Do we kill the catalog?
Locked
123
123
 
|
2018-04-24 7:31 GMT+02:00 Sven Van Caekenberghe <[hidden email]>:
> > >> On 24 Apr 2018, at 00:49, monty <[hidden email]> wrote: >> >> +1. >> >> Replacing is sometimes necessary (Sven's stream work is an obvious example), but I don't see why Nautilus had to be junked rather than gradually evolved into Calypso. And I don't see why the catalog can't be gradually evolved into something better. Killing it off would be discouraging to the users who went through the trouble of adding #catalogXXX methods to their configs and publishing them to the meta repos. Why would they bother adopting whatever replaces the catalog if they think it will just get killed off too? >> >> ___ >> montyos.wordpress.com > > Nice blog, BTW! And I note: Publishing Configurations to Pharo Meta Repositories https://montyos.wordpress.com/2018/04/22/publishing-configurations-to-pharo-meta-repositories/ Thierry |
|
|
In reply to this post by Sven Van Caekenberghe-2
Hi all,
I think that we are all aligned. Stef may be "too strong" in his way of talking, but he raised an issue: we should probably do an iteration on the catalog. And I think we all agree on that? So, what would be a list of tasks for it? (I copy past what Stef put before) - Use Metacello API. Metacello new repository: 'zzz' configurationOf: 'XXX' loadVersion: #'YYY' This point should is easy. I can open an issue and work on it today. - Support for Baseline (no idea what it means)
- What else? Now, I also think about what to do with the meta-repo's. Should Pharo7 see MetaRepo for pharo 5 and pharo 6 and pharo 4? or should only see the one for pharo 7? What about Pharo 6? What are the meta-repo's it is looking for?
|
|
|
In reply to this post by monty-3
hi,
> On 24 Apr 2018, at 00:49, monty <[hidden email]> wrote: > > +1. > > Replacing is sometimes necessary (Sven's stream work is an obvious example), but I don't see why Nautilus had to be junked rather than gradually evolved into Calypso. that’s clearly because you didn’t see nautilus at the inside. In fact, calypso started like that… we wanted to refactor nautilus into something usable, but after losing two months of Denis work, we decided to go other direction, since it was just impossible to do it properly without doing it again. And the effort of doing it from scratch was less than “destroy the house for the inside but let the façade” that would have been working with Nautilus. Also, in the case of Nautilus (or any system browser), there is the adding complexity of refactoring within itself. Not everything can be refactored/patched up to the infinite. > And I don't see why the catalog can't be gradually evolved into something better. Killing it off would be discouraging to the users who went through the trouble of adding #catalogXXX methods to their configs and publishing them to the meta repos. Why would they bother adopting whatever replaces the catalog if they think it will just get killed off too? catalog was a “temporary piece” that stayed for years. it was never intend to stay because it has some important limitations. said that, it can evolve from the quasi-prototype it is now (I should know, it was me who did it) to a real catalog. curiously, I’m thinking one part of making it better would be simplify it, but that’s another story. I wonder why people is so concerned about the use of “killing catalog” phrase. If you know the history of Pharo, we never “kill” anything we do not have a replacement for it so it was never in the table a proposal of just removing it. But to be good it needs: - some better way of collect baselines than force to build a configuration - a better place to be - some kind of validation rules - etc, etc, etc. So well, in this case maybe refactor it is not an option, since many of the assumptions it has are no longer valid. cheers, Esteban > > ___ > montyos.wordpress.com > > >> Sent: Sunday, April 22, 2018 at 9:29 AM >> From: "Torsten Bergmann" <[hidden email]> >> To: [hidden email] >> Cc: "Pharo Development List" <[hidden email]> >> Subject: Re: [Pharo-dev] Do we kill the catalog? >> >> Hi Stef, >> >> do not worry - I'm fully relaxed. It is you - who suggested to kill the catalog - without having >> or proposing a suitable alternative yet. If you want to route the thread to talking about "idiots" >> and "assholes" you go offtopic and I guess I have nothing more to contribute to the discussion. >> >> If it was not clear from my writing: I only expressed my wish that we should better improve it >> instead of killing it. That's all. >> >> Because anytime we have a better idea we get more disruptive in what we provide ... and this >> leads to the unfortunate situation: >> >> - that we replace our own stuff with newer ones again and again (which can but must not be good >> as things get more diverse and complicated over time) >> - that users have to know how things historically have been progressed to find its way >> through Pharo (including the terms and tools we use) >> >> I would not see this as a rant - but rather a true effect we can spot on many places in our >> ecosystem already. >> >> Long story short: I would vote for the "fix" if the sole other option is to "remove". >> >> Bye >> T. >> >> >>> Gesendet: Samstag, 21. April 2018 um 15:31 Uhr >>> Von: "Stephane Ducasse" <[hidden email]> >>> An: "Pharo Development List" <[hidden email]> >>> Betreff: Re: [Pharo-dev] Do we kill the catalog? >>> >>> Torsten >>> >>> Sometimes you should relax. Because if you remember I pushed the >>> catalog (in fact I love it and we need a much more powerful one) >>> and I even improved it. I made it better to automatically add MC repositories. >>> Of course Thierry is exagerating. >>> >>> Now we should fix or remove it. Simple no? >>> >>> Now you can rant in your corner thinking that I'm asshole. >>> This is ok for me. >>> >>> Stef >>> >>> On Sat, Apr 21, 2018 at 3:06 PM, Torsten Bergmann <[hidden email]> wrote: >>>> Even there is no learning from the past - it's deja vu all over again... >>>> >>>> >>>>> Gesendet: Freitag, 20. April 2018 um 08:53 Uhr >>>>> Von: "Stephane Ducasse" <[hidden email]> >>>>> An: "Pharo Development List" <[hidden email]> >>>>> Betreff: Re: [Pharo-dev] Do we kill the catalog? >>>>> >>>>> Hi torsten >>>>> >>>>> the wise shows the moon, the idiot sees his finger. >>>>> >>>>> Stef >>>>> >>>>> >>>>> On Thu, Apr 19, 2018 at 10:38 AM, Torsten Bergmann <[hidden email]> wrote: >>>>>> Now today you had loading trouble with Smacc from Catalog ... and now directly want to >>>>>> kill catalog. Wow! >>>>>> >>>>>> Sometimes I have the impression that our community tries to reinvent itself each day >>>>>> doing it differently (but not only better) instead of extending, improving and supporting >>>>>> what we have have done before. Which sometimes is good ... but not always. >>>>>> >>>>>> Thx >>>>>> T. >>>>>> >>>>>>> Gesendet: Donnerstag, 19. April 2018 um 08:42 Uhr >>>>>>> Von: "Stephane Ducasse" <[hidden email]> >>>>>>> An: "Pharo Development List" <[hidden email]> >>>>>>> Betreff: [Pharo-dev] Do we kill the catalog? >>>>>>> >>>>>>> Hi guys >>>>>>> >>>>>>> What do we do with it? >>>>>>> What alternatives? >>>>>>> >>>>>>> Stef >>>>>>> >>>>>>> >>>>>> >>>>> >>>>> >>>> >>> >>> >> >> > |
|
Administrator
|
In reply to this post by Guillermo Polito
Guillermo Polito wrote
> Should we continue providing configurations that point to the > baselines? I've read some rants against it. > I know it may be "uncomfortable", but publishing some meta-data in a > repository in XXX technology is kind of the same... This seems like an implementation detail. IHMO what people are really complaining about is lack of tool support to map some standard, repeatable trustability system like a Config's #stable and #development to a baseline that refers to an exact snapshot of code (i.e. commitish in git parlance IIUC). ----- Cheers, Sean -- Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html
Cheers,
Sean |
|
|
Hi Sean, Guille,
2018-04-24 15:31 GMT+02:00 Sean P. DeNigris <[hidden email]>: > Guillermo Polito wrote >> Should we continue providing configurations that point to the >> baselines? I've read some rants against it. >> I know it may be "uncomfortable", but publishing some meta-data in a >> repository in XXX technology is kind of the same... Baselines are simple. If using a baseline in the Catalog forces it to be more complex (like need to add a repository url in the baseline), then do not do it and only support configurations. > This seems like an implementation detail. IHMO what people are really > complaining about is lack of tool support to map some standard, repeatable > trustability system like a Config's #stable and #development to a baseline > that refers to an exact snapshot of code (i.e. commitish in git parlance > IIUC). Sort of. A typical git-based development is one tag or branch per pharo version, hence one BaselineOf per supported Pharo version. A single configuration in the Catalog can triage on those baselines, and be used by multiple Pharo versions (4, 5, 6, 7, etc...) This would point out to having only one repository for the Catalog (for all Pharo versions); but the ability for the Catalog to filter out Configurations that are not updated on newer versions. Bonus: if you're tracking Pharo versions in your branches (if you have pharo5.0, pharo6.0, pharo7.0 branches, for example), it may be possible to write a generic url in the configurationOf, like that: repository: 'github://dalehenrich/filetree:pharo', SystemVersion current dottedMajorMinor ,'_dev/repository' Bonus+: you can also play with Unix / Windows dependent branches (because you can also query additional system attributes) Bonus++: Smalltalkhub repositories would still work... Feature: it would be nice to be able to test if the current version-dependent branch or tag exists on the target (i.e. query github), then it would be easy to write something like -> no pharo8.0 branch on repository ? Not available on Pharo8. Thierry > > > ----- > Cheers, > Sean > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html > |
|
|
In reply to this post by EstebanLM
> Sent: Tuesday, April 24, 2018 at 3:42 AM
> From: "Esteban Lorenzano" <[hidden email]> > To: "Pharo Development List" <[hidden email]> > Subject: Re: [Pharo-dev] Do we kill the catalog? > > hi, > > > > On 24 Apr 2018, at 00:49, monty <[hidden email]> wrote: > > > > +1. > > > > Replacing is sometimes necessary (Sven's stream work is an obvious example), but I don't see why Nautilus had to be junked rather than gradually evolved into Calypso. > > that’s clearly because you didn’t see nautilus at the inside. Actually, I did see the inside, to improve its SUnit integration. > In fact, calypso started like that… we wanted to refactor nautilus into something usable, but after losing two months of Denis work, we decided to go other direction, since it was just impossible to do it properly without doing it again. And the effort of doing it from scratch was less than “destroy the house for the inside but let the façade” that would have been working with Nautilus. Also, in the case of Nautilus (or any system browser), there is the adding complexity of refactoring within itself. > Not everything can be refactored/patched up to the infinite. > > > And I don't see why the catalog can't be gradually evolved into something better. Killing it off would be discouraging to the users who went through the trouble of adding #catalogXXX methods to their configs and publishing them to the meta repos. Why would they bother adopting whatever replaces the catalog if they think it will just get killed off too? > > catalog was a “temporary piece” that stayed for years. > it was never intend to stay because it has some important limitations. > said that, it can evolve from the quasi-prototype it is now (I should know, it was me who did it) to a real catalog. > > curiously, I’m thinking one part of making it better would be simplify it, but that’s another story. > > I wonder why people is so concerned about the use of “killing catalog” phrase. If you know the history of Pharo, we never “kill” anything we do not have a replacement for it so it was never in the table a proposal of just removing it. But to be good it needs: > > - some better way of collect baselines than force to build a configuration > - a better place to be > - some kind of validation rules > - etc, etc, etc. > > So well, in this case maybe refactor it is not an option, since many of the assumptions it has are no longer valid. > > cheers, > Esteban > > > > > ___ > > montyos.wordpress.com > > > > > >> Sent: Sunday, April 22, 2018 at 9:29 AM > >> From: "Torsten Bergmann" <[hidden email]> > >> To: [hidden email] > >> Cc: "Pharo Development List" <[hidden email]> > >> Subject: Re: [Pharo-dev] Do we kill the catalog? > >> > >> Hi Stef, > >> > >> do not worry - I'm fully relaxed. It is you - who suggested to kill the catalog - without having > >> or proposing a suitable alternative yet. If you want to route the thread to talking about "idiots" > >> and "assholes" you go offtopic and I guess I have nothing more to contribute to the discussion. > >> > >> If it was not clear from my writing: I only expressed my wish that we should better improve it > >> instead of killing it. That's all. > >> > >> Because anytime we have a better idea we get more disruptive in what we provide ... and this > >> leads to the unfortunate situation: > >> > >> - that we replace our own stuff with newer ones again and again (which can but must not be good > >> as things get more diverse and complicated over time) > >> - that users have to know how things historically have been progressed to find its way > >> through Pharo (including the terms and tools we use) > >> > >> I would not see this as a rant - but rather a true effect we can spot on many places in our > >> ecosystem already. > >> > >> Long story short: I would vote for the "fix" if the sole other option is to "remove". > >> > >> Bye > >> T. > >> > >> > >>> Gesendet: Samstag, 21. April 2018 um 15:31 Uhr > >>> Von: "Stephane Ducasse" <[hidden email]> > >>> An: "Pharo Development List" <[hidden email]> > >>> Betreff: Re: [Pharo-dev] Do we kill the catalog? > >>> > >>> Torsten > >>> > >>> Sometimes you should relax. Because if you remember I pushed the > >>> catalog (in fact I love it and we need a much more powerful one) > >>> and I even improved it. I made it better to automatically add MC repositories. > >>> Of course Thierry is exagerating. > >>> > >>> Now we should fix or remove it. Simple no? > >>> > >>> Now you can rant in your corner thinking that I'm asshole. > >>> This is ok for me. > >>> > >>> Stef > >>> > >>> On Sat, Apr 21, 2018 at 3:06 PM, Torsten Bergmann <[hidden email]> wrote: > >>>> Even there is no learning from the past - it's deja vu all over again... > >>>> > >>>> > >>>>> Gesendet: Freitag, 20. April 2018 um 08:53 Uhr > >>>>> Von: "Stephane Ducasse" <[hidden email]> > >>>>> An: "Pharo Development List" <[hidden email]> > >>>>> Betreff: Re: [Pharo-dev] Do we kill the catalog? > >>>>> > >>>>> Hi torsten > >>>>> > >>>>> the wise shows the moon, the idiot sees his finger. > >>>>> > >>>>> Stef > >>>>> > >>>>> > >>>>> On Thu, Apr 19, 2018 at 10:38 AM, Torsten Bergmann <[hidden email]> wrote: > >>>>>> Now today you had loading trouble with Smacc from Catalog ... and now directly want to > >>>>>> kill catalog. Wow! > >>>>>> > >>>>>> Sometimes I have the impression that our community tries to reinvent itself each day > >>>>>> doing it differently (but not only better) instead of extending, improving and supporting > >>>>>> what we have have done before. Which sometimes is good ... but not always. > >>>>>> > >>>>>> Thx > >>>>>> T. > >>>>>> > >>>>>>> Gesendet: Donnerstag, 19. April 2018 um 08:42 Uhr > >>>>>>> Von: "Stephane Ducasse" <[hidden email]> > >>>>>>> An: "Pharo Development List" <[hidden email]> > >>>>>>> Betreff: [Pharo-dev] Do we kill the catalog? > >>>>>>> > >>>>>>> Hi guys > >>>>>>> > >>>>>>> What do we do with it? > >>>>>>> What alternatives? > >>>>>>> > >>>>>>> Stef > >>>>>>> > >>>>>>> > >>>>>> > >>>>> > >>>>> > >>>> > >>> > >>> > >> > >> > > > > > ___ montyos.wordpress.com |
|
|
In reply to this post by Sven Van Caekenberghe-2
Thanks!
___ montyos.wordpress.com > Sent: Tuesday, April 24, 2018 at 1:31 AM > From: "Sven Van Caekenberghe" <[hidden email]> > To: "Pharo Development List" <[hidden email]> > Subject: Re: [Pharo-dev] Do we kill the catalog? > > > > > On 24 Apr 2018, at 00:49, monty <[hidden email]> wrote: > > > > +1. > > > > Replacing is sometimes necessary (Sven's stream work is an obvious example), but I don't see why Nautilus had to be junked rather than gradually evolved into Calypso. And I don't see why the catalog can't be gradually evolved into something better. Killing it off would be discouraging to the users who went through the trouble of adding #catalogXXX methods to their configs and publishing them to the meta repos. Why would they bother adopting whatever replaces the catalog if they think it will just get killed off too? > > > > ___ > > montyos.wordpress.com > > Nice blog, BTW! > > Your XML work is much appreciated. > > >> Sent: Sunday, April 22, 2018 at 9:29 AM > >> From: "Torsten Bergmann" <[hidden email]> > >> To: [hidden email] > >> Cc: "Pharo Development List" <[hidden email]> > >> Subject: Re: [Pharo-dev] Do we kill the catalog? > >> > >> Hi Stef, > >> > >> do not worry - I'm fully relaxed. It is you - who suggested to kill the catalog - without having > >> or proposing a suitable alternative yet. If you want to route the thread to talking about "idiots" > >> and "assholes" you go offtopic and I guess I have nothing more to contribute to the discussion. > >> > >> If it was not clear from my writing: I only expressed my wish that we should better improve it > >> instead of killing it. That's all. > >> > >> Because anytime we have a better idea we get more disruptive in what we provide ... and this > >> leads to the unfortunate situation: > >> > >> - that we replace our own stuff with newer ones again and again (which can but must not be good > >> as things get more diverse and complicated over time) > >> - that users have to know how things historically have been progressed to find its way > >> through Pharo (including the terms and tools we use) > >> > >> I would not see this as a rant - but rather a true effect we can spot on many places in our > >> ecosystem already. > >> > >> Long story short: I would vote for the "fix" if the sole other option is to "remove". > >> > >> Bye > >> T. > >> > >> > >>> Gesendet: Samstag, 21. April 2018 um 15:31 Uhr > >>> Von: "Stephane Ducasse" <[hidden email]> > >>> An: "Pharo Development List" <[hidden email]> > >>> Betreff: Re: [Pharo-dev] Do we kill the catalog? > >>> > >>> Torsten > >>> > >>> Sometimes you should relax. Because if you remember I pushed the > >>> catalog (in fact I love it and we need a much more powerful one) > >>> and I even improved it. I made it better to automatically add MC repositories. > >>> Of course Thierry is exagerating. > >>> > >>> Now we should fix or remove it. Simple no? > >>> > >>> Now you can rant in your corner thinking that I'm asshole. > >>> This is ok for me. > >>> > >>> Stef > >>> > >>> On Sat, Apr 21, 2018 at 3:06 PM, Torsten Bergmann <[hidden email]> wrote: > >>>> Even there is no learning from the past - it's deja vu all over again... > >>>> > >>>> > >>>>> Gesendet: Freitag, 20. April 2018 um 08:53 Uhr > >>>>> Von: "Stephane Ducasse" <[hidden email]> > >>>>> An: "Pharo Development List" <[hidden email]> > >>>>> Betreff: Re: [Pharo-dev] Do we kill the catalog? > >>>>> > >>>>> Hi torsten > >>>>> > >>>>> the wise shows the moon, the idiot sees his finger. > >>>>> > >>>>> Stef > >>>>> > >>>>> > >>>>> On Thu, Apr 19, 2018 at 10:38 AM, Torsten Bergmann <[hidden email]> wrote: > >>>>>> Now today you had loading trouble with Smacc from Catalog ... and now directly want to > >>>>>> kill catalog. Wow! > >>>>>> > >>>>>> Sometimes I have the impression that our community tries to reinvent itself each day > >>>>>> doing it differently (but not only better) instead of extending, improving and supporting > >>>>>> what we have have done before. Which sometimes is good ... but not always. > >>>>>> > >>>>>> Thx > >>>>>> T. > >>>>>> > >>>>>>> Gesendet: Donnerstag, 19. April 2018 um 08:42 Uhr > >>>>>>> Von: "Stephane Ducasse" <[hidden email]> > >>>>>>> An: "Pharo Development List" <[hidden email]> > >>>>>>> Betreff: [Pharo-dev] Do we kill the catalog? > >>>>>>> > >>>>>>> Hi guys > >>>>>>> > >>>>>>> What do we do with it? > >>>>>>> What alternatives? > >>>>>>> > >>>>>>> Stef > >>>>>>> > >>>>>>> > >>>>>> > >>>>> > >>>>> > >>>> > >>> > >>> > >> > >> > > > > > |
|
Administrator
|
In reply to this post by Thierry Goubier
Thierry Goubier wrote
> If using a baseline in the Catalog forces it to be more complex… I didn't think we were discussing changing baselines themselves, just a better adapter layer than hand-edited Configs Thierry Goubier wrote > A typical git-based development is one tag or branch per pharo > version, hence one BaselineOf per supported Pharo version. Hmm, I didn't know that. I have not been using that workflow, but '*-PlatformPharo60' packages. I wonder what the tradeoffs are… ----- Cheers, Sean -- Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html
Cheers,
Sean |
|
|
Just for the record
Feel free to review or to propose other PRs, Guille |
«
Return to Pharo Smalltalk Developers
|
1 view|%1 views
| Free forum by Nabble | Edit this page |