February 2015 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] get.pharo.org/travis+vm failing: Package 'pharo-vm-core:i386' has no installation candidate

by Markus Fritsche

On 09.02.2015 03:54, Dale Henrichs wrote: > It looks like the following script started failing within the last 5 > days sometime: > > wget --quiet -O - get.pharo.org/travis+vm | bash > > getting this error (on ubunutu12.04): I requested deletion of my recently uploaded packages. Launchpad will take a few hours for the deletion request to proceed though. There are some changes to the way the packages work that I need to sort out before I can make a proper update of the stable ppa. Best regards, Markus

Feb. 10, 2015

Re: [Pharo-dev] Installing Pharo on Ubuntu

by Markus Fritsche

Hi Damien, I updated the pharo launcher with the sources from http://files.pharo.org/platform/launcher/blessed/PharoLauncher-user-0.2.2-2… and get a walkback during startup: Author: MarkusFritsche Date: 2015-02-10T19:28:38.179496+01:00 ByteString(Object)>>doesNotUnderstand: #asLegalSelector PhLLaunchImageCommand class(PhLCommand class)>>specId ByteSymbol(Symbol)>>value: Array(SequenceableCollection)>>collect: PhLCommandToolbar>>commandClasses: PhLTitledTreeModel>>toolbarCommandClasses: PharoLauncher>>initializeImageAndTemplates [ self initializeImageAndTemplates ] in PharoLauncher>>initializeWidgets BlockClosure>>cull: BlockClosure>>cull:cull: [ action cull: anAnnouncement cull: announcer ] in AnnouncementSubscription>>deliver: BlockClosure>>on:do: BlockClosure>>on:fork: AnnouncementSubscription>>deliver: [ subscription deliver: anAnnouncement ] in SubscriptionRegistry>>deliver:to:startingAt: BlockClosure>>ifCurtailed: SubscriptionRegistry>>deliver:to:startingAt: SubscriptionRegistry>>deliver:to: SubscriptionRegistry>>deliver: Announcer>>announce:

Feb. 10, 2015

14616 Package Tags causing Monticello merge conflict

by Ben Coman

I just produced a rough proof of concept to fix this issue. Its not ready to integrate because I don't grasp the architecture. I just hacked and smashed around with the debugger until it worked. It ended up needing only two methods modified, so hopefully it nicely defines the problem for a knowledgeable person to produce a proper solution. About a dozen small issues (e.g. renaming packages) are dependent on this, so can anyone help out? cheers -ben

Feb. 10, 2015

Re: [Pharo-dev] Spotter feature request - fully defined Class>>method lookup

by Ben Coman

Great. Thanks Doru. 100 minute turn around. wow! :) Its 2AM here so I'll try it out tomorrow. Now sorry I forgot to mention that both these forms * Collection>>#do: * Collection>>do: would be useful since individuals vary their use in documentation. Actually I mostly use the latter, but the case that sparked the post used the former. cheers -ben On Wed, Feb 11, 2015 at 1:21 AM, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: > Excellent! > > Thank you Tudor. > Esteban A. Maringolo > > > 2015-02-10 14:00 GMT-03:00 Tudor Girba <tudor(a)tudorgirba.com>: > > Hi, > > > > Indeed. Nice idea. Available in the latest GT-SpotterExtensions-Core: > > Name: GT-SpotterExtensions-Core-TudorGirba.57 > > Author: TudorGirba > > Time: 10 February 2015, 5:59:41.775209 pm > > UUID: 7de47651-7d17-44cd-ac70-467955fa903c > > Ancestors: GT-SpotterExtensions-Core-TudorGirba.56 > > > > added spotter for concrete method reference like: > > Collection>>#do: (thanks Ben for the suggestion) > > > > Cheers, > > Doru > > > > > > > > On Tue, Feb 10, 2015 at 4:15 PM, Ben Coman <btc(a)openinworld.com> wrote: > >> > >> When I am reading Fogbugz issues, often there are method references > using > >> the common convention "MCDefinitionIndex>>#add: for a fully defined > >> reference. > >> > >> It would be nice to be able to copy/paste that fully defined reference > >> into Spotter and get delivered straight to that method. You might still > >> list other implementors of the method, but the fully defined reference > would > >> be at the top, or somehow highlighted. > >> > >> I might get around to trying to code this myself, but I mention it in > case > >> it catches the interest of Spotter core team. > >> > >> cheers -ben > >> > >> > > > > > > > > -- > > www.tudorgirba.com > > > > "Every thing has its own flow" > >

Feb. 10, 2015

Re: [Pharo-dev] Matrix API question for a native english

by Ben Coman

rowCount sounds like its the total number of rows. rowNumber sounds like a particular row. On Wed, Feb 11, 2015 at 12:12 AM, Johann Hibschman <johannh(a)gmail.com> wrote: > rowCount is better. "count" clearly refers to the cardinality. "number" in > post-fix, like "rowNumber", sounds more like an ordinal, closer to > "rowIndex", at least to my ears. "numberOfRows" is once again clearly > cardinal, not ordinal, but it's longer. > > Cheers, > Johann > > On Mon, Feb 9, 2015 at 4:24 PM, stepharo <stepharo(a)free.fr> wrote: > >> thank chris. >> >> I can tell you that I would be even more productive if I would be a >> native speaker.... >> So should we use rowCount for method argument? >> because we got rows: rows meaning rows: rowCount and I rewrote them as >> rows: rowNumber but it could be better to have >> rows: rowCount. >> >> tx >> Stef >> >> Le 9/2/15 22:11, Chris Muller a Ã©crit : >> >> On Mon, Feb 9, 2015 at 3:05 PM, stepharo <stepharo(a)free.fr> wrote: >>> >>>> Hi >>>> >>>> a matrix has rowCount and columnCount and I was wondering if this api >>>> is the >>>> way to get >>>> the number of columns and rows of a matrix. >>>> Would rowNumber and columnNumber be better? >>>> >>> I think rowCount is better than rowNumber for that. >>> >>> You used "number of rows" in your own English right there, so >>> "numberOfRows" would be just as descriptive as rowCount, but not >>> better and certainly not worth changing the API, IMO. >>> >>> >>> >> >> >

Feb. 10, 2015

Re: [Pharo-dev] Spotter idea - exclusion filter and google search syntax

by Ben Coman

On Sun, Feb 8, 2015 at 10:36 AM, Ben Coman <btc(a)openinworld.com> wrote: > > > On Sun, Feb 8, 2015 at 8:53 AM, Nicolai Hess <nicolaihess(a)web.de> wrote: > >> >> >> 2015-02-08 0:32 GMT+01:00 Ben Coman <btc(a)openinworld.com>: >> >>> >>> Just a random idea for Spotter, being able to exclude items from the >>> search using a hyphen/minus sign. >>> >>> I don't have a use case right now, but it "feels" right, so maybe one >>> will turn up soon. Its a broadly used paradigm per google search ( >>> https://support.google.com/mail/answer/7190?hl=en) >>> >>> >>> Also, it would be useful to be able to search on multiple terms similar >>> to a regular web search. >>> >>> A slightly contrived use case is that I want to investigate drag >>> handling of morphs. So I guess entities will have the word "drag" and then >>> maybe "handler" or "handle" but in which order? Will it be #dragHandler or >>> #handleDrag or #handleEventDrag or #handlesMouseOverDragging: ? A single >>> search term restricts me to having to guess and manually retry each >>> possible order. Searching on "drag hand" and getting all possible >>> ordering would be real nice. >>> >>> cheers -ben >>> >> >> +1 for exclusion filter >> (You can already combine search terms - sequentiel. >> 1. search drag and in the result list (implementors) >> 2. search again with the term handler) >> >> > Thanks for the tip, thats cool. Still, doing it from the first window > would be more immediate. > cheers -ben > > > Just to follow up, I have been using that feature and I am finding that after searching for the method name, I can't just press CMD-RightArrow since that takes me to a list of classes. I need to array down several times to get to implementors and then press CMD-RightArrow, only after which I can type my second term. And presumably if I then wanted to do the same two term search for classes, I would need to CMD-LeftArrow, arrow up several times, CMD-RIghtArrow and type the second term again. It seems less friction if I could just type both terms on the first page, and get all search types (i.e. classes/implementors) matching both terms. The paradigm of a web search. cheers -ben

Feb. 10, 2015

Re: [Pharo-dev] Spotter feature request - fully defined Class>>method lookup

by Esteban A. Maringolo

Excellent! Thank you Tudor. Esteban A. Maringolo 2015-02-10 14:00 GMT-03:00 Tudor Girba <tudor(a)tudorgirba.com>: > Hi, > > Indeed. Nice idea. Available in the latest GT-SpotterExtensions-Core: > Name: GT-SpotterExtensions-Core-TudorGirba.57 > Author: TudorGirba > Time: 10 February 2015, 5:59:41.775209 pm > UUID: 7de47651-7d17-44cd-ac70-467955fa903c > Ancestors: GT-SpotterExtensions-Core-TudorGirba.56 > > added spotter for concrete method reference like: > Collection>>#do: (thanks Ben for the suggestion) > > Cheers, > Doru > > > > On Tue, Feb 10, 2015 at 4:15 PM, Ben Coman <btc(a)openinworld.com> wrote: >> >> When I am reading Fogbugz issues, often there are method references using >> the common convention "MCDefinitionIndex>>#add: for a fully defined >> reference. >> >> It would be nice to be able to copy/paste that fully defined reference >> into Spotter and get delivered straight to that method. You might still >> list other implementors of the method, but the fully defined reference would >> be at the top, or somehow highlighted. >> >> I might get around to trying to code this myself, but I mention it in case >> it catches the interest of Spotter core team. >> >> cheers -ben >> >> > > > > -- > www.tudorgirba.com > > "Every thing has its own flow"

Feb. 10, 2015

Re: [Pharo-dev] Spotter feature request - fully defined Class>>method lookup

by Tudor Girba

Hi, Indeed. Nice idea. Available in the latest GT-SpotterExtensions-Core: Name: GT-SpotterExtensions-Core-TudorGirba.57 Author: TudorGirba Time: 10 February 2015, 5:59:41.775209 pm UUID: 7de47651-7d17-44cd-ac70-467955fa903c Ancestors: GT-SpotterExtensions-Core-TudorGirba.56 added spotter for concrete method reference like: Collection>>#do: (thanks Ben for the suggestion) Cheers, Doru On Tue, Feb 10, 2015 at 4:15 PM, Ben Coman <btc(a)openinworld.com> wrote: > When I am reading Fogbugz issues, often there are method references using > the common convention "MCDefinitionIndex>>#add: for a fully defined > reference. > > It would be nice to be able to copy/paste that fully defined reference > into Spotter and get delivered straight to that method. You might still > list other implementors of the method, but the fully defined reference > would be at the top, or somehow highlighted. > > I might get around to trying to code this myself, but I mention it in case > it catches the interest of Spotter core team. > > cheers -ben > > > -- www.tudorgirba.com "Every thing has its own flow"

Feb. 10, 2015

Re: [Pharo-dev] Matrix API question for a native english

by Johann Hibschman

rowCount is better. "count" clearly refers to the cardinality. "number" in post-fix, like "rowNumber", sounds more like an ordinal, closer to "rowIndex", at least to my ears. "numberOfRows" is once again clearly cardinal, not ordinal, but it's longer. Cheers, Johann On Mon, Feb 9, 2015 at 4:24 PM, stepharo <stepharo(a)free.fr> wrote: > thank chris. > > I can tell you that I would be even more productive if I would be a native > speaker.... > So should we use rowCount for method argument? > because we got rows: rows meaning rows: rowCount and I rewrote them as > rows: rowNumber but it could be better to have > rows: rowCount. > > tx > Stef > > Le 9/2/15 22:11, Chris Muller a Ã©crit : > > On Mon, Feb 9, 2015 at 3:05 PM, stepharo <stepharo(a)free.fr> wrote: >> >>> Hi >>> >>> a matrix has rowCount and columnCount and I was wondering if this api is >>> the >>> way to get >>> the number of columns and rows of a matrix. >>> Would rowNumber and columnNumber be better? >>> >> I think rowCount is better than rowNumber for that. >> >> You used "number of rows" in your own English right there, so >> "numberOfRows" would be just as descriptive as rowCount, but not >> better and certainly not worth changing the API, IMO. >> >> >> > >

Feb. 10, 2015

Fwd: [Pharo CI] Jenkins down 10/11 Feb

by Marcus Denker

This is todayâ¦ > > There will a big update of the Inria CI infrastructure. > > On 10 Feb the system will go offline 5PM Paris time, it will > be online the 12th hopefully latest. > > Why? For now the system just supports Win and Linux. > This update will add the base for MacOS vitual servers. > (If you wonder how it works now: we have a non-virtual > Mac in our office that runs all mac buildsâ¦) > > Marcus

Feb. 10, 2015