Adriaan van Os Adrian Kuhn Adrian Lienhard Alan Capewell Alan Knight Alan Lovejoy Andreas Raab Andres Valloud Arden Thomas Bee SmaRT Benoit St-Jean Bernhard Pieber Bert Freudenberg Bill Kerr Bob Nemec Boris Popov Bruce Badger Canol Gökel Cees de Groot Chris Cunnington Chris Thorgrimsson Christoph Budzinski Cincom Smalltalk Clément Béra Coen De Roover Colin Putney Craig Latta Daan van Berkel Dale Henrichs Darius Clarke David A. Smith David Buck (Simberon) Davide Varvello Diego G. Deck Dmitri Zagidulin ESUG news Eddie Eliot Miranda Esteban A. Maringolo Esteban Lorenzano Felipe Bañados-Schwerter Fernando Olivero Francois Stephany GNU Smalltalk Blogs German Arduino Giovanni Corriga Guillermo Polito Gwenael Casaccio Göran Krampe Hilaire Fernandes Holger Hans Peter Freyther Holger Hans Peter Freyther Independent Misinterpretations Isaiah Perumalla JR's Smalltalk 4 You James Christian Ladd James Foster James Robertson Jan van de Sandt Jason Javier Pimás Jeffrey Massung Joachim Tuchel Johan Brichau John P Jordi Delgado Joshua Gargus José A. Ortega-Ruiz Julian Fitzell Julian Lombardi Ken Treis Laurent Laffont Learning Smalltalk Luca Bruno Lukas Renggli Mariano Martinez Peck Marilyn Lombardi Marten Feldtmann Mathieu Suen Matthew Fulmer Michael Davies Michael Lucas-Smith Michael van der Gulik Michal Wyrobeck Milan Zimmermann Nicholas Chen Nicolas Cellier Nicolas Petton Niko Schwarz Norbert Hartl Noury Bouraqadi Panty Jack Pharo Casts Pharo News Pharo Weekly Philippe Mougin Pierce Ng PyPySqueak Blogs Ramon Leon Redline Smalltalk Ricardo Moran Rob Vens Runar Jordahl Sean Allen Sean DeNigris Sean McGinty Self Blogs Silverlight Smalltalk Smalltalk Smalltalk Jobs Smalltalk Medicine Show Squeak Oversight Board SqueakNOS Stefan Krecher Stefan Marr Stephan Wessels Summer of Squeak Suslov Nikolay Suslov Nikolay Tao Starbow The Weekly Squeak Tim Mackinnon Tim Mackinnon Tom Koschate Torsten Bergmann Travis Griggs UK Smalltalk Vassili Bykov Wilkes Joiner Yoshiki Ohshima Yuri Vrancken
+10.42.52.png)
Hi, Spoon 3 beta 5 is released. I’ve still got a bunch of changes pending, for a 3b6 release to follow shortly. This release is just to fix some startup problems on Windows, Linux, and Mac OS. You can also find the Spoon VM changes separated out, in the second “Resources” folder.
What I’d like is for you to just start the app and tell me the results, along with your host platform. Thanks!
I’ve noticed, that the way via Topaz is pretty time consuming: export generated source code, transfer to remote server and then local import via Topaz.
But actually there is a much better way. Copy the generated source code into the clipboard, switch to Jade (which is connected to the remote server), open an empty Workspace, paste the stuff into that Workspace, select all and then do a “FileIn” – and I was not aware, that the expected source code format in this window is: Topaz. Fine
In the 01.02.08 I added a little “Copy all to Clipboard” button to make the whole procedure a little bit faster.
Another stuff I added in this version:
* in some generated source code of methods I used temp variable names which are not that unusual. Actually I got error while compiling my project. Therefore I switch to temp method variable names beginning with an “_”.
* I now created additional methods on the class side of the project class about the model version, the model creation time, the model last saving time, the generator software version.
* most of domain classes in my project have an instance variable named “id”. In REST URL paths there might be a need (for swagger specs) to add the name of the parameter into a path like “/gwc/system/{id}”. When different “id”s of different classes are used the supported framework will not work. Therefore I added method returning extended names to be used in this case like: “/gwc/system/{SystemId}/user/{UserID}”
Marten
The Kapital group in Glasgow seek an infrastructure developer who can also supervise associated UNIX infrastructure changes. Experience in Smalltalk and GemStone is preferred, but someone willing to learn these who is experienced in other OO programming languages, and on other databases, would be considered.
Hernán Morales Durand announced a new fun library with which we can query wikipedia.
Some examples follow:
1) Retrieve in JSON movies from the beautiful Julianne Moore:
| jsonResults |
jsonResults := DBPediaSearch new
setJsonFormat;
timeout: 5000;
query: 'SELECT DISTINCT ?filmName WHERE {
?film foaf:name ?filmName .
?film dbpedia-owl:starring ?actress .
?actress foaf:name ?name.
FILTER(contains(?name, "Julianne"))
FILTER(contains(?name, "Moore"))
}';
execute
To actually get only the titles using NeoJSON:
((((NeoJSONReader fromString: jsonResults) at: #results) at: #bindings)
collect: [ : entry | entry at: #filmName ]) collect: [ : movie | movie
at: #value ]
2) Retrieve in XML which genre plays those crazy Dream Theater guys :
DBPediaSearch new
setXmlFormat;
setDebugOn;
timeout: 5000;
query: 'SELECT DISTINCT ?genreLabel
WHERE {
?resource dbpprop:genre ?genre.
?resource rdfs:label "Dream Theater"@en.
?genre rdfs:label ?genreLabel
FILTER (lang(?genreLabel)="en")
}
LIMIT 100';
execute
More examples are available in DBPediaSearch class side. You can install it
from the Configuration Browser.
If you want to contribute, just ask me and you will be added as contributor.
Best regards,
Hernán
Updates on 3/6/14: Added a listing for the position in Singapore.
Good luck with your job hunting,
James T. Savidge
This blog’s RSS Feed
PUModeller has the whole information of the models. Association are created via OrderedCollection but with Gemstone support I added the possibility to have unique-keyed dictionary realization of associations – nice to have, if you want to look-up elements via unique (not changeable) keys.
But now to the realized swagger supprt. PUM has to transfer parts of its model into the JSON oriented specification of swagger.
Full automation is not possible, but PUM creates help methods to define swagger specs for partial classes (without associations) and full classes (with associations). Associations are less needed, because it would lead to export of large structures. Perhaps not what you want to have if you create answers of REST calls. In addition it creates list models of partial classes – nice to use if you scroll (via rest calls) through a list of elements.
Swagger means also JSON: PUM creates methods defining special mapping for the JSON export of data and offers some special conversions support:
DateTimeWithStamps are converted automatically to UTC and Javascript number. This seems to be practicable for JS clients.
Enumerations are converted from their numerical internal value to their textual representation.
JSON support is based on Neo-JSON.
I like SQLite. SQLite is small, fast and reliable. As mentioned in this old blog post, I was extending the Squeaksource SQLite FFI wrapper to support parameter binding. Since then, Torsten Bergmann has also picked it up and now tends to it on SmalltalkHub.
Taking the chance to learn NativeBoost, with hints from Masashi Umezawa’s PunQLite, I took my SQLite FFI changes and created NBSQLite3. NBSQLite3’s small test suite runs successfully on Pharo 2.0 (OSX, Linux) and the current Pharo 3.0 beta on OSX.
Here’s a multithreading example:
| db db2 sem blocking blocked |
db := NBSQLite3BaseConnection on: 'file::memory:?cache=shared'.
db2 := NBSQLite3BaseConnection on: 'file::memory:?cache=shared'.
sem := Semaphore new.
Transcript open; clear.
blocking := [ :x :sema |
Transcript show: 'Thread-1: begin exclusive...'; cr; flush.
[ x open.
x execute: 'begin exclusive'.
sema wait.
Transcript show: 'Thread-1: releasing exclusive lock.'; cr; flush.
x execute: 'rollback'
] ensure: [ x close ]
] newProcessWith: (Array with: db with: sem).
blocking resume.
Processor yield.
blocked := [ :x |
Transcript show: 'Thread-2: create table ...'; cr; flush.
[[ x open. x execute: 'create table if not exists x (xk integer)' ]
on: NBSQLite3Locked do: [ :ex | Transcript show: ex messageText; cr; flush ]
] ensure: [ x close ]
] newProcessWith: (Array with: db2).
blocked resume.
Processor yield.
sem signal.
To run this example, evaluate “NBSQLite3Example multithreading”. The output looks thusly:
Thread-1: begin exclusive...
Thread-2: create table ...
database schema is locked: main
Thread-1: releasing exclusive lock.
The SQLite shared library is compiled with -DSQLITE_USE_URI=1 and -DSQLITE_THREADSAFE=2.
While working on a small REST system I noticed how boring (and incredable time consuming) all that work is. Typing in the classes, implementing the REST-API calls and then document the whole stuff for all the other developers.
After using the modeling tool now the changes of the model are pretty instant and more or less interactive – but now the documentation is still very time consuming.
Sebastian Heidbrink pointed me to an external tool set named swagger, which could help me saving time.
Swagger works the way, that you annotate your source code with information and then from this information the system generates a JSON based schema model, which defines the parameters location, parameter types and response type for each REST call (and some more stuff).
Several REST-calls may be bundled into defined REST-APIs.
The JSON resources can be static files located on any server. Now with a tool named SwaggerUI one can visualize the API and test it. The last part is really fun debugging !
I actually changed the kind of working with swagger for me. I added several Smalltalk classes in Gemstone to be able to define the JSON based schema model and now added Smalltalk code on the REST class class-side with information, what calls are belonging to the API, what parameters are needed etc.
Then I added a small REST runtime system and for each request now the system is looking for the JSON schema nformation about the executed call, look for the parameter descriptions, retrieves the parameter values, checks them and if all is ok, call the domain specific method.
ZincHTTP now not only serves the API, but also the specification of the API.
After having done this – one thinks about how the Modelling tool can create support for the Swagger Schema – but I will talk about this in a later posting
When receiving a large data HTTP request with a size > 10.000.000 bytes the Zinc HTTP logic might fail due a memory limitation in a primitive. while reading a socket.
To get rid of this problem the receiving logic in Zinc should be changed to attempt several reads to get the request.
When working with Gemstone/s as a REST server system I noticed, that the actual Zinc-HTTP server is crashing now and then.
The reason is an unexpected closing socket throwing a SpSocketError and this exception is NOT catched.
I therefore added this error to:
ZnMultiThreadedServer>>readRequestExceptionSet
and
ZnMultiThreadedServer>>writeResponseExceptionSet
Marten
