- Pharo-dev - lists.pharo.org

Re: [Pharo-dev] STON questions

by stepharo

I do not have the exact points that christophe wants. Now he wants to store package dependencies into package manifest. And I told him to use STON. Now we should probably not haeve an explicit reference to the dependency class Stef > Hi Christophe, > > Sorry for the delay, I was on holiday. > > On 15 Jul 2014, at 14:46, Christophe Demarey <christophe.demarey(a)inria.fr> wrote: > >> Hi, >> >> I read the STON paper and the STON chapter in EnterprisePharo but I did not find answers to some questions: >> â¢ how do you tell STON to not serialize some variables of an object? > To achieve that you have to write your own encoding by overwriting #stonOn: and #fromSton: (see implementors). > >> â¢ is there a way to avoid the class tag? What I mean is that I would like a synthetic STON String. If a STON entry has a class tag, STON knows which class to instantiate. This class may also know which classes to instantiate for its i-var. In this case, the class tag may be omitted. Maybe I should redefine fromSton;, but in this case, is it possible to avoid a manual parsing of the string? > No, that is not possible. I fail to see how this would be useful or needed, but maybe I just don't understand your question. > >> â¢ is there a way to do an alias for an i-var (just like for the class name)? > To achieve that you have to write your own encoding by overwriting #stonOn: and #fromSton: (see implementors). > >> â¢ is there a way to skip the serialization of an i-var if its value is nil (or an empty collection)? > Overwrite #stonShouldWriteNilInstVars for your object to return true. For empty collections there is no provision. > >> Regards, >> Christophe. > Sven > > > >

July 23, 2014

Re: [Pharo-dev] STON questions

by Sven Van Caekenberghe

Hi Christophe, Sorry for the delay, I was on holiday. On 15 Jul 2014, at 14:46, Christophe Demarey <christophe.demarey(a)inria.fr> wrote: > Hi, > > I read the STON paper and the STON chapter in EnterprisePharo but I did not find answers to some questions: > â¢ how do you tell STON to not serialize some variables of an object? To achieve that you have to write your own encoding by overwriting #stonOn: and #fromSton: (see implementors). > â¢ is there a way to avoid the class tag? What I mean is that I would like a synthetic STON String. If a STON entry has a class tag, STON knows which class to instantiate. This class may also know which classes to instantiate for its i-var. In this case, the class tag may be omitted. Maybe I should redefine fromSton;, but in this case, is it possible to avoid a manual parsing of the string? No, that is not possible. I fail to see how this would be useful or needed, but maybe I just don't understand your question. > â¢ is there a way to do an alias for an i-var (just like for the class name)? To achieve that you have to write your own encoding by overwriting #stonOn: and #fromSton: (see implementors). > â¢ is there a way to skip the serialization of an i-var if its value is nil (or an empty collection)? Overwrite #stonShouldWriteNilInstVars for your object to return true. For empty collections there is no provision. > Regards, > Christophe. Sven

July 23, 2014

Re: [Pharo-dev] Project Watch on SThub

by stepharo

THub allows one to "watch" a project, meaning one can get infos on the commits and history of watched projects under in his own home section in "Recent watched events". Unfortunately it does not update, always displays events from 2013 and either it is my profile or a general bug in STHub... Any comments? Do not expect much work around STHub since nicolas moved to another job and we do not have the resources to do that locally. if somebody wants to help it would be a great help. Stef

July 23, 2014

Re: [Pharo-dev] transformmorph and submorph bounds

by Thierry Goubier

2014-07-23 12:09 GMT+02:00 Tudor Girba <tudor(a)tudorgirba.com>: > Hi, > > Ahh. I found it. The proper selector should not be submorphBounds but > positionInWorld. This one works properly. > boundsInWorld :) Thierry

July 23, 2014

Coverage testing options

by Pavel Krivanek

Hi, what are the current working options for the tests coverage reporting? For testing of the minimal image I tried to use the standard code simulator but as you may expect the results are not perfect (it cannot hanlde tests that use should:rise, it's very slow etc.) https://ci.inria.fr/pharo/view/4.0-Bootstrap/job/Pharo-4.0-Bootstrap-Step-2… Cheers, -- Pavel

July 23, 2014

Project Watch on SThub

by Torsten Bergmann

STHub allows one to "watch" a project, meaning one can get infos on the commits and history of watched projects under in his own home section in "Recent watched events". Unfortunately it does not update, always displays events from 2013 and either it is my profile or a general bug in STHub... Any comments?

July 23, 2014

[ANN] ODBC package on STHub

by Torsten Bergmann

The ODBC package is now migrated from SqueakSource to SmalltalkHub and available under http://smalltalkhub.com/#!/~PharoExtras/ODBC I also cleaned up a little bit and made some stuff working on Pharo 3.0 again. You can now als easily load it from the config browser. Repo is open for contributions. Have fun. Thanks T.

July 23, 2014

Re: [Pharo-dev] transformmorph and submorph bounds

by Tudor Girba

Hi, On Wed, Jul 23, 2014 at 11:45 AM, Thierry Goubier <thierry.goubier(a)gmail.com > wrote: > > > > 2014-07-23 11:10 GMT+02:00 Tudor Girba <tudor(a)tudorgirba.com>: > >> Hi, >> >> >> >> >> >>> This seems correct to me: a transformMorph is the one which shift its >>> contents (and reset the origin to its top left corner ?). >>> >> >> Not really. The contract of submorphBounds should be that it returns an >> absolute coordinate, as far as I understand. >> > > In owner coordinates it says. (TransformMorph) > > However, it effectively returns a 0@0 origin. > > >> >> >>> Why do you use a TransformMorph there? >>> >> >> It is used for example in ScrollPane. This means that if you embed a >> TextMorph inside a ScrollPane, you will have the same problem with >> completion: >> >> scrollPane := ScrollPane new. >> pasteUpMorph := PasteUpMorph new. >> scrollPane scroller addMorph: pasteUpMorph. >> scrollPane openInWindow top: 100; left: 100. >> >> pasteUpMorph submorphBounds "(0@0) corner: (50@40)" >> > > Yes. > > What do you want to do about it? Rewrite a scroll pane in a different way? > Or make sure that everything which creates over-morphs ask for global > coordinates? > Ahh. I found it. The proper selector should not be submorphBounds but positionInWorld. This one works properly. > I guess this has also something to do with the glamour bug inside tabs, no? > Not inside tabs, but inside the a Finder or a Pager. Now, completion work properly! Cheers, Doru > Thierry > > -- www.tudorgirba.com "Every thing has its own flow"

July 23, 2014

Re: [Pharo-dev] SmartBreakpoints alpha release

by Tudor Girba

Just a point. We should definitely not throw away TxText :). My message was that we should stop investing in the TextMorph. I suggested Rubric because it is available now and can be integrated with visible improvements over what exists. But, TxText should be the future. Perhaps we can already commit to it for Pharo 4.0 and then we make it a goal for the whole community to get it finished. I would be supportive for such a decision. Doru On Wed, Jul 23, 2014 at 11:46 AM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > Hi, > > Not yet :(. It's the first thing after releasing the current > implementation of GT and Moose. > > Doru > > > On Wed, Jul 23, 2014 at 11:22 AM, stepharo <stepharo(a)free.fr> wrote: > >> >> On 22/7/14 14:43, Tudor Girba wrote: >> >> First of all: Excellent job! >> >> About the editor: I would indeed suggest adopting Rubric as a solution >> for now (until the new Text widget comes around). We use it extensively in >> GT and I can attest that it is better than PluggableTextMorph. >> >> >> doru did you loaded TxText? >> and TxViewContainer example2 >> >> >> Doru >> >> >> On Tue, Jul 22, 2014 at 2:05 PM, Marcus Denker <marcus.denker(a)inria.fr> >> wrote: >> >>> >>> On 22 Jul 2014, at 13:57, stepharo <stepharo(a)free.fr> wrote: >>> >>> >>> On 22/7/14 10:41, Clara Allende wrote: >>> >>> Hi guys! >>> Finally the metaLink-based breakpoints are "ready". I have made some >>> explanations on my blog >>> <https://clariallende.wordpress.com/2014/07/21/smartbreakpoints-alpha-releas…>, >>> but basically now you have the option "break here" in SmartSuggestions menu. >>> So it's still experimental, and even though I have some tests, it would >>> be nice if you can test it, because I can't cover all possible use cases :) >>> And any feedback is most appreciated! >>> >>> Specially, since eventually we will have several kinds of breakpoints, >>> I would like to ask in which way would you like to choose them (a pop menu? >>> a sublist?) >>> >>> Thanks in advance! >>> >>> PS: The script for downloading the code is in the blog post :P >>> PS2: After talking with Anne and Baptiste, things that are not there >>> *yet* but I want to have: >>> * a highlight on the method/messageSend where the breakpoint is inserted >>> * expose persistent breakpoints (it is there but the default is to halt >>> once) >>> * line breakpoints (not sure when, for now the breakpoints are AST based) >>> * integration with TextEditor, once it is done ;) >>> >>> >>> Can you grab igor and sit with him to get this point? >>> >>> Another idea would be to use the editor in Moose (Rubrik?). >>> Better a solution now than to wait forever for perfectionâ¦ >>> >>> Marcus >>> >>> >> >> >> -- >> www.tudorgirba.com >> >> "Every thing has its own flow" >> >> >> > > > -- > www.tudorgirba.com > > "Every thing has its own flow" > -- www.tudorgirba.com "Every thing has its own flow"

July 23, 2014

Re: [Pharo-dev] SmartBreakpoints alpha release

by Tudor Girba

Hi, Not yet :(. It's the first thing after releasing the current implementation of GT and Moose. Doru On Wed, Jul 23, 2014 at 11:22 AM, stepharo <stepharo(a)free.fr> wrote: > > On 22/7/14 14:43, Tudor Girba wrote: > > First of all: Excellent job! > > About the editor: I would indeed suggest adopting Rubric as a solution > for now (until the new Text widget comes around). We use it extensively in > GT and I can attest that it is better than PluggableTextMorph. > > > doru did you loaded TxText? > and TxViewContainer example2 > > > Doru > > > On Tue, Jul 22, 2014 at 2:05 PM, Marcus Denker <marcus.denker(a)inria.fr> > wrote: > >> >> On 22 Jul 2014, at 13:57, stepharo <stepharo(a)free.fr> wrote: >> >> >> On 22/7/14 10:41, Clara Allende wrote: >> >> Hi guys! >> Finally the metaLink-based breakpoints are "ready". I have made some >> explanations on my blog >> <https://clariallende.wordpress.com/2014/07/21/smartbreakpoints-alpha-releas…>, >> but basically now you have the option "break here" in SmartSuggestions menu. >> So it's still experimental, and even though I have some tests, it would >> be nice if you can test it, because I can't cover all possible use cases :) >> And any feedback is most appreciated! >> >> Specially, since eventually we will have several kinds of breakpoints, >> I would like to ask in which way would you like to choose them (a pop menu? >> a sublist?) >> >> Thanks in advance! >> >> PS: The script for downloading the code is in the blog post :P >> PS2: After talking with Anne and Baptiste, things that are not there >> *yet* but I want to have: >> * a highlight on the method/messageSend where the breakpoint is inserted >> * expose persistent breakpoints (it is there but the default is to halt >> once) >> * line breakpoints (not sure when, for now the breakpoints are AST based) >> * integration with TextEditor, once it is done ;) >> >> >> Can you grab igor and sit with him to get this point? >> >> Another idea would be to use the editor in Moose (Rubrik?). >> Better a solution now than to wait forever for perfectionâ¦ >> >> Marcus >> >> > > > -- > www.tudorgirba.com > > "Every thing has its own flow" > > > -- www.tudorgirba.com "Every thing has its own flow"

July 23, 2014