- Pharo-users - lists.pharo.org

Re: [Pharo-users] UFFI and autoRelease

by Serge Stinckwich

I don't understand the default finalizeResourceData: handle handle isNull ifTrue: [ ^ self ]. handle free. handle beNull when handle is an ExternaData, free is defined as in ExternalStructure class as: free "Free the handle pointed to by the receiver" (handle ~~ nil and:[handle isExternalAddress]) ifTrue:[handle free]. handle := nil. but beNull is not defined on nil. I can define my own finalizeResourceData: but the default one, does not work apparently. On Tue, Apr 24, 2018 at 2:06 PM, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: > hi, > > there is nothing like that and I do not recommend messing with the > registry in general. > butâ¦ you can always extend the classes for your use, is around > FFIExternalResourceManager. > > cheers, > Esteban > > > On 24 Apr 2018, at 14:48, Serge Stinckwich <serge.stinckwich(a)gmail.com> > wrote: > > > I'm using autoRelease on some FFIOpaqueObject instances. > > I need to test some behavior when I delete explicitly one of these > objects. âHow to remove these objects from the finalization list, in order > they are not freed two times. > > Something like ignoreFinalization ? > > -- > Serge Stinckwich > UMI UMMISCO 209 (SU/IRD/UY1) > "Programs must be written for people to read, and only incidentally for > machines to execute."http://www.doesnotunderstand.org/ > > > -- Serge Stinckwich UMI UMMISCO 209 (SU/IRD/UY1) "Programs must be written for people to read, and only incidentally for machines to execute."http://www.doesnotunderstand.org/

April 25, 2018

Re: [Pharo-users] SortedCollection>>reverse answers an inconsistent object in Pharo 6

by Denis Kudriashov

2018-04-25 14:26 GMT+02:00 Richard O'Keefe <raoknz(a)gmail.com>: > (aSortFunction value: x value: y) returns a Boolean; > (aSortFunction value: x value: x) returns true. > So you can already set up a SortedCollection using > a SortFunction. So yes, that would work. > Yes. It was original goal to introduce them. I created another thread in dev list "SortedCollection based on SortFunction". > But there is no reason why #reversed (or my preference, > #converse) could not work on sortBlocks. > It will. But I think it is bad approach. Because any time you will need new kind of sort operation you will be forced extend BlockClosure. In general reversed/converse has no meaning for block. With SortFunction only simple extension #asSortFunction is required. > > > On 25 April 2018 at 23:27, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: > >> Hi Richard. >> >> I agree with your proposal. >> But it force me to think that we should completely move to >> SortFunction's. >> In that case SortCollection will have sortFunction variable instead of >> sortBlock. And for your scenario reverse operation will be simple >> expression: "sortFunction := sortFunction reversed". >> >> >> >> 2018-04-23 3:09 GMT+02:00 Richard O'Keefe <raoknz(a)gmail.com>: >> >>> Test case: >>> #(1 2 4) asSortedCollection reverse add: 3; yourself >>> <print it> >>> The answer *should* be aSortedCollection(4 3 2 1) >>> but *is* aSortedCollection(4 2 1 3). >>> This works in Squeak. >>> The problem is that SortedCollection does not define >>> #reverse[d] but simply inherits it(them), and the >>> inherited code pays no attention to the sortBlock. >>> >>> I propose adding the following two methods to >>> SortedCollection: >>> >>> reverseInPlace >>> |a z| >>> a := firstIndex. >>> z := lastIndex. >>> [a < z] whileTrue: [array swap: a with: z. a := a + 1. z := z - 1]. >>> sortBlock := sortBlock >>> ifNil: [[:x :y | y <= x]] >>> ifNotNil: [[:x :y | sortBlock value: y value: x]]. >>> ^self >>> >>> reversed >>> ^self copy reverseInPlace >>> >>> The ANSI method is #reverse, not #reversed, but Pharo >>> defines #reverse to call #reversed, and OrderedCollection overrides >>> #reversed, so SortedCollection *must* override #reversed. >>> >>> #reverseInPlace is the name Squeak uses for the other >>> method. It also has a definition in SequenceableCollection, which is >>> not but is >>> equivalent to >>> >>> reverseInPlace >>> |a z| >>> a := 1. >>> z := self size. >>> [a < z] whileTrue: [self swap: a with: z. a := a + 1. z := z - 1].l >>> ^self >>> >>> r >>> >>> >> >

April 25, 2018

Re: [Pharo-users] SortedCollection>>reverse answers an inconsistent object in Pharo 6

by Richard O'Keefe

(aSortFunction value: x value: y) returns a Boolean; (aSortFunction value: x value: x) returns true. So you can already set up a SortedCollection using a SortFunction. So yes, that would work. But there is no reason why #reversed (or my preference, #converse) could not work on sortBlocks. On 25 April 2018 at 23:27, Denis Kudriashov <dionisiydk(a)gmail.com> wrote: > Hi Richard. > > I agree with your proposal. > But it force me to think that we should completely move to SortFunction's. > In that case SortCollection will have sortFunction variable instead of > sortBlock. And for your scenario reverse operation will be simple > expression: "sortFunction := sortFunction reversed". > > > > 2018-04-23 3:09 GMT+02:00 Richard O'Keefe <raoknz(a)gmail.com>: > >> Test case: >> #(1 2 4) asSortedCollection reverse add: 3; yourself >> <print it> >> The answer *should* be aSortedCollection(4 3 2 1) >> but *is* aSortedCollection(4 2 1 3). >> This works in Squeak. >> The problem is that SortedCollection does not define >> #reverse[d] but simply inherits it(them), and the >> inherited code pays no attention to the sortBlock. >> >> I propose adding the following two methods to >> SortedCollection: >> >> reverseInPlace >> |a z| >> a := firstIndex. >> z := lastIndex. >> [a < z] whileTrue: [array swap: a with: z. a := a + 1. z := z - 1]. >> sortBlock := sortBlock >> ifNil: [[:x :y | y <= x]] >> ifNotNil: [[:x :y | sortBlock value: y value: x]]. >> ^self >> >> reversed >> ^self copy reverseInPlace >> >> The ANSI method is #reverse, not #reversed, but Pharo >> defines #reverse to call #reversed, and OrderedCollection overrides >> #reversed, so SortedCollection *must* override #reversed. >> >> #reverseInPlace is the name Squeak uses for the other >> method. It also has a definition in SequenceableCollection, which is not >> but is >> equivalent to >> >> reverseInPlace >> |a z| >> a := 1. >> z := self size. >> [a < z] whileTrue: [self swap: a with: z. a := a + 1. z := z - 1].l >> ^self >> >> r >> >> >

April 25, 2018

Re: [Pharo-users] SortedCollection>>reverse answers an inconsistent object in Pharo 6

by Richard O'Keefe

As I pointed out yesterday, #sort and its immediate relatives are not alone. Look at #isSortedBy: . When do you want (#(1 1 1) asSortedCollection: b) isSortedBy: b to answer false? There are several reasons for sorting. One is so that you can use binary search. It's rather disconcerting that the binary search offered in Squeak and Pharo has nothing to do with sortBlocks. In several other Smalltalks it does. It is not terribly hard to write a binary search using a sortBlock that acts like #<=. It is not terribly hard to write a binary search using a sort block that acts like #<. It is harder than it needs to be to write a binary search that can tolerate either, and the one Smalltalk I know that does this copes by calling the sortBlock twice as often as it would need to if it knew one way or the other. As it happens, my own Smalltalk library lets me do sortBlock := sortBlock converse, which makes it easy to do the right thing. It is possible to implement this so that the result is as cheap to call as the original.

April 25, 2018

Re: [Pharo-users] SortedCollection>>reverse answers an inconsistent object in Pharo 6

by Denis Kudriashov

Hi Richard. I agree with your proposal. But it force me to think that we should completely move to SortFunction's. In that case SortCollection will have sortFunction variable instead of sortBlock. And for your scenario reverse operation will be simple expression: "sortFunction := sortFunction reversed". 2018-04-23 3:09 GMT+02:00 Richard O'Keefe <raoknz(a)gmail.com>: > Test case: > #(1 2 4) asSortedCollection reverse add: 3; yourself > <print it> > The answer *should* be aSortedCollection(4 3 2 1) > but *is* aSortedCollection(4 2 1 3). > This works in Squeak. > The problem is that SortedCollection does not define > #reverse[d] but simply inherits it(them), and the > inherited code pays no attention to the sortBlock. > > I propose adding the following two methods to > SortedCollection: > > reverseInPlace > |a z| > a := firstIndex. > z := lastIndex. > [a < z] whileTrue: [array swap: a with: z. a := a + 1. z := z - 1]. > sortBlock := sortBlock > ifNil: [[:x :y | y <= x]] > ifNotNil: [[:x :y | sortBlock value: y value: x]]. > ^self > > reversed > ^self copy reverseInPlace > > The ANSI method is #reverse, not #reversed, but Pharo > defines #reverse to call #reversed, and OrderedCollection overrides > #reversed, so SortedCollection *must* override #reversed. > > #reverseInPlace is the name Squeak uses for the other > method. It also has a definition in SequenceableCollection, which is not > but is > equivalent to > > reverseInPlace > |a z| > a := 1. > z := self size. > [a < z] whileTrue: [self swap: a with: z. a := a + 1. z := z - 1].l > ^self > > r > >

April 25, 2018

Re: [Pharo-users] SortedCollection>>reverse answers an inconsistent object in Pharo 6

by Erik Stel

Why does the sortBlock has to answer true for the same (identical) object? I have not been able to find where it is specified. And in different locations I see #> being used as a comparator for the sortBlock. So why is it bad (as you wrote)? (It seems you describe that SortedCollection should create a collection with a Total Order, whereas I describe a situation where a Strict Order is also valid.) -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 25, 2018

Re: [Pharo-users] Launcher not launching

by Jan BlizniÄenko

Hello Thank you for quick reply. Yes, it seems like the same bug and the linked current alpha version works fine. Switched to it and did not encounter even any other problems so far. Thank you. Regards, Jan demarey wrote > Hello, > >> Le 25 avr. 2018 Ã 11:57, Jan BlizniÄenko < > bliznjan@.cvut > > a Ã©crit : >> >> Hello >> >> On Windows 10, I updated launcher from the one on >> https://ci.inria.fr/pharo/view/Launcher/job/Package-and-Publish-Launcher-st… >> (1.0.1) to the one now on official website http://pharo.org/web/download >> (1.1), but the new launcher does not open any images. >> I can create, rename, remove images or do whatever with them, but they do >> no >> launch ... their VM gets downloaded, launcher closes (if it is set to >> close >> on launch - it does not close if not set so), then nothing - the image >> does >> not open, no new Pharo.exe process in list of processes in Windows. >> Tried disabling AV, (re)installing to different directory, selecting >> different directories for images and vms, trying various Pharo versions >> (from 3 to 7) of images, still nothing. It does not work with both old >> and >> new images (downloaded by old working launcher or by this new one). >> There are no exceptions, no crash dumps, nothing in stderr. >> >> Those images can be opened by double clicking on them or by dragging and >> dropping them on C:\Users\User\Documents\Pharo\vms\61-x86\Pharo.exe (or >> similar - the same executable as Launcher should use). Also if I install >> back the old launcher (1.0.1) again, it works. >> >> The new launcher does not work on both of my machines (both Windows 10 >> that >> are quite similar in config and installed applications). >> >> For now, I will stick with 1.0.1, but I would like to find a reason for >> this >> problem. > > I think the problem is > https://github.com/pharo-project/pharo-launcher/issues/68 > Could you try the latest version for windows? > http://files.pharo.org/pharo-launcher/win-alpha/ > Iâm waiting to finish app signing on windows before publishing it. > > Thanks for the report, > Christophe -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 25, 2018

Re: [Pharo-users] Launcher not launching

by Christophe Demarey

Hello, > Le 25 avr. 2018 Ã 11:57, Jan BlizniÄenko <bliznjan(a)fit.cvut.cz> a Ã©crit : > > Hello > > On Windows 10, I updated launcher from the one on > https://ci.inria.fr/pharo/view/Launcher/job/Package-and-Publish-Launcher-st… > (1.0.1) to the one now on official website http://pharo.org/web/download > (1.1), but the new launcher does not open any images. > I can create, rename, remove images or do whatever with them, but they do no > launch ... their VM gets downloaded, launcher closes (if it is set to close > on launch - it does not close if not set so), then nothing - the image does > not open, no new Pharo.exe process in list of processes in Windows. > Tried disabling AV, (re)installing to different directory, selecting > different directories for images and vms, trying various Pharo versions > (from 3 to 7) of images, still nothing. It does not work with both old and > new images (downloaded by old working launcher or by this new one). > There are no exceptions, no crash dumps, nothing in stderr. > > Those images can be opened by double clicking on them or by dragging and > dropping them on C:\Users\User\Documents\Pharo\vms\61-x86\Pharo.exe (or > similar - the same executable as Launcher should use). Also if I install > back the old launcher (1.0.1) again, it works. > > The new launcher does not work on both of my machines (both Windows 10 that > are quite similar in config and installed applications). > > For now, I will stick with 1.0.1, but I would like to find a reason for this > problem. I think the problem is https://github.com/pharo-project/pharo-launcher/issues/68 Could you try the latest version for windows? http://files.pharo.org/pharo-launcher/win-alpha/ Iâm waiting to finish app signing on windows before publishing it. Thanks for the report, Christophe

April 25, 2018

Launcher not launching

by Jan BlizniÄenko

Hello On Windows 10, I updated launcher from the one on https://ci.inria.fr/pharo/view/Launcher/job/Package-and-Publish-Launcher-st… (1.0.1) to the one now on official website http://pharo.org/web/download (1.1), but the new launcher does not open any images. I can create, rename, remove images or do whatever with them, but they do no launch ... their VM gets downloaded, launcher closes (if it is set to close on launch - it does not close if not set so), then nothing - the image does not open, no new Pharo.exe process in list of processes in Windows. Tried disabling AV, (re)installing to different directory, selecting different directories for images and vms, trying various Pharo versions (from 3 to 7) of images, still nothing. It does not work with both old and new images (downloaded by old working launcher or by this new one). There are no exceptions, no crash dumps, nothing in stderr. Those images can be opened by double clicking on them or by dragging and dropping them on C:\Users\User\Documents\Pharo\vms\61-x86\Pharo.exe (or similar - the same executable as Launcher should use). Also if I install back the old launcher (1.0.1) again, it works. The new launcher does not work on both of my machines (both Windows 10 that are quite similar in config and installed applications). For now, I will stick with 1.0.1, but I would like to find a reason for this problem. Thank you Jan -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 25, 2018

[Ann] Iceberg 0.7.3

by Guillermo Polito

Hi all, We just introduced Iceberg 0.7.3 into Pharo. This new version solves several issues people had when contributing to Pharo in general and others. I've made a screencast on how to contribute to Pharo, and added some tutorial and glossary in Iceberg's wiki. Below the links and the changes log. Thanks for making your issue reports ^^. We will schedule a backporting to Pharo6.1 in a couple of weeks, once we have another iteration. Cheers, Guille # Contribute to Pharo with Iceberg 0.7.3 - wiki page: https://github.com/pharo-vcs/iceberg/wiki/Contribute-to-Pharo-with-Iceberg-… - Screencast: https://www.youtube.com/watch?v=PK2yCu2rWCc&feature=youtu.because This screencast shows how to use Iceberg 0.7.3 to contribute to latest Pharo 7, by cloning from scratch your own (potentially outdated) clone. This video will guide you through: - solve the "Missing local repository" by cloning - solve the "Fetch required" that appears because the fork is out of date - solve the "Detached working copy" by creating a new branch - use the pharo plugin to create an issue branch - make some changes, commit them and push them to your fork - make a pull request If you already have a clone, you can just import an existing repository from your disk. The rest of the instructions should be the same. If you don't have a fork, create one before doing anything else. New videos to come with more explanations. # Iceberg Glossary https://github.com/pharo-vcs/iceberg/wiki/Iceberg-glossary # 0.7.3 ChangesLog - #693 <https://github.com/pharo-vcs/iceberg/pull/693>: fixing problem when the announcement of changes in a package is nil. (bug, important) - #682 <https://github.com/pharo-vcs/iceberg/issues/682>: update update script (Pharo6-backport) - #681 <https://github.com/pharo-vcs/iceberg/issues/681>: Author name is not correctly decoded (UI, enhancement) - #680 <https://github.com/pharo-vcs/iceberg/issues/680>: Repair actions dialog should explain the situation (UI, enhancement, important) - #679 <https://github.com/pharo-vcs/iceberg/issues/679>: Create branch repair action should show create branch from issue if available (Pharo plugin, UI) - #678 <https://github.com/pharo-vcs/iceberg/issues/678>: Iceberg does not correctly recognize packages (bug, important) - #677 <https://github.com/pharo-vcs/iceberg/issues/677>: Unknown commit can only be solved with repair action (Model, bug, important) - #676 <https://github.com/pharo-vcs/iceberg/issues/676>: Unknown commit state is not properly shown (UI, bug, important) - #667 <https://github.com/pharo-vcs/iceberg/issues/667>: Merging branches with file deletion doesn't seem possible (bug, important, merge) - #666 <https://github.com/pharo-vcs/iceberg/issues/666>: Iceberg sometimes add a $ to commit message (Model, bug) - #662 <https://github.com/pharo-vcs/iceberg/issues/662>: The Pharo plugin has to do the fetch only if required. (Pharo plugin, enhancement) - #660 <https://github.com/pharo-vcs/iceberg/pull/660>: Add a star badge in the push button in case of a new branch () - #658 <https://github.com/pharo-vcs/iceberg/issues/658>: Progress Bar during checkout of Repository. (UI, enhancement) - #657 <https://github.com/pharo-vcs/iceberg/issues/657>: Confusing UI: Ghost Test maybe should be default? (UI, enhancement) - #656 <https://github.com/pharo-vcs/iceberg/issues/656>: Typo in Progress Bar (UI, bug) - #651 <https://github.com/pharo-vcs/iceberg/issues/651>: When pushing a new branch the list of commit is empty (UI, bug) - #650 <https://github.com/pharo-vcs/iceberg/issues/650>: "Create new branch from issue" is leaving image in detached working copy state () - #649 <https://github.com/pharo-vcs/iceberg/issues/649>: The VM is crashing when trying to use Iceberg from a saved image (bug) - #645 <https://github.com/pharo-vcs/iceberg/issues/645>: IceLibgitRepository should use "utilities" protocol instead of "utils" (clean up) - #644 <https://github.com/pharo-vcs/iceberg/issues/644>: Unnecessary temp in IceTipGitHubRepositoryPanel>>#getGitHubRepository" (clean up) - #643 <https://github.com/pharo-vcs/iceberg/issues/643>: LGit_GIT_ERROR: error authenticating: no auth sock variable loading a Tonel based baseline dependency in smalltalkCI (bug, important) - #639 <https://github.com/pharo-vcs/iceberg/issues/639>: Create pull request fails if branch has not been pushed (Github plugin, bug) - #636 <https://github.com/pharo-vcs/iceberg/issues/636>: The pull button should have a star badge when it has no upstream (UI, enhancement, important) - #634 <https://github.com/pharo-vcs/iceberg/issues/634>: Travis is not building Pull Requests (CI, bug, important) - #624 <https://github.com/pharo-vcs/iceberg/issues/624>: add "add remote" to remote picking on pull/push browsers (enhancement) - #610 <https://github.com/pharo-vcs/iceberg/pull/610>: Fix for issue #608 <https://github.com/pharo-vcs/iceberg/issues/608> and #607 <https://github.com/pharo-vcs/iceberg/issues/607>. (clean up) - #609 <https://github.com/pharo-vcs/iceberg/issues/609>: New branch repair action does not refresh dirty packages (UI, bug) - #583 <https://github.com/pharo-vcs/iceberg/issues/583>: Creating a tonel project does not commit the properties file (Model, bug) - #542 <https://github.com/pharo-vcs/iceberg/issues/542>: Bug during commit of a new package (Model, bug) -- Guille Polito Research Engineer Centre de Recherche en Informatique, Signal et Automatique de Lille CRIStAL - UMR 9189 French National Center for Scientific Research - *http://www.cnrs.fr <http://www.cnrs.fr>* *Web:* *http://guillep.github.io* <http://guillep.github.io> *Phone: *+33 06 52 70 66 13

April 25, 2018