I would be strongly interested in using Coral as a replacement for
puppet/chef/salt
Stephan
The build was red, referring to dh83/ast-Interpreter (empty).
Changed the config to use Pharo/AST-Interpreter
Stephan
Camillo,
While fixing some ci builds, I noticed that at least some of your smalltalkhub
repositories are empty (petitmarkdown, ast-interpreter). What happened
with the code? ast-interpreter I could find back, but petitmarkdown not.
Stephan
How is git support for Pharo going to work?
Because currently it takes extra effort.
I noticed some Pharo users moving their code to
github, making it less accessible.
I know a few different ways to get git installed on a mac:
(finkproject, brew, macports, xcode, git-osx-installer, git-scm.com)
Max wrote:
I’m working on libgit2 bindings for Pharo. If we can include libgit2 in the VM
environment there will be no dependencies to a Git installation. I’ll be
visiting Lille for two weeks starting on monday and one goal is to advance
that work to a stage where we can start doing something
Max wrote:
I don’t quite follow. What do you mean by methods?
When mapping methods to files, there is the problem that on disk files take a
minimum amount of space,
often 4KB. This slows writing a package to disk down a lot.
Stephan
Max wrote
In my opinion there are other factors, like readability, maintenance etc, that
are maybe more important than speed.
I’m not totally convinced. The readability from the outside is limited by
having one file per method.
Mainstream tooling assumes one file per class. The emails from
Eliot wrote
But the real things to learn are the Method Finder and the browser.
In case you are wondering why no beginner seems to know about Method Finder:
Your post made me take a look at it for the first time (after near a decade of
smalltalk). Interesting class comment. I’m pretty sure no
At the moment I'm moving, so I have very little time to do something about it.
Current state is described in a thread from end of last year:
http://forum.world.st/have-a-question-for-morphic-Designer-in-pharo-td4731307.html
Stephan
On the magritte and seaside lists Bruce Prior explains that loading
stable versions of Seaside and Magritte doesn't result in
having a working combination.
project: 'Magritte3' with: [
spec
Max wrote:
Good idea. What would we put there? I can’t think of anything else than
globals…
If globals is the only thing then I’m not sure that it’s worth the effort.
Make sure the code loads (not runs) in gemstone. Then you know.
At least the calls to primitives should be moved.
Stephan
?
Stephan Eggermont
ConfigurationOfForum done.
Stephan
Sven wrote:
That is great, Stephan!
Would you say that all the important libraries that you depend on in Pharo 2.0
are already available for Pharo 3.0 or not ? Which ones are missing from your
standpoint ?
Pier3 with add-ons.
I want to be able to deploy a Pier website like you showed in
After typing cmd-x, we get a kind of modal behavior: cmd-m does something else
than before we typed cmd-x. This should be visible in the UI. In a class
browser we could highlight which method, class, package would be removed when
the relevant keys were to be pressed, or the title/status bar of
Torsten wrote
Please make it a setting - so one can see the creation message with
poolDictionaries in Nautilus
as before in the browser and is able to fill out the template.
Please don’t. Settings as a replacement for design are evil
Stephan
Anyone?
Stephan
I’ll be there from friday evening, and plan to have diner there.
Last years we had diner at a nice place near the south station,
which Andy found. I don’t remember the name, but know how
to get there. If I remember correctly, parking there was
less easy. I would be interested in having diner
Bestlem wrote
I think this is a misreading of Apple's directory design. It is for more
than preferences and caches but for all app-specific files
No, it is not. Library is meant for shared libraries, like vm plugins,
and preferences and caches. Nothing more.
Apple is changing this with the app
I’ve taken a better look at what users is doing.
would classesUsing be a more specific name for that?
Stephan
I tried loading Deltawerken in Pharo 3.0 and noticed
it is now impossible to load code where a class side method is defined named
users.
DEUser classusers
^self subclassResponsibility
Stephan
Igor,
The way you describe the role of punishment in education is not in line with
current research. Most learning happens trough copying the behavior of others,
and punishment has a number of negative consequences on character development,
making it a non-suitable instrument. You might want
I’ve asked the author of Gravel to show it to us at Fosdem2014.
He wants to, but might not be able to make it due to expected
family expansion.
Stephan
Sebastian wrote
Cheap idea for us:
Good idea for a bit later, and for the new parts.
We already have a number of well-known problems for new users.
If we don’t eliminate them first, we’re not going to find much new
issues. They will be hidden behind:
- no discoverable way of creating classes
If anyone wants the solution for
https://pharo.fogbugz.com/f/cases/10551/All-Monticello-Repositories-are-added-to-all-Working-Copies
before Jan 26, please remember Camillo is working on his PhD till then. (needs
reviewing)
Stephan
Doru wrote:
I do not get that. As far as I can tell, the fix is already integrated :).
No, it needs to be resolved to Fix to be included.
Stephan
Doru wrote:
Hmm, strange. It was available in 30666, but it is not available anymore in
30669.
Was it reverted because it was not good, or was it because of another issue?
It was never there. Did you by change load the slice in your 30666?
Stephan
Stef wrote:
my doc stopped me because I’m exhausted so see you in two weeks.
Please take care of yourself.
Stephan
Hilaire wrote:
I asked for the changes implication and see a lot not
related to PolygonMorph. Is it normal?
The first comment is
mainly cleaning deadcode.
That should serve as a warning.
Most slices have a better focus, but there are a few
only describing the most important fix. Finer
The point I apparently failed to make is that I miss the abstraction above the
widget level.
Stephan
Jannik wrote
- In a browser, when I did a change in the source code of a method,
then I select another method, the changes are lost, without asking if I want
to save changes.
Thanks Nicolai, for the fast results. I retested.
Resolved, Fix to Include
Stephan
Marcus wrote:
On 31 Dec 2013, at 11:58, Tudor Girba [hidden email] wrote:
I think this should not be a global preference at all. Most of the time, you
actually want to commit in the same repository where you always do. Only
from time to time do you need to switch and copy to another repo.
Hi Doru,
I like very much how easy it is to modify and adapt these tools.
The higher abstraction level provided by Glamour really pays off here.
In the past week I also looked at Nautilus, MCWorkingCopyBrowser and
the UiDesigner and the difference is shocking.
Stephan
Sean wrote:
I would've fixed it already, but since someone went through the trouble to add
a setting, I figured maybe there was a use case I was missing...
Please go ahead, I won't have any more time to do it today or tomorrow. Design,
no choice!
Stephan
Norbert wrote:
To send or receive a broadcast packet you need to put special options on the
socket on the native side (setsockopt). You need to look if the SO_BROADCAST
option is mentioned anywhere. If not it won’t work without adding them. And
for sending them you need almost special
Stef wrote:
Doru wrote:
We need to reconsider how we approach continuous integration for
pharo-based images.
can you expand ?
The most important thing, from my POV, is to not break things silently when
cleaning up
the system. Having a system integration test that tells you you are no
Doru wrote:
This is clearly an issue of Grease. And it is not essential either as there
are no senders to it.
There are senders in Seaside. They look not so difficult to change.
So, I see no reason not to rename it in Grease.
It is a compatibility layer, so you wouldn't be able to know
I notice that several builds are green, though I added a breaking test to Pharo
3.
That probably means that those builds don't run all of the Pharo tests.
Most builds probably don't need to run all tests (all the time), but it might be
useful to be able to mark tests as integration tests, that
Torsten,
For unit tests building on green builds might be enough, but what about
integration/smoke tests?
Concrete case: I added a test to check that no class implements packages
That is a test that should be run by all builds, but is not. The consequence is
that Moose and Pharo3 based builds
Downloaded 20628 288 VM, run on a Mavericks mac
Browse the slice:
Segmentation fault Mon Dec 9 10:12:26 2013
https://github.com/pharo-project/pharo-vm.git Commit:
71adca113a9aff2876288927e5c1c86bf7ac13e2 Date: 2013-12-07 05:00:47 -0800 By:
Philippe Back philippeb...@gmail.com Jenkins build
Hmm, not easily reproducible.
Stephan
I found a workaround. See
https://pharo.fogbugz.com/f/cases/12379/Shape-is-missing-a-comment
For me, the Nautilus create a slice and commit” doesn’t work. Creating the
slice from the
monticello browser and saving it there does work.
And a new version of the 12324 slice is committed
https://ci.inria.fr/pharo/job/Pharo-3.0-Issue-Validator/4185//artifact/validationReport.html
Stef wrote
Now for example you can load code in a background process now if
you modify objects that may be impacted by changes then you can get in
trouble.
Here is a live example of doing that: This is how we load Delphi code in Moose.
http://www.flickr.com/photos/35787002@N03/10609428345/
In
Hi Torsten,
Just take a look at:
http://smalltalkhub.com/mc/StephanEggermont/DeprecationFinder/main
It is deliberately rate-limited.
It doesn’t find the teams and their projects. The API you can find by
downloading the SmalltalkHub image from ci. That wants a MongoDB
running with default
Just tried in 30624.
Looks good to me. Works a lot faster.
Is not trivial to load by hand, it needs search strings entered (nil isn’t
empty)
The way the search is forked doesn’t interfere with the UI thread?
Stephan
Would it make sense to check the search string for complex strings (*\?) and
switch to
regex searching then? That should be a simple extension (later).
Stephan
Ben wrote:
who put a ô in the code at the first place ? :P
Doesn’t happen often, I’m happy to observe. Strings in code
with interesting characters are a much more common problem,
though. Made it impossible to import MCs into Gemstone.
Stephan
I managed to break it.
Open monticello, open the 30inbox
type sl return
backspace backspace return
Still needs a nil check it seems
Stephan
Ben wrote:
It used to be regex, but regex are slow (and far slower than just string
comparison)
Yes, I saw that. They are also useful. Checking if the search string contains
regex
characters is not slow. Only doing regex checking when they are there solves the
main slowness, without losing
### Here's what I've been up to since the last WhatsUp:
- put PetitDelphi in the moose ci
- complained about losing code (issue 12324)
- added the twitter bootstrap magritte demo to the new bootstrap code
- try to understand the problems with packages
- committed a fix and test for packages that
Esteban wrote:
sorry, but I still do not understand why this is a problem.
can you elaborate why it should not be more implementations?
It breaks the system. You can no longer use senders.
So currently, all Moose and Seaside based 3.0 images are
somewhat unusable.
Stephan
I tried committing/copying an updated slice, but am unable to get it into the
Pharo30Inbox,
even though I can copy packages in it. Is this a known issue?
Stephan
Nice!
I directly tried it loading into 30618 with te latest vm on OS X Mavericks.
When loading the configuration as shown in the manual, there was a complaint
about
a class not found yet. The examples showed up fine though. They look great.
The first one with the square/cube I couldn’t get the
Here’s a new slice, the corresponding packages are in the inbox.
SLICE-Issue-12324-There-should-be-no-class-side-implementations-of-packages-except-for-ClassDescription-StephanEggermont.2.mcz
Description: Binary data
Mariano wrote:
You will only have problems with class side #packages.
Are all those 35 class side?
5 out of 35. 4 of which are in Pharo 3
Stephan
Stef wrote:
The real problem is that a CLASS now responds to the packages message.
SUnitAPIDocumentation also implements packages class-side.
If I try to add a instance side method #bla I get a debugger
(but the method is added). Searching for implementors then
also returns a debugger. I
I could copy the separate packages, but not the slice.
Please note that the added test is expected to fail on all builds containing
Grease-Core.
Stephan
Hi Sean,
Didi you compare it to Metaceller in Moose?
Stephan
The smalltalkhub build has been red for quite some time now. If I want to run a
hub,
should I take the latest build?
Stephan
Can anyone put it in place? Can’t seem to commit it
Stef wrote:
Do you think that if an external library redefines the semantics of #class, it
will be a Pharo bug?
I don’t know. I know I have 35 implementers of #packages in my Moose image, and
191 senders.
And I know that in Magritte 3 we changed from using #description to
#magritteDescription
Chris wrote:
Sean wrote:
In general in UIs, he best approach is to do what the user requests without
hesitation, and make everything undoable.
Ah, thanks for saying that. People need to be reminded. I think the
tablet craze has lurched the world back into modal UI's being ok”.
I don’t
The configuration of XMLWriter refers to the wrong repository.
That breaks the moose build.
I don’t seem to have commit rights to PharoExtras, so
ConfigurationOfXMLWriter-StephanEggermont.41.mcz
Description: Binary data
Nice!
kilon wrote:
any links why newcomers like me should be impressed ? Whats the advantages ?
Then we can now have more compact versions of
FreeTypeSettings LabelMorph MorphTreeNodeMorph MorphTreeListManager
ThemeSettings MultistateButtonMorph PharoUserPermissions
Camillo wrote:
I love these scripts :P..
maybe you should make gists [1] out of them :)
Good idea
https://gist.github.com/StephanEggermont/7598946
Nicolas wrote:
It has to be more simple than that…
Probably. I don’t do this often enough.
Why not searching for allReferences to true and false, and gather the class of
referents?
How is that going to give me the values of the instVars?
This would also remove the limitation of first 10
hernan/BioSmalltalk
Is what DeprecationFinder me tells
Stephan
Verstuurd vanaf mijn iPhone
I have for you
GraphViz-hfm.46.mcz
GraphViz-jrp.42.mcz
GraphViz-jrp.41.mcz
As soon as my ADSL is up again.
The first is on smalltalkhub, btw
Stephan
Verstuurd vanaf mijn iPhone
Doru wrote:
I need a bit of help with the issue of the SpecDebugger relying on the default
inspector being implemented in Spec.
Hmm, one gets used to having a decent inspector. Going back to the Pharo 3
default is not funny…
The GTInspector might be a bit of an ugly duckling, but there really
With the QCMagritte image (from the pharo contributions ci).
That has build-in support for translations. Just start it and open
the web demo
Stephan
How could we recognize the using of a deprecated stream?
With a tool like DeprecationFinder it should be easy to walk the repositories
and find out where they are used.
Stephan
Norbert wrote:
You may have a look at PParser#delimitedBy:
Thank you.
Stephan
Hi Sean,
In PetitDelphi, we defined a few convenience methods
like plusFor:separatedByToken: that you might find useful.
Stephan
Hi! We're sending this automatic email twice a month, to give the community an
opportunity to easily know what's happening and to coordinate efforts. Just
answer informally, and feel free to spawn discussions thereafter!
### Here's what I've been up to since the last WhatsUp:
Put out CFP
I’d strongly prefer auto-generated mails to be moved to the issue tracker list.
Anything that doesn’t have decent subjects has no place in this mailing list
[pharo-project/pharo-core]
is no decent subject, and neither is
[pharo-project/pharo-core] 612dc4: 30542
Stephan Eggermont
SmalltalkHub uses Seaside 3.0.8, referring to an old version of Grease.
Upgrading to a newer Seaside should solve that problem, AFAIK
Stephan
Commit dates seem to be wrong. I committed a few packages in Moose/PetitDelphi
today,
and find them back as 17/05/2013
Stephan
Nico wrote:
The creation of the MC packages are from today too?
Name: PetitDelphi-StephanEggermont.129
Author: StephanEggermont
Time: 24 October 2013, 3:54:21.159 pm
UUID: b310f98a-41cf-4dd6-8be4-d1954ae69e4b
Ancestors: PetitDelphi-StephanEggermont.128
Moved some AST classes
(from pharo -dev)
PharoVM-AcceptanceTest loads Moose-Stable. That is based on a snapshot.
As the snapshot flattens the configuration, it is basically unpatchable.
Stephan
Moose.Test.HismoImporter.MooseHismoImporterTest.testEvery15Days 64 ms
3
Moose.Test.HismoImporter.MooseHismoImporterTest.testEvery2Years 52 ms
3
Moose.Test.HismoImporter.MooseHismoImporterTest.testEvery6Months51 ms
3
Don't know where they come from. The implementation
There has been a fixed package on smalltalkhub since 10 days,
so it looks like a configuration problem
Name: Moose-Test-HismoImporter-StephanEggermont.5
Author: StephanEggermont
Time: 20 September 2013, 2:45:53.012 pm
UUID: 766c5b47-daa6-4b51-af8f-091539186c59
Ancestors:
Hi Marcel,
my goal is to find the big ones to take a closer look at. Thus, next steps
would be to fetch the latest MCZ of all projects and then analyze their
contents. :)
That is in DeprecationFinder. (load in a Moose 4.8 after loading NeoJSon).
Current analysis is just getting the
And please don't all at the same time download and run the DeprecationFinder
code,
until Nicolas tells you he wants to do a load test!
Stephan
Yes, it is dead. The number of answers is getting lower and lower, the
questions get more and more detailed.
http://stackoverflow.com/questions/18946662/if-pypy-is-6-3-times-faster-than-cpython-why-not-just-use-faster-interpreter
Hmm yes, that is a question that needs to be closed. Interesting,
Kilon wrote:
Excuse me I have just an allergy in the word dead so I could not resist. I
remember for 3 days in a row there was some people popping in #smalltalk irc
channel asking , why smalltalk is dead, how dead it is, why the corpse still
smells and what the autopsy revealed :D. I find
Hannes wrote:
In reply to this post by Stephan Eggermont
On 9/23/13, Stephan Eggermont [hidden email] wrote:
Kilon wrote
Here is a radical suggestion you probably don't want to hear.
Close down Pharo users mailing list, redirect everyone to stackoverflow.
Definitely
These projects and packageversions on Smalltalkhub
use the #metaclass selector.
This might refer to a deprecated selector, replaced by #theMetaClass
The packageversions were selected as latest+not merged
gokr/DeltaStreams System-gk.143
StephaneDucasse/ArchiveTraits TraitsUI-al.19
Track and tag your change/Fix/Enhancement:
• Add an entry to the Issues tracker describing your change (include
the package version name of the package). Set the status to Fixed.
That doesn't work. Possible status changes are 'Working on' and 5 variations on
'Work needed'.
And 11628 is
Camille wrote:
You need to click on 'resolve' instead of 'edit' I think.
Ah, thank you. That works.
Stephan
Yes, smalltalkhub is nice. And installing your own is very easy. I think more
people should do so.
It doesn't seem to be as solid as ss3, but can be installed much easier,
and on legacy platforms (windows).
Running your own mirror should not take that much space and capacity, and makes
you
less
Around 12.00 we managed to load and analyze upto 100 mczs/minute from
Smalltalkhub.
We hoped that we would not be able to overload Smalltalkhub from a home ADSL
connection, using a single (non multi-threaded) Moose image.
However, this seems to coincide with some failing jenkins builds. We
Igor wrote:
i think you should contact Nico about how to run tests
using some test server first, and only if it passes, on deployed one :)
Oh, the code was working fine ;) Really no different to copying
monticello packages from one repo to another. It's just that
I try to find all invocations in
Sven wrote
StHub is an important shared community resource.
Definitely.
And this is a stop-the-line problem.
Having a single image on an ADSL connection bring down
smalltalkhub doing one mcz request at a time means there is a
serious dos risk.
This is my packagecache from the offending run:
We noticed smalltalkhub is currently down. We have been generating
some load, walking smalltalkhub a bit today, to see how we could get
access to the latest versions of all packages. That should help us
determine which projects use deprecated methods.
Diego Stephan
Found some improvements for the wiki page on fogbugz.
I don't see a way to make changes myself.
- refer to the smalltalkhub repositories, explain the squeaksource
history;
- the flow is not so clear in the difference between describing the
contribution and where to put it (attachment, paste
works.
But that doesn't mean that I have all the answers...
Stephan
When running Pharo 30366 tests on Pharo VM 6.0-pre (21.0) on a 10.6.8 mac
I get different results when I open the image and run the tests compared to
when I saved and quit the image and then re-run the tests.
2nd time results:
11287 run, 11232 passes, 12 skipped, 26 expected failures, 6
Hmm, and somewhat later the image freezes
601 - 700 of 739 matches
Mail list logo