RPackage classes name

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

RPackage classes name

Franck Warlouzet
Hello,

Currently I am working on groups in Nautilus, and I am actually reimplementing them because I could not just fix them (There are a lot of bugs which make them almost unusable in practice), it is too complicated for nothing. Doing this I somehow have to use RPackage announcements and I am a little bit confused by their name.
When you remove a package, an announcement RPackageUnregistered is raised, but I do not know if there is a difference with RPackageRemoved (which by the way does not exist, but I was expecting a name like that). What is this registration thing ? There is no RPackageRegistered but there is RPackageCreated.

So I am confused. Does someone know if I should create an announcement RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It does not seem consistent and so it is confusing. Or can someone explain to me the notion of registration for the RPackages ?

Thanks,

Franck
Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

Nicolai Hess


2015-07-18 15:48 GMT+02:00 Franck Warlouzet <[hidden email]>:
Hello,

Currently I am working on groups in Nautilus, and I am actually reimplementing them because I could not just fix them (There are a lot of bugs which make them almost unusable in practice), it is too complicated for nothing. Doing this I somehow have to use RPackage announcements and I am a little bit confused by their name.
When you remove a package, an announcement RPackageUnregistered is raised, but I do not know if there is a difference with RPackageRemoved (which by the way does not exist, but I was expecting a name like that). What is this registration thing ? There is no RPackageRegistered but there is RPackageCreated.

So I am confused. Does someone know if I should create an announcement RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It does not seem consistent and so it is confusing. Or can someone explain to me the notion of registration for the RPackages ?

Thanks,

I don't know.
I would rename RPackageCreated to RPackageRegistered.


nicolai


 

Franck

Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

Thierry Goubier
In reply to this post by Franck Warlouzet
Hi Franck,

Le 18/07/2015 15:48, Franck Warlouzet a écrit :

> Hello,
>
> Currently I am working on groups in Nautilus, and I am actually
> reimplementing them because I could not just fix them (There are a lot
> of bugs which make them almost unusable in practice), it is too
> complicated for nothing. Doing this I somehow have to use RPackage
> announcements and I am a little bit confused by their name.
> When you remove a package, an announcement RPackageUnregistered is
> raised, but I do not know if there is a difference with RPackageRemoved
> (which by the way does not exist, but I was expecting a name like that).
> What is this registration thing ? There is no RPackageRegistered but
> there is RPackageCreated.
>
> So I am confused. Does someone know if I should create an announcement
> RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It
> does not seem consistent and so it is confusing. Or can someone explain
> to me the notion of registration for the RPackages ?

I'd say that, at the moment, the RPackage code is a bit half-way through
a complete handling of all those aspects; some decisions about packages
are left outside the RPackage code, in Monticello in particular.

So objects tracking packages changes should also have a look into the
Monticello related announcements which are (digging through the
AltBrowser code)... MCWorkingCopyCreated, MCWorkingCopyModified,
MCWorkingCopyDeleted, in addition to RPackageCreated,
RPackageUnregistered. For example, AltBrowser, for its package
categories and browsing environments (aka groups), tracks all of them.

I haven't checked, but I suspect that only Monticello can really delete
a package (by unloading it) and that RPackageOrganizer will react to it
by unregistering the package. Another possibility is the removal of a
system category.

So renaming RPackageUnregistered as RPackageRemoved requires significant
changes in RPackageOrganizer, for which I'd say beware: this is a very
good way of confirming that Pharo5 is alpha software ;)

Thierry

Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

Franck Warlouzet
Hello,

Thanks for the explanation ! So Nicolai is probably right, RPackageCreated should be named RPackageRegistered to be consistent.

Franck

> Date: Sun, 19 Jul 2015 13:47:37 +0200

> From: [hidden email]
> To: [hidden email]
> Subject: Re: [Pharo-dev] RPackage classes name
>
> Hi Franck,
>
> Le 18/07/2015 15:48, Franck Warlouzet a écrit :
> > Hello,
> >
> > Currently I am working on groups in Nautilus, and I am actually
> > reimplementing them because I could not just fix them (There are a lot
> > of bugs which make them almost unusable in practice), it is too
> > complicated for nothing. Doing this I somehow have to use RPackage
> > announcements and I am a little bit confused by their name.
> > When you remove a package, an announcement RPackageUnregistered is
> > raised, but I do not know if there is a difference with RPackageRemoved
> > (which by the way does not exist, but I was expecting a name like that).
> > What is this registration thing ? There is no RPackageRegistered but
> > there is RPackageCreated.
> >
> > So I am confused. Does someone know if I should create an announcement
> > RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It
> > does not seem consistent and so it is confusing. Or can someone explain
> > to me the notion of registration for the RPackages ?
>
> I'd say that, at the moment, the RPackage code is a bit half-way through
> a complete handling of all those aspects; some decisions about packages
> are left outside the RPackage code, in Monticello in particular.
>
> So objects tracking packages changes should also have a look into the
> Monticello related announcements which are (digging through the
> AltBrowser code)... MCWorkingCopyCreated, MCWorkingCopyModified,
> MCWorkingCopyDeleted, in addition to RPackageCreated,
> RPackageUnregistered. For example, AltBrowser, for its package
> categories and browsing environments (aka groups), tracks all of them.
>
> I haven't checked, but I suspect that only Monticello can really delete
> a package (by unloading it) and that RPackageOrganizer will react to it
> by unregistering the package. Another possibility is the removal of a
> system category.
>
> So renaming RPackageUnregistered as RPackageRemoved requires significant
> changes in RPackageOrganizer, for which I'd say beware: this is a very
> good way of confirming that Pharo5 is alpha software ;)
>
> Thierry
>
Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

Thierry Goubier
Le 19/07/2015 13:58, Franck Warlouzet a écrit :
> Hello,
>
> Thanks for the explanation ! So Nicolai is probably right,
> RPackageCreated should be named RPackageRegistered to be consistent.

Yes, Nicolai is right. RPackageCreated is announced in two places:
        RPackageOrganizer>>registerPackage:
and
        RPackageOrganizer>>ensureExistAndRegisterPackageNamed:

(i.e. both are "registerPackage" methods).

A simple class rename refactoring should be enough ;)

Thierry

>
> Franck
>
>  > Date: Sun, 19 Jul 2015 13:47:37 +0200
>  > From: [hidden email]
>  > To: [hidden email]
>  > Subject: Re: [Pharo-dev] RPackage classes name
>  >
>  > Hi Franck,
>  >
>  > Le 18/07/2015 15:48, Franck Warlouzet a écrit :
>  > > Hello,
>  > >
>  > > Currently I am working on groups in Nautilus, and I am actually
>  > > reimplementing them because I could not just fix them (There are a lot
>  > > of bugs which make them almost unusable in practice), it is too
>  > > complicated for nothing. Doing this I somehow have to use RPackage
>  > > announcements and I am a little bit confused by their name.
>  > > When you remove a package, an announcement RPackageUnregistered is
>  > > raised, but I do not know if there is a difference with RPackageRemoved
>  > > (which by the way does not exist, but I was expecting a name like
> that).
>  > > What is this registration thing ? There is no RPackageRegistered but
>  > > there is RPackageCreated.
>  > >
>  > > So I am confused. Does someone know if I should create an announcement
>  > > RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It
>  > > does not seem consistent and so it is confusing. Or can someone explain
>  > > to me the notion of registration for the RPackages ?
>  >
>  > I'd say that, at the moment, the RPackage code is a bit half-way through
>  > a complete handling of all those aspects; some decisions about packages
>  > are left outside the RPackage code, in Monticello in particular.
>  >
>  > So objects tracking packages changes should also have a look into the
>  > Monticello related announcements which are (digging through the
>  > AltBrowser code)... MCWorkingCopyCreated, MCWorkingCopyModified,
>  > MCWorkingCopyDeleted, in addition to RPackageCreated,
>  > RPackageUnregistered. For example, AltBrowser, for its package
>  > categories and browsing environments (aka groups), tracks all of them.
>  >
>  > I haven't checked, but I suspect that only Monticello can really delete
>  > a package (by unloading it) and that RPackageOrganizer will react to it
>  > by unregistering the package. Another possibility is the removal of a
>  > system category.
>  >
>  > So renaming RPackageUnregistered as RPackageRemoved requires significant
>  > changes in RPackageOrganizer, for which I'd say beware: this is a very
>  > good way of confirming that Pharo5 is alpha software ;)
>  >
>  > Thierry
>  >


Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

Franck Warlouzet
Ok I will do that, thanks again

Fanck

> Date: Sun, 19 Jul 2015 15:12:06 +0200

> From: [hidden email]
> To: [hidden email]
> Subject: Re: [Pharo-dev] RPackage classes name
>
> Le 19/07/2015 13:58, Franck Warlouzet a écrit :
> > Hello,
> >
> > Thanks for the explanation ! So Nicolai is probably right,
> > RPackageCreated should be named RPackageRegistered to be consistent.
>
> Yes, Nicolai is right. RPackageCreated is announced in two places:
> RPackageOrganizer>>registerPackage:
> and
> RPackageOrganizer>>ensureExistAndRegisterPackageNamed:
>
> (i.e. both are "registerPackage" methods).
>
> A simple class rename refactoring should be enough ;)
>
> Thierry
>
> >
> > Franck
> >
> > > Date: Sun, 19 Jul 2015 13:47:37 +0200
> > > From: [hidden email]
> > > To: [hidden email]
> > > Subject: Re: [Pharo-dev] RPackage classes name
> > >
> > > Hi Franck,
> > >
> > > Le 18/07/2015 15:48, Franck Warlouzet a écrit :
> > > > Hello,
> > > >
> > > > Currently I am working on groups in Nautilus, and I am actually
> > > > reimplementing them because I could not just fix them (There are a lot
> > > > of bugs which make them almost unusable in practice), it is too
> > > > complicated for nothing. Doing this I somehow have to use RPackage
> > > > announcements and I am a little bit confused by their name.
> > > > When you remove a package, an announcement RPackageUnregistered is
> > > > raised, but I do not know if there is a difference with RPackageRemoved
> > > > (which by the way does not exist, but I was expecting a name like
> > that).
> > > > What is this registration thing ? There is no RPackageRegistered but
> > > > there is RPackageCreated.
> > > >
> > > > So I am confused. Does someone know if I should create an announcement
> > > > RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It
> > > > does not seem consistent and so it is confusing. Or can someone explain
> > > > to me the notion of registration for the RPackages ?
> > >
> > > I'd say that, at the moment, the RPackage code is a bit half-way through
> > > a complete handling of all those aspects; some decisions about packages
> > > are left outside the RPackage code, in Monticello in particular.
> > >
> > > So objects tracking packages changes should also have a look into the
> > > Monticello related announcements which are (digging through the
> > > AltBrowser code)... MCWorkingCopyCreated, MCWorkingCopyModified,
> > > MCWorkingCopyDeleted, in addition to RPackageCreated,
> > > RPackageUnregistered. For example, AltBrowser, for its package
> > > categories and browsing environments (aka groups), tracks all of them.
> > >
> > > I haven't checked, but I suspect that only Monticello can really delete
> > > a package (by unloading it) and that RPackageOrganizer will react to it
> > > by unregistering the package. Another possibility is the removal of a
> > > system category.
> > >
> > > So renaming RPackageUnregistered as RPackageRemoved requires significant
> > > changes in RPackageOrganizer, for which I'd say beware: this is a very
> > > good way of confirming that Pharo5 is alpha software ;)
> > >
> > > Thierry
> > >
>
>
Reply | Threaded
Open this post in threaded view
|

Re: RPackage classes name

stepharo
Thanks for looking at this issue.

Stef

Le 19/7/15 15:15, Franck Warlouzet a écrit :
Ok I will do that, thanks again

Fanck

> Date: Sun, 19 Jul 2015 15:12:06 +0200
> From: [hidden email]
> To: [hidden email]
> Subject: Re: [Pharo-dev] RPackage classes name
>
> Le 19/07/2015 13:58, Franck Warlouzet a écrit :
> > Hello,
> >
> > Thanks for the explanation ! So Nicolai is probably right,
> > RPackageCreated should be named RPackageRegistered to be consistent.
>
> Yes, Nicolai is right. RPackageCreated is announced in two places:
> RPackageOrganizer>>registerPackage:
> and
> RPackageOrganizer>>ensureExistAndRegisterPackageNamed:
>
> (i.e. both are "registerPackage" methods).
>
> A simple class rename refactoring should be enough ;)
>
> Thierry
>
> >
> > Franck
> >
> > > Date: Sun, 19 Jul 2015 13:47:37 +0200
> > > From: [hidden email]
> > > To: [hidden email]
> > > Subject: Re: [Pharo-dev] RPackage classes name
> > >
> > > Hi Franck,
> > >
> > > Le 18/07/2015 15:48, Franck Warlouzet a écrit :
> > > > Hello,
> > > >
> > > > Currently I am working on groups in Nautilus, and I am actually
> > > > reimplementing them because I could not just fix them (There are a lot
> > > > of bugs which make them almost unusable in practice), it is too
> > > > complicated for nothing. Doing this I somehow have to use RPackage
> > > > announcements and I am a little bit confused by their name.
> > > > When you remove a package, an announcement RPackageUnregistered is
> > > > raised, but I do not know if there is a difference with RPackageRemoved
> > > > (which by the way does not exist, but I was expecting a name like
> > that).
> > > > What is this registration thing ? There is no RPackageRegistered but
> > > > there is RPackageCreated.
> > > >
> > > > So I am confused. Does someone know if I should create an announcement
> > > > RPackageRemoved or rename RPackageCreated into RPackageRegistered ? It
> > > > does not seem consistent and so it is confusing. Or can someone explain
> > > > to me the notion of registration for the RPackages ?
> > >
> > > I'd say that, at the moment, the RPackage code is a bit half-way through
> > > a complete handling of all those aspects; some decisions about packages
> > > are left outside the RPackage code, in Monticello in particular.
> > >
> > > So objects tracking packages changes should also have a look into the
> > > Monticello related announcements which are (digging through the
> > > AltBrowser code)... MCWorkingCopyCreated, MCWorkingCopyModified,
> > > MCWorkingCopyDeleted, in addition to RPackageCreated,
> > > RPackageUnregistered. For example, AltBrowser, for its package
> > > categories and browsing environments (aka groups), tracks all of them.
> > >
> > > I haven't checked, but I suspect that only Monticello can really delete
> > > a package (by unloading it) and that RPackageOrganizer will react to it
> > > by unregistering the package. Another possibility is the removal of a
> > > system category.
> > >
> > > So renaming RPackageUnregistered as RPackageRemoved requires significant
> > > changes in RPackageOrganizer, for which I'd say beware: this is a very
> > > good way of confirming that Pharo5 is alpha software ;)
> > >
> > > Thierry
> > >
>
>