Smalltalk › Frameworks & Tools › SciSmalltalk

Configurationproblem

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

9 messages Options

stepharo

Configurationproblem

Hi werner

I'm trying to understand why
- some packages are dirty (chromosone, server....)
- some classes got lost
- the extensions we did with didier are not loaded.

I will restart from my last configuration (which was bogus but I know
what I did)
and see if I can get something stable.

Then I will have a look or you can explain to me what you did.

Stef

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

stepharo

Re: Configurationproblem

After I will include

Math-Tests-Random-WernerKassens.12

Stef

Le 23/3/16 12:17, stepharo a écrit :

> Hi werner
>
> I'm trying to understand why
> - some packages are dirty (chromosone, server....)
> - some classes got lost
> - the extensions we did with didier are not loaded.
>
>
> I will restart from my last configuration (which was bogus but I know
> what I did)
> and see if I can get something stable.
>
> Then I will have a look or you can explain to me what you did.
>
> Stef
>

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

werner kassens-2

Re: Configurationproblem

In reply to this post by stepharo

Hi Stephane

On Wed, Mar 23, 2016 at 12:17 PM, stepharo <[hidden email]> wrote:

I'm trying to understand why
- some packages are dirty (chromosone, server....)

i noticed that too but i have to admit i did not yet think about this. i simply assumed it would be something outside of scismalltalk and would <stupid grin> go away by itself if i wait a bit ( because i noticed this somewhere else too, if i remember correctly)

- some classes got lost
- the extensions we did with didier are not loaded.

i did not notice this, but i did not look for this, what exactly got lost or is not loaded? and which config number should i use to look into this (StephaneDucasse.164)? (at the moment i have visitors, hence not too much time.)
and Math-PrincipleComponentAnalysis was loaded btw.

werner

werner kassens-2

Re: Configurationproblem

Hi Stephane,
during preparing a fond de volaille i remembered that i also occasionally lost some objects in packages. if you press changes in the monticello browser before saving a package you will in those cases see that for monticello the objects are gone although at the same time they are existent for the system browser and the rest of the system. but i dont remember what i did to make them visible to monticello again (perhaps flush the cache?).
werner

On Wed, Mar 23, 2016 at 1:19 PM, werner kassens <[hidden email]> wrote:

Hi Stephane

On Wed, Mar 23, 2016 at 12:17 PM, stepharo <[hidden email]> wrote:
I'm trying to understand why
- some packages are dirty (chromosone, server....)
i noticed that too but i have to admit i did not yet think about this. i simply assumed it would be something outside of scismalltalk and would <stupid grin> go away by itself if i wait a bit ( because i noticed this somewhere else too, if i remember correctly)
- some classes got lost
- the extensions we did with didier are not loaded.
i did not notice this, but i did not look for this, what exactly got lost or is not loaded? and which config number should i use to look into this (StephaneDucasse.164)? (at the moment i have visitors, hence not too much time.)
and Math-PrincipleComponentAnalysis was loaded btw.

werner

stepharo

Re: Configurationproblem

In reply to this post by stepharo

Ok so the bar is green

I merged back some changes, used the most recent baseline.
I will check a bit your changes werner now if you can tell what you were
trying to achieve it will help me.

Else I will start to work on Polymath.

Stef

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

werner kassens-2

Re: Configurationproblem

On Wednesday, March 23, 2016 at 5:44:56 PM UTC+1, stepharo wrote:

I will check a bit your changes werner now if you can tell what you were
trying to achieve it will help me.

Hi Stephane,
i was trying to make a package for interval-calculation (for optimization and such). unfortunately i had to overwrite some standard pharo methods. of course this is bad style for a public package and the user should at least be able to choose whether he wants to use such a package or not. hence i thought i make a config for this package and put that config into the standard scismalltalk-image so that one can easily check out the package but it is not loaded by default.

because of this ugly overwriting of standard methods, it is of course possible that even a standard image will crash and i have no possibilty to check that for 64bit pc's. it does not happen (at least for now) with 32bit pc's. that was the reason for my original question. but even if it does not crash now, that does not mean that it wont happen in the future, hence it was probably a bad idea to upload it, and i would suppose it should not be part of scismalltalk or respectively polymath. overwriting standard methods is probably too nasty style.
werner

stepharo

Re: Configurationproblem

Ok then the first step I would do is propose a separated configuration so that we can identify the problem.
Why not simple provide another class of Interval

SmartInterval, OptimizedInterval

Like that no overwrite.

Stef

Le 24/3/16 11:55, werner kassens a écrit :

On Wednesday, March 23, 2016 at 5:44:56 PM UTC+1, stepharo wrote:
I will check a bit your changes werner now if you can tell what you were
trying to achieve it will help me.

Hi Stephane,
i was trying to make a package for interval-calculation (for optimization and such). unfortunately i had to overwrite some standard pharo methods. of course this is bad style for a public package and the user should at least be able to choose whether he wants to use such a package or not. hence i thought i make a config for this package and put that config into the standard scismalltalk-image so that one can easily check out the package but it is not loaded by default.

because of this ugly overwriting of standard methods, it is of course possible that even a standard image will crash and i have no possibilty to check that for 64bit pc's. it does not happen (at least for now) with 32bit pc's. that was the reason for my original question. but even if it does not crash now, that does not mean that it wont happen in the future, hence it was probably a bad idea to upload it, and i would suppose it should not be part of scismalltalk or respectively polymath. overwriting standard methods is probably too nasty style.
werner

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

werner kassens-2

Re: Configurationproblem

Hi Stephane,
yes, i had made a separate class of interval and also a separated config. i had encapsulated the overwriting (some math operations in Number classes) for pharo 5.0 in package:
Math-RealIntervalCompatibility-WernerKassens.4
the problem i have is twofold:
1. it is absolutely unclear to me, whether it makes any sense at all to upload a project that uses this nasty style of overwriting standard methods? if one does this of course one should not load it by default but should let the user decide whether he wants to load it, therefore i made the config. but perhaps having a project that does this overwriting is generally unwanted in scismalltalk? (if it is unwanted that would make sense to me.)
2. i have no 64bit pc and can't test whether a pharo5.0 spur image would crash _in the moment_ Math-RealIntervalCompatibility-WernerKassens.4 would be loaded. although i do not really expect it, from my experience with these things it _is_ a distinct possibility.
thanks for going with me through these problems.
werner

On Thursday, March 24, 2016 at 2:14:19 PM UTC+1, stepharo wrote:

Ok then the first step I would do is propose a separated configuration so that we can identify the problem.
Why not simple provide another class of Interval

SmartInterval, OptimizedInterval

Like that no overwrite.

Stef

Le 24/3/16 11:55, werner kassens a écrit :

On Wednesday, March 23, 2016 at 5:44:56 PM UTC+1, stepharo wrote:
I will check a bit your changes werner now if you can tell what you were
trying to achieve it will help me.

Hi Stephane,
i was trying to make a package for interval-calculation (for optimization and such). unfortunately i had to overwrite some standard pharo methods. of course this is bad style for a public package and the user should at least be able to choose whether he wants to use such a package or not. hence i thought i make a config for this package and put that config into the standard scismalltalk-image so that one can easily check out the package but it is not loaded by default.

because of this ugly overwriting of standard methods, it is of course possible that even a standard image will crash and i have no possibilty to check that for 64bit pc's. it does not happen (at least for now) with 32bit pc's. that was the reason for my original question. but even if it does not crash now, that does not mean that it wont happen in the future, hence it was probably a bad idea to upload it, and i would suppose it should not be part of scismalltalk or respectively polymath. overwriting standard methods is probably too nasty style.
werner

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="v7FfAsGtBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;">scismalltalk...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href='https://groups.google.com/d/optout';return true;" onclick="this.href='https://groups.google.com/d/optout';return true;">https://groups.google.com/d/optout.

stepharo

Re: Configurationproblem

Hi Stephane,
yes, i had made a separate class of interval and also a separated config. i had encapsulated the overwriting (some math operations in Number classes) for pharo 5.0 in package:
Math-RealIntervalCompatibility-WernerKassens.4
the problem i have is twofold:
1. it is absolutely unclear to me, whether it makes any sense at all to upload a project that uses this nasty style of overwriting standard methods? if one does this of course one should not load it by default but should let the user decide whether he wants to load it, therefore i made the config. but perhaps having a project that does this overwriting is generally unwanted in scismalltalk? (if it is unwanted that would make sense to me.)

I'm pragmatic :)
In general I try to avoid overwriting and we should minimize their use. Now may be for your special case you need it
and ok people should accept the risk. This is why having a good test suite is important.

2. i have no 64bit pc and can't test whether a pharo5.0 spur image would crash _in the moment_ Math-RealIntervalCompatibility-WernerKassens.4 would be loaded.

Wait we do not have a Pharo 64 bits image so we cannot test it either.

although i do not really expect it, from my experience with these things it _is_ a distinct possibility.
thanks for going with me through these problems.
werner

On Thursday, March 24, 2016 at 2:14:19 PM UTC+1, stepharo wrote:

Ok then the first step I would do is propose a separated configuration so that we can identify the problem.
Why not simple provide another class of Interval

SmartInterval, OptimizedInterval

Like that no overwrite.

Stef

Le 24/3/16 11:55, werner kassens a écrit :

On Wednesday, March 23, 2016 at 5:44:56 PM UTC+1, stepharo wrote:
I will check a bit your changes werner now if you can tell what you were
trying to achieve it will help me.

Hi Stephane,
i was trying to make a package for interval-calculation (for optimization and such). unfortunately i had to overwrite some standard pharo methods. of course this is bad style for a public package and the user should at least be able to choose whether he wants to use such a package or not. hence i thought i make a config for this package and put that config into the standard scismalltalk-image so that one can easily check out the package but it is not loaded by default.

because of this ugly overwriting of standard methods, it is of course possible that even a standard image will crash and i have no possibilty to check that for 64bit pc's. it does not happen (at least for now) with 32bit pc's. that was the reason for my original question. but even if it does not crash now, that does not mean that it wont happen in the future, hence it was probably a bad idea to upload it, and i would suppose it should not be part of scismalltalk or respectively polymath. overwriting standard methods is probably too nasty style.
werner

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a moz-do-not-send="true" href="javascript:" target="_blank" gdf-obfuscated-mailto="v7FfAsGtBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;">scismalltalk...@googlegroups.com.
For more options, visit <a moz-do-not-send="true" href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href='https://groups.google.com/d/optout';return true;" onclick="this.href='https://groups.google.com/d/optout';return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "SciSmalltalk" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.