April 2015 - Pharo-users - lists.pharo.org

Non-evaluating and / or

by Jan B.

Hello I would like to ask how to do non-evaluating and. As far as I know, & should be evaluating and and: should be non-evaluating. However, I need to find a way how to use it properly. In "conventional" languages I would expect to work it like in following example... false and: Transcript open Since false is... false... there is no reason to evaluate Transcript open. But when I run this one in playground, transcript opens. Any help is appreciated. Jan -- View this message in context: http://forum.world.st/Non-evaluating-and-or-tp4817211.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

April 3, 2015

Re: [Pharo-users] Voyage performance issue

by Esteban Lorenzano

could be. Write is not a fast action now with the mongo driver. Take into account that each object generates a graph of dictionaries who are committed to the DB. Depending on the complexity of that graph, you will have larger write times. you can try measuring time of two separated operations: [ serialized := VORepository current serializer serialize: theObject ] timeToRun. [ VORepository current resolver insert: theObject serialized: serialized withId: OID nextOID ] timeToRun. that might give you a better idea of what happens (you can also use the TimeProfiler to see where your time is consumed) if you have to write such collections in a regular basis, from there you can explore ways to speed up the process. Sorry not being able to help more, but the amount of information you give me is like going to the doctor and say âit hurts!â :P cheers, Esteban > On 03 Apr 2015, at 01:25, Josefina Arcidiacono <josefine2_94(a)hotmail.com> wrote: > > Hi! > I'm using Voyage with MongoDB. > I need to store an object with a large collection. > Since both are voyageRoot, there is a Mongo table with all the collection > objects and another with the first object pointing to them. > With 20,000 objects in the collection it's taking 30 min to save. > Â¿Is that normal? > > Josefina > > > > -- > View this message in context: http://forum.world.st/Voyage-performance-issue-tp4817087.html > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. >

April 3, 2015

Re: [Pharo-users] OrderedCollection remove:

by Ramon Leon

On 04/01/2015 11:22 PM, Peter UhnÃ¡k wrote: > col := #(1 2 3 4 5) asOrderedCollection. > col do: [ :each | col remove: each. ]. > col As a general rule in Smalltalk, a lot of hassle can be avoided avoiding loops. If you think you need a loop, stop and find a better way, it likely already exists, in this case as someone else already mentioned either #removeAll: or #removeSuchThat:. Smalltalk isn't procedural, loops aren't your bread and butter. -- Ramon Leon

April 3, 2015

Re: [Pharo-users] Devnology presentation slides

by Sven Van Caekenberghe

+10 > On 03 Apr 2015, at 00:36, phil(a)highoctane.be wrote: > > Nice slideware!â > > And interesting points at that. > > Phil

April 3, 2015

Re: [Pharo-users] Glorp with PostgresV2 on Pharo 4

by Sven Van Caekenberghe

> On 03 Apr 2015, at 06:58, Pierce Ng <pierce(a)samadhiweb.com> wrote: > > Now that we know the PostgresV2 driver can talk to our database, using the > Monticello browser, open the PostgresV2 repository and load the package > GlorpDriverPostgreSQL. Here I had to edit > NativePostgresDriver>>connectionArgsFromCurrentLogin: to comment out > the second last line: > > connectionArgs clientEncoding: aLogin encodingStrategy asSymbol > > This is because GlorpDatabaseLoginResource class>defaultPostgreSQLLocalLogin > does not specify encodingStrategy, meaning it is nil and will respond > to #asSymbol with DNU. Yes, we have to fix that, but you also set it manually, no real need to patch the code: createLogin ^ Login new database: PostgreSQLPlatform new; username: 'sven'; password: ''; connectString: 'localhost:5432_sven'; encodingStrategy: #utf8; yourself Interesting articles ! Sven

April 3, 2015

Glorp with PostgresV2 on Pharo 4

by Pierce Ng

I'm on a roll. :-) I've written another blog post on installing Glorp with PostgresV2 on Pharo 4. To avoid this message looking like spam mail promoting my blog and devoid of content, here's the slightly edited blog post, in its original Markdown format sans JSON metadata. Outside of Smalltalk, create the database 'sodbxtest', user 'sodbxtest' with password 'sodbxtest': # su postgres -c psql postgres=# create role sodbxtest with password 'sodbxtest' login; CREATE ROLE postgres=# create database sodbxtest; CREATE DATABASE postgres=# \q # In Smalltalk, firstly, install PostgresV2: Gofer it smalltalkhubUser: 'PharoExtras' project: 'PostgresV2'; package: 'ConfigurationOfPostgresV2'; load. ((Smalltalk at: #ConfigurationOfPostgresV2) project version: '2.4') load. Open Test Runner and runs the PostgresV2 tests. On my Linux Mint machine, using a vanilla PostgreSQL 9.3 installation, 23 of 24 tests passed, and TestPGConnection>>#testNotify2 erred. Now that we know the PostgresV2 driver can talk to our database, using the Monticello browser, open the PostgresV2 repository and load the package GlorpDriverPostgreSQL. Here I had to edit NativePostgresDriver>>connectionArgsFromCurrentLogin: to comment out the second last line: connectionArgs clientEncoding: aLogin encodingStrategy asSymbol This is because GlorpDatabaseLoginResource class>defaultPostgreSQLLocalLogin does not specify encodingStrategy, meaning it is nil and will respond to #asSymbol with DNU. Next, in a playground, execute the following: GlorpDemoTablePopulatorResource invalidateSetup. GlorpDatabaseLoginResource defaultLogin: GlorpDatabaseLoginResource defaultPostgreSQLLocalLogin Open Test Runner and run the Glorp tests. Tested on Linux Mint 17.

April 3, 2015

Re: [Pharo-users] Spring CampSmalltalk

by rjkwok

EstebanLM wrote > you should subject your mail: Spring CampSmalltalk at Nanaimo, Columbia, > Canada For the sake of accuracy, that should be Nanaimo,* British* Columbia, Canada. Sebastion Heidbrink wrote > Please let us also know if youâre interested. Final exams finish on the 25th for me, and I'm just across the water in the lower mainland. I'll see if I can catch the ferry over on the 26th and meet you guys. I'm pretty new to Smalltalk/Pharo, but I'd love to see what you're all making with it! -- View this message in context: http://forum.world.st/Spring-CampSmalltalk-tp4816420p4817118.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

April 3, 2015

Voyage performance issue

by Josefina Arcidiacono

Hi! I'm using Voyage with MongoDB. I need to store an object with a large collection. Since both are voyageRoot, there is a Mongo table with all the collection objects and another with the first object pointing to them. With 20,000 objects in the collection it's taking 30 min to save. Â¿Is that normal? Josefina -- View this message in context: http://forum.world.st/Voyage-performance-issue-tp4817087.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

April 2, 2015

Re: [Pharo-users] Oracle on Linux

by phil＠highoctane.be

This discussion about database connectivity feels like the early days of JDBC. Things got solved when people wrote the type 4 drivers in Java, making it all nice to use and all. Given our resources, I doubt we can get there very fast. But maybe could we leverage the Java work in a way or another. Would it be possible to parse the Java code and have a kind of code rewriter to target Pharo so that we could have those drivers? Phil On Fri, Apr 3, 2015 at 12:19 AM, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: > My position is that OpenDBX as a library, independently of Pharo, is a > dead end. It might have been a good candidate in the past, but today it has > no real use. > > But it is also true that for Pharo it is the only option you have to run > it on Linux. > > Regards, > > ps: I would also avoid Oracle altogether as well, but given the fact it > seems to be a requirement I didn't bring it to the discussion. > El abr 2, 2015 11:04 AM, "Norbert Hartl" <norbert(a)hartl.name> escribiÃ³: > >> >> Am 02.04.2015 um 13:55 schrieb Esteban Lorenzano <estebanlm(a)gmail.com>: >> >> >> On 02 Apr 2015, at 13:13, Norbert Hartl <norbert(a)hartl.name> wrote: >> >> >> Am 02.04.2015 um 12:52 schrieb Esteban A. Maringolo <emaringolo(a)gmail.com >> >: >> >> I don't know if you need full fledged access or just to read a few tables. >> >> If it the latest you can use PostgreSQL Foreign Data Wrappers (FDW) and >> access the pgsql instance using the existing PGv2 driver. >> >> I know it might look convoluted, and maybe it is, but you avoid using DBX. >> >> I would also avoid using oracle which is a prerequisite, right? :) So DBX >> is to avoid? Sorry, I didn't care very much about SQL engines the last >> years. >> >> >> why DBX is to avoid? >> >> >> I don't know. I read the sentence "I know it might look convoluted, and >> maybe it is, but you avoid using DBX." as avoiding DBX is something >> positive. So I want to know as well. >> >> in linux, it is your only chance for connecting to oracle (in windows you >> could also use odbc). >> and it should be working, AFAIK >> >> I'm fine with using DBX as long as it works. I understood also that it is >> one of the higher prioritized tasks for the consortium. >> >> Norbert >> >> >> Esteban >> >> >> Norbert >> >> Regards. >> >> Ps: FDW are amazing. >> El abr 2, 2015 6:07 AM, "Norbert Hartl" <norbert(a)hartl.name> escribiÃ³: >> >>> I see from time to time topics like SQL popping up. Most of the time >>> DBXTalk seems to be a requirement. It is also something that is considered >>> necessary for a lot of people. Now have the need to access an oracle >>> database. Is that possible? And if yes does it also work on linux? >>> >>> Norbert >>> >> >> >> >>

April 2, 2015

Re: [Pharo-users] Devnology presentation slides

by phil＠highoctane.be

Nice slideware!â And interesting points at that. Phil

April 2, 2015