Ben Coman | 3 Sep 06:10 2014

nil pointersTo size

I am curious... what does it mean that
    nil PointersTo size  
is a large number? Intuitively I'd have thought the answer would be zero.

Pharo build 40196 -->  ~101,000
Pharo build 30856 -->  ~103,000
Pharo build 20628 -->  ~80,000
Squeak 4.5 -->  ~40,000

cheers -ben

Sebastian Sastre | 3 Sep 00:49 2014
Yuriy Tymchuk | 2 Sep 20:47 2014

Any changes with fonts?


when I cannot see system fonts anymore after selecting Free Type font checkbox in settings. Are there any
changes to fonts? Of something broke down?



NonInteractiveTranscript and Debugging issue in headless image

I've got this error when the debugger is triggered in a headless image.

So, #MNU.

I had a look at the NonInteractiveTranscript and it has this method missing indeed in 3.0.

I've added a basic one but we could benefit from enhanced debugging abilities on that front.

As I am running on a server without any graphics abilities, it would be welcome to have this fixed.

I added this to the startup script to patch the thing:

Author fullName: 'PhilippeBack'. "Avoid yet another question..."

NonInteractiveUIManager compile: 'openDebuggerOn: process context: context label: title contents: contentsStringOrNil fullView: bool 

        ^ self nonInteractiveWarning: ''Opening Debugger''' classified: #'ui-requests'.

So far so good, but it turned to:

An attempt to use interactive tools detected, while in non-interactive mode
Opening Debugger
SpecDebugger class>>openOn:context:label:contents:fullView:
GRPharoPlatform>>openDebuggerOn: in Block: [ process debug: anError signalerContext title: an...etc...

And the vm stopped again.

Looks like:

NonInteractiveUIManager>>nonInteractiveWarning: aWarnMessage
^ ErrorNonInteractive signal: aWarnMessage

isn't too friendly.

How would one get debugging info shown and just proceed or close the debugger?

Closing the image is really hard to swallow here.



Torsten Bergmann | 2 Sep 16:53 2014

Surrounding text in editors


1. When I open a workspace and type


   in it. Now I select the whole text and hit the $" key
   to make it a comment. This works, the workspace


   on both sides and I can continue to work with my comment.

2. When I open a workspace and type


   in it. Now I select the whole text and hit the $' key
   to make it a string. This does not works, the workspace
   again shows:


   on both sides - which is a comment, but I would have expected a string:


Is this a known problem. Is it related to Windows only - can one
check on other platforms? Should we open an issue?

Seems to work well for brackets and braces by the way: 
   Selecting Hello with an additional $[ key gives correctly [Hello] 
   Selecting Hello with an additional $( key gives correctly (Hello) 


Yuriy Tymchuk | 2 Sep 16:21 2014

#literals in RBSmalllintContext

Hi, does anybody know what #literals in RBSmalllintContext stand for? Because it’s calculated in a very
weird way with the use of literalSemaphore and literalProcess variables.


Yuriy Tymchuk | 2 Sep 14:18 2014

Rewrite rules doc?

Hi guys. Is there any doc about rewrite rules besides ?


GitHub | 2 Sep 11:25 2014


  Branch: refs/tags/40196
GitHub | 2 Sep 11:25 2014

[pharo-project/pharo-core] d45fa4: 40196

  Branch: refs/heads/4.0
  Commit: d45fa418c5b78c474497721b97328f43706c7e4b
  Author: Jenkins Build Server <board@...>
  Date:   2014-09-02 (Tue, 02 Sep 2014)

  Changed paths:
    A OpalCompiler-Core.package/extension/TApplyingOnClassSide/instance/
    A OpalCompiler-Core.package/extension/TApplyingOnClassSide/instance/
    A OpalCompiler-Core.package/extension/TClass/instance/
    A OpalCompiler-Core.package/extension/TClass/instance/
    A ScriptLoader40.package/ScriptLoader.class/instance/pharo - scripts/
    A ScriptLoader40.package/ScriptLoader.class/instance/pharo - updates/
    M ScriptLoader40.package/ScriptLoader.class/instance/public/
    A Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/class/accessing/
    A Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/class/defaults/
    M Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/
    A Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/instance/actions/
    M Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/instance/menu/
Tool-ConfigurationBrowser.package/MetacelloConfigurationBrowserPane.class/instance/private accessing/

  Log Message:
13941 [FEATURE]: Configuration Browser: Switch to Any Repository

13938 Opal, custom compilers and class side methods

Sven Van Caekenberghe | 2 Sep 09:47 2014

Pharo Project Catalog Should look at Pharo 3


It seems the Pharo Project Catalog CI job looks at Configurations in MetaRepoForPharo2. Would it not be
better if it (also) looked at MetaRepoForPharo3/main ?


Tim Mackinnon | 2 Sep 00:41 2014

Jenkins builds not properly preserving artifacts?

Guys - I’ve noticed that when a build completed on the Inria infrastructure, for a short time the console output correctly shows all the artefacts that were used to create that build - e.g. the slices loaded etc. However after a short time - the output then reverts to something like below (which isn’t as useful): 

Started by upstream project "[8mha:AAAAqh+LCAAAAAAAAP9b85aBtbiIQTGjNKU4P08vOT+vOD8…. originally caused by: [URLTrigger] A change within the response URL invocation (log) Building remotely on pharo-contribution-winxp in workspace C:\builds\workspace\PharoLauncher Deleting project workspace… done---
I’ve looked at a few other jobs (I noticed this for Pharo Launcher) and they seem the same. I think the jobs aren’t configured properly? - As normally (at least in other languages like Java), the intent is that the console output reflects what was on the console when the job was run. I have a hunch that something is overwriting these files before they are archived away.
As we don’t seem to have other details of what went into a build job (which is another issue - I would expect the changes page for a build to also show this info in a better format).
Do you think this could be fixed?