September 2017 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Calypso question

by Denis Kudriashov

2017-09-24 9:30 GMT+02:00 Stephane Ducasse <stepharo.self(a)gmail.com>: > What is the relationship between a cursor and ClyEnvironmentContent > and ClyEnvironment? > If you are asking about ClyNavigationEnvironment then it is kind of facade to query the system. It is responsible to select specific scope of environment. And it maintains cache for them. scope := ClySystemNavigationEnvironment selectScope: ClyPackageScope of: {Object package} Concrete kind of environment represents concrete kind of system. ClySystemNavigationEnvironment represents general Pharo system. With scope you evaluate queries: cursor := scope query: (ClyMessageSenders of: #(do:)_. And cursor provides enumeration interface to items of query result where result itself is an instance of concrete type of ClyEnvironmentContent. So cache of ClyNavigationEnvironment has multiple scopes. Every scope instance has cache of multiple ClyEnvironmentContent instances. Each content knows in what scope it was built and what environment it represents (and also what query built it). Cursor instance references single environment content instance. You can have multiple cursors which point to same environment content instance. Is it clear? > On Sun, Sep 24, 2017 at 9:16 AM, Stephane Ducasse > <stepharo.self(a)gmail.com> wrote: > > On Sat, Sep 23, 2017 at 10:56 AM, Denis Kudriashov <dionisiydk(a)gmail.com> > wrote: > >> > >>>> > >>>> another question > >>>> > >>>> packageScope query: (ClyMessageSenders of: #(do: x)) > >>>> > >>>> It is returning all the senders of do: and the senders of x > >>>> or > >>>> returning all the sender of do: and x > >>> > >>> > >>> It returns methods which sent #do: or #x > >> > >> > >> This logic was needed to support "Senders button" when multiple methods > are > >> selected. I think it is expected behaviour in that case. > > > > Yes but this is not when you read the code > > > > packageScope query: (ClyMessageSenders of: #(do: x)) > > > > > > packageScope query: (ClyMessageSenders anyOf: #(do: x)) > > > > is much clearer because it means what it does. > > > > > > stef > > > > One day you should try to write documentation on something that you > > should reverse engineer first and you will see. > > Nothing is clear. > > > > > >> > >>> > >>> > >>>> > >>>> > >>>> > >>>> Stef > >>>> > >>> > >> > >

Sept. 25, 2017

Re: [Pharo-dev] Calypso question

by Denis Kudriashov

2017-09-24 9:15 GMT+02:00 Stephane Ducasse <stepharo.self(a)gmail.com>: > anyOf: means > senders of any of the following message = or > > because of: with a list could mean that you expect to find the senders > of all the message in the list. > > Am I clear? > Yes. But is it good english to say "Senders anyOf: #(do: x)"? Should not it be "Senders ofAny: #(do x)"? > > Stef > > > On Sat, Sep 23, 2017 at 10:37 AM, Denis Kudriashov <dionisiydk(a)gmail.com> > wrote: > > 2017-09-23 7:54 GMT+02:00 Stephane Ducasse <stepharo.self(a)gmail.com>: > >> > >> Ok what is localClassScope? > > > > > > From LocalClassScope you will see only things which are defined in class > > itself. > > > > There are also ClassHierarchyScope subclasses. They show you all things > > which are visible from related classes depending on what type of > hierarchy > > they represent (there are ClySubclassScope, ClyInheritedTraitScope, etc). > > > > I will add this notes to class comments. > > > > Before I introduced ClassHierarchyScope there was only ClassScope (with > > instance side and class side subclasses). Then I extract LocalClassScope. > > > >> > >> > >> so may be we should rename it anyOf: > > > > > > I not understand it. > > > >> > >> > >> On Fri, Sep 22, 2017 at 9:54 PM, Denis Kudriashov <dionisiydk(a)gmail.com > > > >> wrote: > >> > Hi > >> > > >> > 2017-09-22 20:41 GMT+02:00 Stephane Ducasse <stepharo.self(a)gmail.com > >: > >> >> > >> >> Hi denis > >> >> > >> >> I'm doing a pass on the architecture of calypso chapter :) as a way > to > >> >> force me to understand it better. > >> >> > >> >> Now I do not understand why the following does not work. > >> >> > >> >> | env classScope | > >> >> env := ClySystemNavigationEnvironment currentImage. > >> >> classScope := env selectScope: ClyClassScope of: {Point. Collection}. > >> >> classScope allClasses > >> >> > >> > > >> > Probably it is not implemented. Each scope should implement what can > be > >> > retrieved from it. > >> > If something was not needed for browser implementation then it will > not > >> > work. (I not implemented it). > >> > > >> > But try also with ClyLocalClassScope. I did little refactoring and it > >> > can be > >> > not reflected in any doc/comment. > >> > > >> >> > >> >> > >> >> another question > >> >> > >> >> packageScope query: (ClyMessageSenders of: #(do: x)) > >> >> > >> >> It is returning all the senders of do: and the senders of x > >> >> or > >> >> returning all the sender of do: and x > >> > > >> > > >> > It returns methods which sent #do: or #x > >> > > >> >> > >> >> > >> >> > >> >> Stef > >> >> > >> > > >> > > > >

Sept. 25, 2017

Esteban's ChangeLog week of 18 September 2017

by estebanlm＠gmail.com

Hello! This is my weekly ChangeLog, from 18 September 2017 to 24 September 2017. You can see it in a better format by going here: http://log.smallworks.eu/web/search?from=18/9/2017&to=24/9/2017 ChangeLog ========= 22 September 2017: ------------------ * Yesterday I worked on a new version of [metacello](https://github.com/Metacello/metacello/tree/pharo-6.1_dev) for Pharo 7.0 that incorporates latest fixes and specially, adds support for [tonel](https://github.com/pharo-vcs/tonel). After some tries, it is ready to be loaded and older Pharo versions can update executing: ---- Metacello new baseline: 'Metacello'; repository: 'github://metacello/metacello:pharo-6.1_dev/repository'; onConflict: [:ex | ex allow]; get; load ---- (do not fear, even if is a 6.1 branch, it will load properly also on Pharo 3.0+) Now, I discovered a remaining problem, [https://github.com/Metacello/metacello/issues/460](https://github.com/Metac… (I will fix this today) but most of the time this will be painless. 20 September 2017: ------------------ * ... and today we make linux i386 and x86 [headless VM](https://github.com/estebanlm/opensmalltalk-vm/tree/add-minheadless-vm) to build and (aparently), work :) 19 September 2017: ------------------ * Today Ronie and I finished the [build of real headless VMs](https://github.com/estebanlm/opensmalltalk-vm/tree/add-minheadless-vm) for macOS (both i386 and x86) versions. Now, we stopped here because linux is more complicated (oh yes, at least once :P): since all configuration is made with autoconf and Ronie's' builds are made with CMake, now we need to adapt more than we wanted. Anyway, tomorrow we'll do another pass... or maybe we will start with Windows, to see if this time also there is easier ;) cheers! Esteban

Sept. 25, 2017

Re: [Pharo-dev] Non-deterministic ombu errors

by Martin Dias

On Sun, Sep 24, 2017 at 5:49 AM, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > > On 23 Sep 2017, at 22:42, Jan BlizniÄenko <bliznjan(a)fit.cvut.cz> wrote: > > > > Hello > > > > I am having very similar errors from time to time when manually loading > code > > into Pharo from filetree (gitfiletree), but only in one of my projects > which > > contains non-ASCII characters like ÄÅ¡ÄÅÅ¾Ã½Ã¡ÃÃ©Å¯ÃºÅ¥Ä. For the very same > reason > > (not able to find why sometimes the error is there and sometimes is > not), I > > did not report it yet, but it is quite annoying. Retrying loading the > > package helps every time. > > I think was reported before. I believe Ombu somewhere takes an arbitrary > chunk of bytes out of an UTF-8 encoded file (like the last 100 of something > like that) and then starts reading that. With UTF-8, a variable length > encoding, that can be dangerous because you could end up in the middle of a > character, not between encoded character boundaries. This would also > explain the randomness, as it is highly dependent on the contents. > > ZnUTF8Encoder>>#backOnStream: can be used to make sure you are really at > the start of an encoded character. > Hello This issue was reported as: https://pharo.fogbugz.com/f/cases/20112/Epicea-can-fail-while-reading-non-a… My apologies for the bug... I introduced that code as an i/o optimization but wasn't properly tested. Thnks Sven, but I finally used stream's #basicNext (as Henrik suggested in fogbugz) which skips forward. Martin

Sept. 24, 2017

Re: [Pharo-dev] XPath chapter...

by ghoetker

Thank you, Stef, for providing this (and all else you do). As a very new user, I agree completely with your comment about the importance of documentation. I addition to the "Command/Package" based approach exemplified in this and the other books, I'd encourage the community to also consider a "Recipes/Cookbook" approach. I've linked to two examples below, but the basic idea is an organized structure of Problem: (E.g., I want to list all of the PDFs in a directory and its sub-directories). Solution: [Sample code, perhaps heavily commented to explain. May have a basic and a more advanced code snippet] Explanation: [An explanation of the sample code, if not provided in its comments] This approach is especially helpful when users may not be familiar with the ecosystem, meaning they wouldn't even know where to start looking for an answer. Also, some people learn best starting from specific problems. I feel bad recommending that "Someone should do this..." when I'm not in a position to help currently (heaviest teaching period of the year), but offer this up for consideration. https://books.google.com/books?id=wVsnCgAAQBAJ&pg=PA192&lpg=PA192&dq=swift+… <https://books.google.com/books?id=wVsnCgAAQBAJ&pg=PA192&lpg=PA192&dq=swift+…> https://iswift.org/cookbook/get-directory-contents <https://iswift.org/cookbook/get-directory-contents> ----- Glenn Hoetker ghoetker(a)me.com http://hoetker.faculty.asu.edu -- Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html

Sept. 24, 2017

XPath chapter...

by Stephane Ducasse

Hi I'm fully convinced that we can have the best language if we do not let people find their way super easily to do what they want to do we will fail. Documentation is KEY and I feel a bit alone. Since I always wanted to learn XPath I started to write a chapter on XPath. I hope that I did not write too many mistakes. Pull requests are welcome here https://github.com/SquareBracketAssociates/Booklet-Scraping I would love to have one booklet on SAX and DOM handling. Stef

Sept. 24, 2017

STON question :)

by Stephane Ducasse

Hi Sven For Moose we are wondering if we can use STON to save/load models instead of MSE. - I'm sure we can :) Now my little friends would like to kill two or three birds with the same stone :) Would it be possible to use STON to handle cycles but to generate JSON compatible in the following sense: Class { "id": 123, "name" : "A" } => { "id": 123, "type" : "Class", "name" : "A" } It means that we could be able to output an extra field and when we read that the materialiser would create an entity based on the type field for example. Am I thinking something totally wrong? Stef

Sept. 24, 2017

Re: [Pharo-dev] catalog seems down?

by Esteban Lorenzano

> On 24 Sep 2017, at 10:48, Norbert Hartl <norbert(a)hartl.name> wrote: > > Esteban, > > this sounds as the catalog will be generated on each request. This is architectural asking for problems. > The catalog is updated once per day. At that time the response content of the catalog should be written to a file. The file can be served from a frontend web server. This way it is very reliable and can scale easily to 100 requests/s or more. all that is true. but catalog was a temporal solution that became not so temporal. Iâm still waiting to kill itâ¦ and I do not want to spend much time on it :) Esteban > > Norbert > >> Am 24.09.2017 um 10:07 schrieb Esteban Lorenzano <estebanlm(a)gmail.com>: >> >> is online now. >> but see that *it could* have been offline for a short period of timeâ¦ I added an auto-restart job that checks catalog status each hour, so it may happen that it was down during some time. >> >> Esteban >> >>> On 24 Sep 2017, at 09:12, Stephane Ducasse <stepharo.self(a)gmail.com> wrote: >>> >>> >>> >

Sept. 24, 2017

Re: [Pharo-dev] Non-deterministic ombu errors

by Sven Van Caekenberghe

> On 23 Sep 2017, at 22:42, Jan BlizniÄenko <bliznjan(a)fit.cvut.cz> wrote: > > Hello > > I am having very similar errors from time to time when manually loading code > into Pharo from filetree (gitfiletree), but only in one of my projects which > contains non-ASCII characters like ÄÅ¡ÄÅÅ¾Ã½Ã¡ÃÃ©Å¯ÃºÅ¥Ä. For the very same reason > (not able to find why sometimes the error is there and sometimes is not), I > did not report it yet, but it is quite annoying. Retrying loading the > package helps every time. I think was reported before. I believe Ombu somewhere takes an arbitrary chunk of bytes out of an UTF-8 encoded file (like the last 100 of something like that) and then starts reading that. With UTF-8, a variable length encoding, that can be dangerous because you could end up in the middle of a character, not between encoded character boundaries. This would also explain the randomness, as it is highly dependent on the contents. ZnUTF8Encoder>>#backOnStream: can be used to make sure you are really at the start of an encoded character. > Jan > > > gcotelli wrote >> Hi, >> >> I'm having some non-deterministic errors in the Travis builds, the error >> is >> always something like: >> >> *OmFileStoreReadingError: Reading entry from File @ >> pharo-local/ombu-sessions/Pharo.4p8buz6mww48y1noxrolgwkl1.ombu@17432579: >> Invalid utf8 input detected* >> >> If I retried the build it works without any change. I think it happens >> during the code loading phase. >> >> Is anyone experiencing this issue? >> >> Here's an extract of the travis log: > > > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html

Sept. 24, 2017

Re: [Pharo-dev] catalog seems down?

by Norbert Hartl

Esteban, this sounds as the catalog will be generated on each request. This is architectural asking for problems. The catalog is updated once per day. At that time the response content of the catalog should be written to a file. The file can be served from a frontend web server. This way it is very reliable and can scale easily to 100 requests/s or more. Norbert > Am 24.09.2017 um 10:07 schrieb Esteban Lorenzano <estebanlm(a)gmail.com>: > > is online now. > but see that *it could* have been offline for a short period of timeâ¦ I added an auto-restart job that checks catalog status each hour, so it may happen that it was down during some time. > > Esteban > >> On 24 Sep 2017, at 09:12, Stephane Ducasse <stepharo.self(a)gmail.com> wrote: >> >> >>

Sept. 24, 2017