Em 05-04-2011 18:05, Mariano Martinez Peck escreveu:


On Mon, Apr 4, 2011 at 5:19 PM, Marcus Denker <marcus.denker@inria.fr> wrote:

On Apr 4, 2011, at 5:05 PM, Mariano Martinez Peck wrote:



On Sat, Mar 12, 2011 at 8:19 PM, Mariano Martinez Peck <marianopeck@gmail.com> wrote:
So....here is the list of things I guess are pending:



In general, for all pending things adding entries in the bug tracker is the only way ti make sure they are not forgotten.

1) "Release" all configurations included in Pharo (this is, ConfigurationOfPharo and all it dependencies). I mean, we should NOT use or depend on #development versions.

so...can someone check this is done?  GUillermo and me did it once...but it seems a lot of stuff was updated once again

I just gave up and took the image yesterday. Like you did for 1.1: Just build it once, don't care that it is repeatable.

Sorry, I didn't understand. The Pharo 1.1 was repeatable (in the same core) since all the Metacello versions were frozen.
 

 
2) Update the one click image to use latest cog images and "pharo prepared vms". There are some issues regarding this...

Marcus, which is the status of this one ?

I updated the one-click image with the last cog VM. 

For the one-click, it is identical to 1.1 with the only change that I put the windows VM I got from Torsten.


But 1.1 was not-Cog  (Cog 1.1.1 was cog). So...we are shipping the Pharo One Click 1.2  with the Interpreter VM instead of Cog ?
or did you mean 1.1.1 ?  if so, it would be good to update to latest Cog. The cog version used for 1.1.1 had several problems (I think the run:with:in: crashes the image -> TestCoverage)
In linux (Fedora 14) when using diacritics '`~^", cog messes inputs (at least pt_BR) and fail to interpret properly keyboard entries (like áéíóúàãõâê, etc). Thus typing ' + a won't give á but a?

It's not a pharo problem but it's cog one. I suggest traditional vm is hold in use while cog don't address these input methods problem.



 
3) Write down all the things/summary done in Pharo 1.2

consider as done in nobody else comes with something better,
 
5) Update  SystemVersion current -->> Pharo1.2rc2 of 27 January 2011 update 12340


SystemVersion current ->> Pharo1.2.1 of 2 April 2011 update 12345

so..it is ok, isn't it ? can we forget this item ?
 
6) Condense sources? condense changes?  #cleanUpForRelease


Marcus, is that necessary ?

DEVImageCreator >> cleanUps
   
    ScriptLoader new cleanUpForRelease.
    "In Pharo 1.1 condenseChanges is VERY slow because of the new method trailers. So condensingChanges each time even when it is not needed is not a good idea anymore. So in case a condenseChanges is needed it should be done explicitly.
    Smalltalk condenseChanges."
    Smalltalk garbageCollect.
 
I did that and I put that comment. Now that they are faster and that we have Hudson, should I put back the Smalltalk condenseChanges ?    is there anything else needed?  do we need a condenseSources ?



CondenseChange and CondenseSource is broken right now.


Ok, we can live withoput that :)
 
Marcus


--
Marcus Denker  -- http://www.marcusdenker.de
INRIA Lille -- Nord Europe. Team RMoD.




--
Mariano
http://marianopeck.wordpress.com