Matthew,

This has nothing to do with the bug you posted. Here is the original method:

= aVersion

^ numbers = aVersion numbers

It tests equality and not identity. If aVersion is not a VersionNumber, which is the situation happening when streaming in data (version = String), then aVersion will not understand #numbers.

Here is my fix:

= aVersion

^ (aVersion isKindOf: VersionNumber) and: [numbers = aVersion numbers]

Which protects against that kind of business.

I now look to see how fast and efficient the bug processing loop is and see this make it into the image. I have a feeling that it is not so efficient.

Rob

----- Original Message ----- From: "Matthew Fulmer" tapplek@gmail.com To: squeak-dev@lists.squeakfoundation.org Sent: Monday, November 05, 2007 8:45 AM Subject: Re: [BUGFIX] SqueakMap is broken in 3.10

On Mon, Nov 05, 2007 at 08:34:41AM -0800, Rob Withers wrote:

...

This dang VersionNumber class has a bad equality operation.. Here's a fix.

http://bugs.squeak.org/view.php?id=2788

-- Matthew Fulmer -- http://mtfulmer.wordpress.com/ Help improve Squeak Documentation: http://wiki.squeak.org/squeak/808

I'll just go ahead and preface this with my own claims of 'tiredness' or at least frustration.

First the initial report was sent to the mailing list with something resembling the old scheme of bug/fix reporting (although not quite matching) rather than being posted to http://bugs.squeak.org/ . I sighed deeply but decided to simply ignore it.

Matthew kindly tried to move things in that direction, but with little apparent success.

Now I notice that you (Goran) don't once mention below the option of adding a report on Mantis under the SqueakMap category.

It has now been three years since we started with Mantis and deprecated the old report to list method. And yet new reports still occasionally appear on the list.

Yes, it's true that this is still the functionality that appears in the image. I'll accept that as the primary reason new reports continue to appear on the list.

All I can say is "Is anyone itching enough yet to actually build a system we will use?" I think it's pretty clear by now that Mantis does not represent such a system. I'll be the first to admit that Mantis doesn't cause me to want to scratch very much and so I have no confidence that I can define a system that will be accepted.

Too be more than a little blunt, it's time for all of us to put up or shut up. A few of us have complained about Mantis in the past. I'm not going to name names but I think it's time for those who don't feel like Mantis is a workable solution to start seriously thinking about the shape of a significantly more appropriate solution.

Ken

On Mon, 2007-11-05 at 22:22 +0200, goran@krampe.se wrote:

Yeah, I noticed this too - and fixed it too - and ended up forgetting I did it. :)

Let me see - in a perfect world (my perfect world) - this email (yours) would more or less be sent to me since I am a co-maintainer of VersionNumber on SM (using some nice mapping between PI and SM etc that actually is in place but not in use) and I would put the fix in and publish a new version of that darn package (which btw is a quite poor class all in all, when you look really, really close you realize that you can only branch ONCE per version number, ouch!!).

Eh, ok, so I just tried and... hmmph. Ok, so VersionNumber is an old .st file on SM. I made a new release of that. But... hmmm, so I notice two methods on VersionNumber are class extensions from SMBase - which could be moved into VersionNumber now that I am messing with it (haven't done that earlier).

(ok, so I tried using an old 3.8-6665 image and did "SMSqueakMap bootStrap" in it and then upgrade all - seems to end up fine with the fix in place)

But SMBase... has diverted off in 3.10 with changes that I am fairly sure are not backwards compatible with 3.8 (ChangeSorter vs ChangeOrganizer).

Ok, I give up there. I can't say I understand how SM is supposed to be maintained anymore - feel free to enlighten me. In fact - feel free to explain to us all how the new process/model for bug fixing, maintenance, packages etc is meant to work (IIRC that was the purpose of 3.10).

And no, not being sarcastic - possibly a bit tired.

regards, Gran

Mantis integration with the tools would be handy, I must admit... Think, having found a method with problems (doesn't cover larger areas of problem, I admit) to submit a report, categorised by default to its package, with, perhaps, a new better version... from a menu option in the browsers...

Or perhaps the debugger stack trace, rather than just an email option. Maybe we need something between Mantis and the tools that could match similar bug reports...

Just my thoughts at this time.

-----Original Message----- From: squeak-dev-bounces@lists.squeakfoundation.org [mailto:squeak-dev-bounces@lists.squeakfoundation.org]On Behalf Of Edgar J. De Cleene Sent: 05 November 2007 10:11 PM To: The general-purpose Squeak developers list Subject: Re: Mantis Experiment a Failure? Are we ready yet to move on? (was Re: [BUGFIX] SqueakMap is broken in 3.10)

El 11/5/07 6:46 PM, "Ken Causey" ken@kencausey.com escribió:

...

I'll just go ahead and preface this with my own claims of 'tiredness' or at least frustration.

First the initial report was sent to the mailing list with something resembling the old scheme of bug/fix reporting (although not quite matching) rather than being posted to http://bugs.squeak.org/ . I sighed deeply but decided to simply ignore it.

Matthew kindly tried to move things in that direction, but with little apparent success.

Now I notice that you (Goran) don't once mention below the option of adding a report on Mantis under the SqueakMap category.

It has now been three years since we started with Mantis and deprecated the old report to list method. And yet new reports still occasionally appear on the list.

Yes, it's true that this is still the functionality that appears in the image. I'll accept that as the primary reason new reports continue to appear on the list.

All I can say is "Is anyone itching enough yet to actually build a system we will use?" I think it's pretty clear by now that Mantis does not represent such a system. I'll be the first to admit that Mantis doesn't cause me to want to scratch very much and so I have no confidence that I can define a system that will be accepted.

Too be more than a little blunt, it's time for all of us to put up or shut up. A few of us have complained about Mantis in the past. I'm not going to name names but I think it's time for those who don't feel like Mantis is a workable solution to start seriously thinking about the shape of a significantly more appropriate solution.

Ken

Not a failure. I find Mantis very valuable and think (as I put in some mail) some bugs reports deserve showing to students for how a bug was find , what several Squeakers have to say about, proposed and developed solutions, etc.

If good Squeaker forget do reports or he is too busy , we could wait some more time. Or someone here notice and do the report.

Maybe some could do a Squeak front end to Mantis for easier and to hand way ? Edgar

P.S. And Squeakmap always works in 3.10 ...

The obvious question is whether gjallar is ready to be used for this?

tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Useful random insult:- His head whistles in a cross wind.

More positively, I did not notice mantis use was fading. Ken, don't you have some good stats about mantis showing it's not vain to repeat, repeat and repeat again?

So again, reporting a bug at http://bugs.squeak.org/ is easy. It is not much longer than posting on this list. Just spend a minute or two browsing related reports, it might be instructive.

Don't use "had no time" excuse. Mantis does save time indeed, far much than costs!

Of course we cannot force anyone to use it, only ask.. So please every one is kindly invited to submit.

Ken Causey a écrit :

I'll just go ahead and preface this with my own claims of 'tiredness' or at least frustration.

First the initial report was sent to the mailing list with something resembling the old scheme of bug/fix reporting (although not quite matching) rather than being posted to http://bugs.squeak.org/ . I sighed deeply but decided to simply ignore it.

Matthew kindly tried to move things in that direction, but with little apparent success.

Now I notice that you (Goran) don't once mention below the option of adding a report on Mantis under the SqueakMap category.

It has now been three years since we started with Mantis and deprecated the old report to list method. And yet new reports still occasionally appear on the list.

Yes, it's true that this is still the functionality that appears in the image. I'll accept that as the primary reason new reports continue to appear on the list.

All I can say is "Is anyone itching enough yet to actually build a system we will use?" I think it's pretty clear by now that Mantis does not represent such a system. I'll be the first to admit that Mantis doesn't cause me to want to scratch very much and so I have no confidence that I can define a system that will be accepted.

Too be more than a little blunt, it's time for all of us to put up or shut up. A few of us have complained about Mantis in the past. I'm not going to name names but I think it's time for those who don't feel like Mantis is a workable solution to start seriously thinking about the shape of a significantly more appropriate solution.

Ken

Thanks Rob.

Ken

On Mon, 2007-11-05 at 14:53 -0800, Rob Withers wrote:

I followed the directions and added a bug report to Mantis. It was auto-assigned to you Goran.

http://bugs.squeak.org/view.php?id=6754

I figure this ought to go into the update stream, if that's still around.

Rob