-weight:bold; } #contents { width:70%; height:auto; text-align:justify; margin:0 0 0 16%; padding:5px 10px 10px; } #contents h1 { margin-top:0; font-size:20px; } #contents h2 { margin-bottom:5px; font-size:15px; } #contents h3 { margin-bottom:5px; font-size:13px; } #contents hr { border:none; border-bottom:1px dotted black; } #contents p { width:90%; } #bottom { padding-left:5px; padding-right:5px; clear:both; border-top:1px solid black; background:white; margin:0; } .TabSelected { background-color:#360484; color:white; width:95px; text-align:center; } .TabUnselected { background-color:#D0C4E2; width:95px; text-align:center; } .TabSpacer { width:5px; } .TabLine { background-color:#360484; } #batch { text-align:center; margin-top:8px; } .evenRow { background-color:#E7E1F0; } .largeLabel { width:120px; } .mediumLabel { width:100px; } .smallLabel { width:80px; } .activityTable { margin-left:auto; margin-right:auto; } #window-titlebar { background-color:#D0C4E2; margin-bottom:10px; width:100%; } #window-title { text-align:right; width:66%; } #window-close { text-align:right; } .column-choice select { min-width:15em; }

5 5

[Pharo-project] Initial experience coming from Squeak
by Sean P. DeNigris April 10, 2010

April 10, 2010

I love the clean look with all the great developer features pre-installed (shout, OCompletion, etc.). It's been a pleasure to work with. And, a few things stood out that didn't seem to work as well. I'm sharing them here because I wonder what your opinions are. Also, what is the policy about contributing "good ideas" i.e. new/altered features that are not bug fixes? Is it best to discuss them on the list like this, or submit a changeset right off the bat per http://code.google.com/p/pharo/wiki/HowToContribute? Suggestions: * Why shift for meta? On a Macbook (with one button), it's awkward to press shift+alt+click, and also disorienting to go back and forth from Squeak which is just alt+click. Is there a reason to not keep the convention? * 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) * 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. * 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. * Single/multi-line popups - when you're extracting a method via "refactor source," you get a huge multi-line popup window, which eats crs, so after you enter the name and hit return, you are still in the popup and must take your hands off the keyboard to click "ok." Now, fair enough, this might actually be a good thing in other circumstances, but - when you reference an unknown class, you get an "unknown variable" popup, choose "define new class," click through the category selection, you get a one-line popup to edit a paragraph-long class definition, lol! So: - "refactor source->extract method" should be a single-line popup - and, "define new class->edit class definition" should be a multi-liner. Thanks for an awesome environment, and I will continue to contribute any way I can :) Sean P. DeNigris -- View this message in context: http://n4.nabble.com/Initial-experience-coming-from-Squeak-tp1799044p179904… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

6 12

[Pharo-project] potential bugfix for orphan OCompletion menus
by Romain Robbes April 10, 2010

April 10, 2010

Dear list, I just published a version of OCompletion with a possible fix for the bug in which OCompletion menus would stay on the screen after a browser is closed. The issue here is inherited from eCompletion: the menu morph is not a child of the window, but a top-level morph. To adress this, I implemented a timeout mechanism: if their was no keystroke in the editor which spawned the menu for the last 5 seconds, the menu will close automatically. You can hence try Ocompletion-RomainRobbes.54 on squeaksource, and tell me what you think of the solution, or of the length of the timeout (I tried to stay conservative, maybe something like 3 seconds would work better). Cheers, Romain -- Romain Robbes http://romain.robb.es rrobbes(a)dcc.uchile.cl

5 12

[Pharo-project] [update 1.1] #11310
by StÃ©phane Ducasse April 9, 2010

April 9, 2010

11310 ----- - Issue 2286: Class definition while compiling method - Issue 2131: Handling Window events. Thanks Henrik - I looked at the code and I would not be able to fix that. again some stuff to learn. Stef

2 1

[Pharo-project] Communication with VM guys (part three)
by StÃ©phane Ducasse April 9, 2010

April 9, 2010

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 >

1 0

[Pharo-project] bizarre window cycling behavior
by Sean P. DeNigris April 9, 2010

April 9, 2010

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.

1 0

[Pharo-project] Communication with VM guys (part two)
by StÃ©phane Ducasse April 9, 2010

April 9, 2010

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.) >

1 0

[Pharo-project] Communication with VM guys (part one)
by stephane ducasse April 9, 2010

April 9, 2010

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 >

1 0

Re: [Pharo-project] [Vm-dev] Fwd: loading library with FFI fails (was Linux vm: assertions and strategy)
by Schwab,Wilhelm K April 9, 2010

April 9, 2010

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

1 0

Re: [Pharo-project] [Vm-dev] Fwd: loading library with FFI fails (was Linux vm: assertions and strategy)
by Schwab,Wilhelm K April 9, 2010

April 9, 2010

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

1 0