September 2010 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] WebClient License Update

by Andreas Raab

On 9/9/2010 1:15 AM, Henrik Johansen wrote: > > On Sep 9, 2010, at 4:27 50AM, Andreas Raab wrote: > >> Phillipe wrote: >>> Did you change your position towards #squeakToUtf8 and string >>> concatenation (I didn't follow the entire previous thread)? >> >> Implicitly yes, but let's discuss this. You're saying you want WebClient "without overrides" but is this really what you mean? The only reason these methods are marked overrides is so that they don't kill your system if the methods themselves get added to the system at some point. For example, Grease currently adds extension methods for Collection>>sorted but these methods are already in Squeak and while nothing harmful happens if you load Grease, unloading it will destroy several methods. As a consequence it is vastly advantageous to mark methods as overrides if they have even the slightest possibility to conflict (but of course if you'd rather have them straightforward extensions, I have no problems with that). >> >> As for integrating the changes itself, I think that we're talking two very different issues here. I suspect that the only objection to #squeakToUtf8 and #utf8ToSqueak is that they have "Squeak" in their name. I don't think you'd have any objection if they would've been called #utf8Encoded and #utf8Decoded. Which, honestly, is a childish attitude from my point of view. > > Yeesh. > I can only speak for myself, but my objection to the methods is more that I don't think knowing how to change its contents to a specific encoding should be the responsibility of the String class. > Yes, there are plenty of other comparable delegations of responsibility, but in this case it is easily avoided. > Sure though, it's a bit faster in some cases than the alternatives using convertToWithConverter:, or convertToEncoding: (at least with a symbol -> converter cache added), > but with the general improvements to UTF8TextConverter the speedup is no longer measured in the orders of magnitudes, where it might feel worth it (to me). Which isn't really the argument. The utf8 encoding and decoding methods are by far the most commonly used conversion methods. They are similar to methods such as #zipped/#unzipped, or #base64Encoded/#base64Decoded. I actually have some code that literally uses: data := aString squeakToUtf8 zipped base64Encoded. Compare that with the (exactly) equivalent version that doesn't have the utility methods: (Base64MimeConverter mimeEncode: (((GZipWriteStream on: (RWBinaryOrTextStream on: String new)) nextPutAll: (aString convertToWithConverter: UTF8TextConverter new); close; encodedStream) contents) readStream) contents Thus, the basic claim here is that it's worthwhile to have shorthand methods for conversion to/from utf8 because it's so prevalent. As for why it's called squeakToUtf8 and utf8ToSqueak, that's simply because there's a family of conversion methods in String which use the squeakToXXX/xxxToSqueak pattern (remnants of which you can still find in Pharo). These methods got first introduced around Squeak 3.8 to deal with various issues during m17n integration and have been in Squeak ever since. >> Put differently, the change adds nothing but robustness to the system. There is really no data to back up the FUD about all the changes that "may" go wrong, but from my experience there is ample evidence to show that logging and error handling involving string concatenation is highly susceptible to this kind of problem and generally not very well tested and difficult to QA. And the major use of string concatenation is right in these areas. > > While I see some value in your arguments, in addition to Pharo, at least the VW, GST and ST-X dialects, plus the ANSI standard, do not allow it. > IE, if you want to write your code the least bit portable, you include the ugly as/printString sends. Unless you convince people to change things. Would make a fine addition to Grease, too :-) Cheers, - Andreas

Sept. 11, 2010

Re: [Pharo-project] Mark and space parity

by Schwab,Wilhelm K

Tony, I *almost* understand that, and greatly appreciate your reply. Compiling a new vm is getting to be routine, mostly because I find its diagnostic output to be far less than I typically want when sorting out a problem. Certainly the Windows vm goes out of its way to fail on valid settings and should be changed to allow parity values 3 and 4; the Linux vm might also benefit from relaxing its checks. The interesting part is that I don't need to send to the device, only read from it. If my notes are correct, it is 7S2. Gary, re programming by barcode, are you thinking that I might be able to get the scanner to use different settings? I wondered about that last night. There is a fairly extensive book full of barcodes that I believe are for changing the scanner's behavior. In your experience, are such changes stored in persistent store (flash, eprom, etc.), or are they transient such that they typically have to be restored each time the scanner loses power? If I can get it to use something friendly to the vm and stay that way, that might be a good solution. Bill ________________________________________ From: Tony Garnock-Jones [tonyg(a)lshift.net] Sent: Saturday, September 11, 2010 1:44 PM To: The general-purpose Squeak developers list Cc: Schwab,Wilhelm K; vm-dev(a)lists.squeakfoundation.org; Pharo-project(a)lists.gforge.inria.fr Subject: Re: [squeak-dev] Mark and space parity Hi Bill, Based on http://www.lothosoft.ch/thomas/libmip/markspaceparity.php, it looks like you're set if you need 7S1. If you need 8S1, however, the options for linux are - flip between 8N1 and 8E1 depending on the byte to send, or - use the undocumented CMSPAR flag. Given that the SerialPort class in Squeak uses 0, 1, 2 for no, even and odd parity respectively, and the sqUnixSerial.c driver doesn't support any other kinds of parity specification (see the table parityTypeDecode), ruling out the use of CMSPAR, your only option (aside from modifying the SerialPlugin or writing your own) is to flip between 8N1 and 8E1 for each byte. Unfortunately it doesn't look like there's a way to get at tcsetattr without also doing an open, so there goes that idea as well. In summary: 7S1 ought to be OK (use 8N1 with a cleared 8th bit), but I suspect that 8S1 is out of reach without changes to the SerialPlugin. Regards, Tony On 11/09/2010 09:42, Schwab,Wilhelm K wrote: > I ran across a device (barcode scanner) that uses space parity and dug into vm code to see if it was indeed going to be a problem. This is not a debate on whether or not the designers of the device could have chosen something else: clearly they should have. The question is whether our vms should be able to communicate with it, and they should where possible. > > It is not clear whether Linux even supports space parity, though there are hints that recent kernels do so. Windows certainly does support it. Have any of you made the required changes and had success with it? On which operating system(s)? > > Bill > > >

Sept. 11, 2010

Re: [Pharo-project] WebClient License Update

by Andreas Raab

On 9/9/2010 3:34 AM, StÃ©phane Ducasse wrote: >> For string concatenation on the other hand, we're basically talking about dispersing a whole load of FUD about all the things that "may" go wrong. Fact is, nothing actually *does* go wrong with the change, but the change does fix situations that are *very* difficult to handle otherwise. One of the unfortunate realities of string concatenation is that it's very often used around error messages and logging, often in corner cases that aren't well tested, like: >> >> <some impossible condition> ifTrue:[ >> self log: 'Found impossible condition: ', foo. >> ]. > > I do not buy this argument. We should fix the code and not break the libraries. Yes, that is the usual FUD. Can you clarify how that change breaks any libraries, where and when exactly? Because it doesn't. > Your argument is bogus. Which one? I made about five different arguments from that this change doesn't break anything, over it preserving the type semantics, that string concatenation often occurs in less tested code, that it is error-prone for no good reason, up until the efficiency argument. Or is it simply "I don't like this guy so matter what he says I will not listen"? > We do not want such kind of extensions that opens the door to really sloppy code. And that would be sloppy code because ... ? It's more efficient to write it that way? Is that your measure of sloppyness? If not how do you distinguish "sloppy" code from other perfectly well accepted patterns like the type coercions happening in arithmetic? You *could* write "4 asFloat + 3.0" instead of "4 + 3.0" you know. > Then we should avoid overrides as much as possible. I wholeheartedly agree but did you actually follow the Grease discussion? The problem of not being able to remove methods after they have been added to the base is quite real. The only way to address this problem currently is by using overrides. Cheers, - Andreas

Sept. 11, 2010

Re: [Pharo-project] Compiling the VM for MacOSX

by John M McIntosh

On 2010-09-11, at 7:34 AM, Alexandre Bergel wrote: > Thanks John. It works well! > I can update the pharo book about this, but I am not sure how. The instruction given by the book to compile the OSX VM is fairly different from what you said (which works!) The instructions seems to target building a VM from April of 2009. Lots has changed since then to make the process of building a Cocoa OS-X VM to match a distributed pre-built VM version more of a one-step process. Otherwise it's just too hard to mesh the pieces together between an arbitrary squeak image, vmmaker, and the source from squeakvm.org -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

Sept. 11, 2010

Re: [Pharo-project] Compiling the VM for MacOSX

by John M McIntosh

On 2010-09-11, at 7:42 AM, Alexandre Bergel wrote: > >> If you want a fast vm i recommend to compile in deployment, but that take more time. > > How do I compile in deployment? By inspecting the project and selecting the "Build" tab? Ok, there is a bit of learning that has to take place since I'm going to assume if you want to build your own VM, then you should be familiar with XCode, otherwise you should just use the pre-built binaries. The Build tab you refer to lets you change the compile, link, deployment settings. In the Menu Bar the "Project" menu has "Set Active..." to let you choose which project to build and the type, etc... > > Cheers, > Alexandre > -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

Sept. 11, 2010

Re: [Pharo-project] Anything similiar to DabbleDB still available?

by Yanni Chiu

Friedrich Dominicus wrote: > I wonder if anyone knows of anything like DabbleDB now that that has > closed their doors? I just checked, and there's NO new announcement that they've closed. It's still showing the June 10 announcement that no new accounts are being accepted, and that current customers will continue to run, until further notice. On their discussion board, is a thread about their future. It also has discussion about various alternatives. See: https://www.dabbledb.com/help/forum/viewtopic.php?id=4253&p=7

Sept. 11, 2010

[Pharo-project] [update 1.2] #12137

by Marcus Denker

12137 ----- Issue 2862: Cleanup related to retractable changesets (part2) -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Sept. 11, 2010

[Pharo-project] [update 1.2] #12136

by Marcus Denker

12136 ----- Issue 2862: Cleanup related to retractable changesets (part1) -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Sept. 11, 2010

Re: [Pharo-project] Pharo-project Digest, Vol 29, Issue 84

by Eliot Miranda

Hi Ralph, On Sat, Sep 11, 2010 at 1:05 AM, Ralph Boland <rpboland(a)gmail.com> wrote: > > Hi, i don't know why things are done the way they are... > > Is your question how the exception mechanism works, or why it is > > needed to be done with an exception? as opposed to say doing it > > some other way? > > > The direct answer to your question is that given it is implemented in > > terms of exceptions, the signal is needed to trigger the > > defaultAction, or to allow someone higher up in the stack to trap it > > and redirect the notification. > > > cheers, > > Mike > > > I am no expert on Exceptions but I do understand this. > What I don't understand is WHY ProgressInitiationException > is used and why a signal is sent to its instance to get the ball rolling. > > Could not SystemProgressMorph be modified to handle this? > > Alternatively, could not some other new (non-exception) class handle > the logic where an instance of that class is created and gets the ball > rolling without sending a signal to itself? For me, currently, this > is the more intuitive way to do things. Why use signals when they > are not needed? (Or are they needed for some reason unknown to me?) > > The modifications I am planning to make to the ProgressDisplay > package look easier if I get rid of the signal handling. > But I don't know what I will be breaking or what functionality will > be lost if I get rid of the signal handling. > So I need to know WHY? > I tried to answer your question the first time round but as I looked at the code I found that I didn't understand it, and so gave up on my explanation :). It seemed to me that the reason for using a signal was to handle nesting. One raises the progress initiation signal in the wrapper to the action block to get progress to display in the right progress bar. If the signal is uncaught then its defaultAction would open up a new progress bar with one progress strip and initiate processing of the progress block, providing a handler for the progress initiation signal. If the signal is caught (by the handler provided at the first level) then the existing progress bar is extended with a new progress strip for the nested progress. At least that's how I thought it worked but I couldn't work out how the handler worked. If I were writing it I would still use the signal to find if the progress was nested but I /would not/ put the code in the defaultAction. Instead the signal would answer if it was nested and all the code would be at the point where the signal is raised. For this one can use a simple Notification, using its tag to disambiguate it e.g. [Notification new tag: #progressBar; signal] on: Notification do: [:ex| ex tag == #progressBar ifTrue: [ex resume: true]. ex pass] true So e.g. outerBarOrNil := Notification new tag: #progressBar; signal. ... if nested add to the existing bar otherwise create a new bar ... [actionBlock value: bar] on: Notification do: [:ex| ex tag == #progressBar ifTrue: [ex resume: bar]. ex pass] my 2Â¢ If you can make the tests pass with simpler code then go for it! best Eliot > Regards, > > Ralph Boland > > > On Fri, Sep 10, 2010 at 4:37 AM, Ralph Boland <rpboland(a)gmail.com> > wrote: > > > I posted the following earlier to the Squeak list but didn't get a > responce. > > > Hopefully someone on the Pharo list can help me. ?I need this > information > > > so I can make some planned changes to the ProgressDisplay package I > released > > > a while back. > > > > > > > ----------------------------------------------------------------------------------------------------------------- > > > > > > A SystemProgressMorph is used to display progress while some > computation > > > is being done. ?Furthermore the progress bars can be stacked to show > progress > > > of component computations within a computation. > > > This I understand and understand most of the code. > > > But not All. > > > What I have trouble with is understanding why a > ProgressInitiationException > > > is needed. ?In particular, String>>displayProgressAt:from:to:during: > > > does: > > > > > > ProgressInitiationException > > > > display: self > > > at: aPoint > > > from: minVal > > > to: maxVal > > > during: workBlock. > > > > > > > > > This method sets a few variables and then does: > > > > > > ^self signal > > > > > > This signal is caught where upon a SystemProgressMorph is started or > > > extended with > > > an additional progress bar. > > > > > > But I don't understand why sending a signal to a > ProgressInitiationException > > > is needed to get the ball rolling, > > > > > > Can someone please explain why things are done the way they are? > > > > > > > > > Regards, > > > > > > Ralph Boland > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Sept. 11, 2010

[Pharo-project] Pharo sprint now?

by Alexandre Bergel

Who wants to code? Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Sept. 11, 2010