Squeak-dev November 2005

squeak-dev@lists.squeakfoundation.org

163 participants
269 discussions

IMPORTANT: box1 will be no more at 27 November!!!
by Cees De Groot 07 Nov '05

07 Nov '05

This reminder will keep continuing in roughly a weekly fashion until then :) If you have services running on box1, or notice that something is pointing to box1, please take action. I will make a backup copy of box1 to box2 just a couple of days prior to shutdown, but it's everyone's own duty to check. Known services: squeakmap is still running there. That's why Goran is Cc'ed on this post ;) Regards, Cees

2 1

'self error: ..' considered harmful?
by Cees De Groot 07 Nov '05

07 Nov '05

I'm not sure whether I mentioned this - but I think one of the more useful areas that the Teams could be working on is to get rid of all the calls to #error:.... It could be that I'm the only one disturbed by this, but we have at least the possibility to work with exception hierarchies, but I find myself almost always having to catch Error in exception-handling code, because that's all the information that comes out of the basic class library...

4 4

Re: Design advice: lists and containment
by Chris Muller 06 Nov '05

06 Nov '05

> would also simplify my code to let each bookmark have a user instance > variable that "points" back to the user. > > 1) Is that kosher? Is it good practice to have a bookmark hold on to its > user? Well, a bookmark strikes me as very nearly a "value" object (with no extrinsic references) but you may have good reason for doing it. > 2) To add a bookmark to a user, it now takes two steps: > > user bookmarks add: bookmark. > bookmark user: user. I like to let my domain keep these relationships maintained, something like. User>>addBookmark: aBookmark "Clean up if moving aBookmark from a prior user.." aBookmark user ifNotNil: [ aBookmark user removeBookmark: aBookmark ]. bookmarks add: aBookmark. aBookmark user: self > 3) When I remove a bookmark from a user's collection, will the > bookmark be GC'd? I think so, but I want to be sure. Only if the Bookmark is not referenced. ;) > 4) The accessor method on user that returns the bookmarks means that > callers can manipulate the list. My first instinct is to say, "Don't > worry about it. I'm not developing a library, so I will know how to > use the collection properly." For me, this is the key to the answer to the question. I don't believe in writing code meant to guard against (potentially) incorrect code elsewhere; if its all your own code then you'll make it "compatible" with itself. You can't prevent misuse everywhere anyway, so there may be little point in copy of the collection that will help with misuse in merely one place. > Should I worry? Should the accessor > method return a copy of the bookmarks, so manipulation of the contents > won't affect the user's bookmark list? If so, then I have to provide > addBookmark: and removeBookmark: methods, right? If you returned a copy, right. - Chris

1 0

Re: MC passwords in images?
by Chris Muller 06 Nov '05

06 Nov '05

Sorry to say this again; images are (almost always) a personal thing, therefore sharing them is sharing something personal and should be done with great care or not at all; all this discussion trying to hash it out just proves it. So I still think the problem to solve is "ultimate image configuration." If someone gives me an image, its pretty much guaranteed I will want my own personalization (i.e., my preferred tools, etc.). So if I have to do image configuration anyway, why wouldn't I just configure it with whatever you're trying to deliver to me in the image you're sharing in the first place? The answer is, we don't have the tools to do it (or, we probably do but most of us aren't using them). We are using Monticello to share code but what to share other objects in the image? Now, I realize some lower-level work such as Traits may require image sharing, but the general-case, sharing objects such as Projects, should be solved generally. - Chris > On 11/5/05, Andreas Raab <andreas.raab(a)gmx.de> wrote: > > If there is no > > sensitive data in the image, then there is nothing to remember about > > sensitive data. > > And is there? Apart from the obvious bit? What about that > company-internal presentation in that project nested three levels > deep? It'd be nice if you could mark a project 'sensitive' and > scrub&save would notice it, no? Bummer if you hand out an image with > salary figures but the odd MC password safely in an external file ;) ...

2 1

RE: A wish for the File Team
by Peter Crowther 06 Nov '05

06 Nov '05

> From: [...] Craig Latta > Flow, being part of Spoon since Spoon 1a1 of 14 > February 2004, uses the > Spoon license, a license trivially derived from the 22 July > 1999 version of the BSD license [...] Craig, how have you managed to get rid of the terms of the Squeak license for the parts of Spoon that are derived from Squeak? - Peter

1 0

Re: My first morph
by Bob.Cowdery＠CGI-Europe.com 06 Nov '05

06 Nov '05

This was what I wanted to do: It's basically a frequency display consisting of 9 digits in 3 sets of 3 digits separated by '_' like 999_999_999 except the font size and colour will change for each group. The first basic requirement is that when I mouse over this I want each digit to highlight in some way, perhaps a box around it. When highlighted I want the mouse wheel to change that digit up or down by one for each click. It's very important to me that there is no flicker on this. I can't immediately see how to hook the mouse wheel events - is it possible to do so. Bert Freudenberg wrote: >> You will get these as keyboard events - Ctrl-Up/Down or Cmd-Up/Down, >> depending on the VM. See, e.g., ScrollPane>>scrollByKeyboard: Thanks Bert, got me past the first hurdle. This is basically working, very well actually. I don't know if this was the correct way to do things but I built a single digit morph and put them plus a few simple string morphs into an alignment morph. Now I need to know how to do user defined events such that when a digit overflows or underflows the next digit in line receives an event and can increment or decrement itself. The event can go direct or probably more sensible to send it to the parent which is my top level display class not the alignment morph. Any hints appreciated. Thanks Bob *** Confidentiality Notice *** Proprietary/Confidential Information belonging to CGI Group Inc. and its affiliates may be contained in this message. If you are not a recipient indicated or intended in this message (or responsible for delivery of this message to such person), or you think for any reason that this message may have been addressed to you in error, you may not use or copy or deliver this message to anyone else. In such case, you should destroy this message and are asked to notify the sender by reply email.

1 0

A wish for the File Team
by stéphane ducasse 06 Nov '05

06 Nov '05

hi guys This is years that craig is telling us that Flow is cool and better. Now I would like to know if we could build a group of person that really evaluate it: - is it good? - can we build a migration plan to use it? - does it make sense? - how much will it cost us (VM modification)? So is this target for the File team? Stef

6 16

pre-[ANN] KryptOn (was: MC passwords in images?)
by Chris Muller 06 Nov '05

06 Nov '05

> Oh, and is anyone out there interested in implementing a password > manager that stores passwords *outside* of an image in a > reasonably secure location? I have been working on this very problem for the last few months and have (what I think to be) a very nice solution which I call "KryptOn". For best results, "passwords" (more correctly, private-keys) are intended to be stored on one of those USB microdrive devices. My wife just bought me this one for anniversary gift: http://www.knifecenter.com/kc_new/store_detail.html?s=VN53975 KryptOn can be leveraged by any program to provide highly-transparent security based on the principles suggested for Croquet at http://minnow.cc.gatech.edu/squeak/3770 and it never leaves your passwords exposed in a saved image; it first overwrites the bytes and then dereferences them during #shutDown of the image. I have put some information about it on the Swiki: http://minnow.cc.gatech.edu/squeak/5785 Questions or comments are very welcome. - Chris

1 0

Failing test in 3.9a-latest: DosFileDirectoryTests>>testFileDirectoryNonExistence
by Cees De Groot 06 Nov '05

06 Nov '05

testFileDirectoryNonExistence "Hoping that you have 'C:' of course..." FileDirectory activeDirectoryClass == DosFileDirectory ifFalse:[^self]. self should: [(FileDirectory basicNew fileOrDirectoryExists: 'C:')] raise: InvalidDirectoryError. (initials indicate author is Andreas Raab) My question: why should this raise an error?

2 1

Re: Design advice: lists and containment
by Jim Menard 06 Nov '05

06 Nov '05

Everyone, Thank you for the thoughtful advice. In the past, I've usually come down on the side of "hand out the bookmarks list; trust the caller to be nice". The thing that tripped me up this time is that I found myself writing code that had a bookmark, but not the owner of the bookmark. How, in the bookmark, do I implement behavior like "remove me from the user's list" for example? I understand now that is the user's responsibility. I realize that I asked my question about back-linking before fully understanding who was responsible for what. I've applied the Law of Demeter in the past, and it led to hard-to-follow code: messages end up getting passed through three or four objects, and then back down to the original object. Jim -- Jim Menard, jim.menard(a)gmail.com, jimm(a)io.com http://www.io.com/~jimm

1 0

← Newer
1
...
18
19
20
21
22
23
24
...
27
Older →

Jump to page: