ignite-devel — For developers only.

Re: [Ignite-devel] using cvs effectively

[Fernando F. <th...@gm...>]

i agree.
personally, i'd rather have the full cvs functionality (or subversion) at=
=20
your disposal (syncronize, check out, etc) instead of having things called=
=20
StartCodingSession EndCodingSession, that do some cvs tasks automaticaly.=
=20
actually, i like to have the option of commiting and updating only parts of=
=20
the code, and not all the things that were changed, for instance some crazy=
=20
idea i started working on but don't want to commit yet. anyway, it would=20
definitely be better if we work on the actual program instead of having to=
=20
create a cvs interface.
nano

On 7/17/05, Ricardo Birmann <ric...@gm...> wrote:
>=20
> dear lazy ignite developers,
>=20
> I have been giving our "ignite developers interface" some though
> lately and am starting to give up on the whole idea of integrating cvs
> directly into visualworks smalltalk; instead of taking such a tough
> track towards our goal (developing ignite itself), we should start
> using one of the many cvs interfaces already available, such as:
>=20
> http://sourceforge.net/projects/cvsgui/
>=20
> ....and give our "IDI" the sole responsibility of filling in and out
> our class files into the local working directory (that is been kept in
> synchrony with the cvs repository)
>=20
> yes, yes, it'd be great to just click on "Start/End Coding Session"
> without needing another program; but the trouble we would be going
> through (if we were actually developing anything at all) to implement
> the whole functionality of cvs into such embedded interface could not
> be payed off by what we'd gain from it... simple as that
>=20
> any thoughts on this?
>=20
> Ricardo
>=20
> PS: ... if we are not coding, we should at least try to talk about
> ignite... there's stil some hope for it to become a serious
> project.... hehehe
>=20
>=20
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick
> _______________________________________________
> Ignite-devel mailing list
> Ign...@li...
> https://lists.sourceforge.net/lists/listinfo/ignite-devel
>

[Ignite-devel] using cvs effectively

[Ricardo B. <ric...@gm...>]

dear lazy ignite developers,

I have been giving our "ignite developers interface" some though
lately and am starting to give up on the whole idea of integrating cvs
directly into visualworks smalltalk; instead of taking such a tough
track towards our goal (developing ignite itself), we should start
using one of the many cvs interfaces already available, such as:

http://sourceforge.net/projects/cvsgui/

....and give our "IDI" the sole responsibility of filling in and out
our class files into the local working directory (that is been kept in
synchrony with the cvs repository)

yes, yes, it'd be great to just click on "Start/End Coding Session"
without needing another program; but the trouble we would be going
through (if we were actually developing anything at all) to implement
the whole functionality of cvs into such embedded interface could not
be payed off by what we'd gain from it... simple as that

any thoughts on this?

Ricardo

PS: ... if we are not coding, we should at least try to talk about
ignite... there's stil some hope for it to become a serious
project.... hehehe

[Ignite-devel] reply-to

[Ricardo B. <ric...@gm...>]

the "reply-to" field of the headers of every email sent to this list
are pointing to the user who sent the email, instead of back to the
list...

i will try check (and fix) this later today

Re: [Ignite-devel] desenvolvimento

[Ricardo B. <ric...@gm...>]

> for all that i know of cvs (and i use quite a bit) it doesn't make sense =
to
> uptade after you commit, because you've just uploaded all your info, so e=
ven
> if you update, cvs isn't going to download anything, cause nothing's
> changed. but our program is still going to file in everything, even thoug=
h
> the image is what it's supposed t be...
> the only case in wich i think you have to update and commit is when there=
's
> somebody coding at the same time as you, and commited something after you
> updated, but i think we should resolve this by ending the beginCodingSess=
ion
> endCodingSession routines, and do the whole sync business, where you can =
see
> exactly what you should commit and what you have to update. i think it ma=
kes
> more sense.
>

it might

>  to me it seems that you don't need more than one object in your system
> doing the mediation and executing the shell cvs commands. in the current
> situation what happens is that we instantiate a mediator for every
> interaction (update, commit (wich updates at the end, so instantiates
> another mediator)) and it's up to garbage collection to clean it up.
>  please correct me if i'm wrong.
>

it used to be a singleton, but i dont remember why i changed this
(there was a good reason, i'm sure)... still... i have to take the
time to take a look and will probably do it whithin the next few
days...

i'll post back then....

see you

ricardo

Re: [Ignite-devel] desenvolvimento

[Roberto W. <rhw...@gm...>]

real!

On 6/23/05, Fernando Freire <th...@gm...> wrote:
>=20
> yeah! real time...
>=20
> On 6/23/05, Roberto Winter < rhw...@gm...> wrote:
> >=20
> > ok...
> >=20
> > On 6/23/05, Fernando Freire < th...@gm...> wrote:
> > >=20
> > > opa!
> > >=20
> > > On 6/23/05, Roberto Winter <rhw...@gm...> wrote:
> > > > hello.=20
> > > > shouldn't this be in english?=20
> > > >=20
> > >=20
> > > whatever
> > >=20
> > > > Q. why do we executeCVSUpdate on the CVSMediator>>executeCVSCommit=
=20
> > > message.=20
> > > > A. if I remember correctly, the CVS system has some bizarre working=
s=20
> > > that
> > > > obliges you to do this. it has something to do with keeping the=20
> > > local (or
> > > > remote) "image" of the file structure updated... maybe you should=
=20
> > > check CVS=20
> > > > docs... apart from that I can't remeber very well.=20
> > > >=20
> > >=20
> > > for all that i know of cvs (and i use quite a bit) it doesn't make=20
> > > sense to uptade after you commit, because you've just uploaded all yo=
ur=20
> > > info, so even if you update, cvs isn't going to download anything, ca=
use=20
> > > nothing's changed. but our program is still going to file in everythi=
ng,=20
> > > even though the image is what it's supposed t be...
> >=20
> >=20
> > well. i really don't remember. you should probably check the code for=
=20
> > that message (the executeCVSUpdate) perhaps it has to do more with the=
=20
> > things it does (smalltalk-wise) than CVS. maybe the Birmann can help us=
 out=20
> > here! I don't have the code here, so that does not make is easier!...
> >=20
>=20
> well, the only thing that executeCVSUpdate does, smalltalk wise, is file=
=20
> in all the .st(s) in the rootDirectory, but that doesn't make much sense=
=20
> since you're the one commiting your coding state, so your image should be=
=20
> all good...
>=20

the only tihng I have to say right now is that I clearly remember a=20
discussion about this whole thing, but I really don't remember the outcome,=
=20
but apparently the solution was leaving that message call there... so...=20
maybe leave it there until we remember something or you find out why it=20
shuold be there (and hopefully come up with a better solution!). you can=20
also try removing and risking creating a bug that wasn't there. if you=20
really want to dig deep: remove it and see (real deep) what happens... that=
=20
could work out.

the Birmann is the one that told me to write to the list...., we're waiting=
=20
> for his contact. good thing you wrote though, or i would be here sitting =
and=20
> waiting....
>=20

hehehe=20

 the only case in wich i think you have to update and commit is when there'=
s=20
> > > somebody coding at the same time as you, and commited something after=
 you=20
> > > updated, but i think we should resolve this by ending the beginCoding=
Session=20
> > > endCodingSession routines, and do the whole sync business, where you=
=20
> > > can see exactly what you should commit and what you have to update. i=
 think=20
> > > it makes more sense.=20
> > >=20
> > > > Q. why isn't CVSMediator a singleton?=20
> > > > Q. can't remember. why would it be a Singleton?=20
> > > >=20
> > >=20
> > > to me it seems that you don't need more than one object in your syste=
m=20
> > > doing the mediation and executing the shell cvs commands. in the curr=
ent=20
> > > situation what happens is that we instantiate a mediator for every=20
> > > interaction (update, commit (wich updates at the end, so instantiates=
=20
> > > another mediator)) and it's up to garbage collection to clean it up.=
=20
> > > please correct me if i'm wrong.
> >=20
> >=20
> > if you see it like that, than probably you are right. like I said, I=20
> > don't remember, so we probably didn't even think of that... change it i=
f you=20
> > want to and see fit. we can always change it back if we need!
> >=20
>=20
> ok=20
>=20
> > roberto.=20
> > >=20
> > > nano
> >=20
> >=20
> > roberto=20
> >=20
>=20
> nano
>=20

roberto=20

>=20
> > > > On 6/23/05, Fernando Freire < th...@gm...> wrote:=20
> > > > >=20
> > > > > ent=E3o, as quest=F5es eram as seguintes:
> > > > > porque que no final do m=E9todo CVSMediator>>executeCVSCommit a=
=20
> > > gente=20
> > > > > manda a mensagem executeCVSUpdate?
> > > > > porque CVSMediator n=E3o =E9 singleton?
> > > > > acho que era isso por enquanto...
> > > > >=20
> > > > >=20
> > > > > -------------------------------------------------------=20
> > > > > SF.Net <http://SF.Net> email is sponsored by: Discover Easy Linux=
=20
> > > Migration Strategies=20
> > > > > from IBM. Find simple to follow Roadmaps, straightforward=20
> > > articles,=20
> > > > > informative Webcasts and more! Get everything you need to get up=
=20
> > > to=20
> > > > > speed, fast.
> > > > http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick=20
> > > <http://ads.osdn.com/?ad_idt77&alloc_id%16492&opclick>
> > > > > _______________________________________________=20
> > > > > Ignite-devel mailing list
> > > > > Ign...@li...=20
> > > > > https://lists.sourceforge.net/lists/listinfo/ignite-devel
> > > > >=20
> > > >=20
> > > >=20
> > >=20
> >=20
> >=20
>

[Ignite-devel] desenvolvimento

[Fernando F. <th...@gm...>]

yeah! real time...

On 6/23/05, Roberto Winter <rhw...@gm...> wrote:
>=20
> ok...
>=20
> On 6/23/05, Fernando Freire < th...@gm...> wrote:
> >=20
> > opa!
> >=20
> > On 6/23/05, Roberto Winter <rhw...@gm...> wrote:
> > > hello.=20
> > > shouldn't this be in english?=20
> > >=20
> >=20
> > whatever
> >=20
> > > Q. why do we executeCVSUpdate on the CVSMediator>>executeCVSCommit=20
> > message.=20
> > > A. if I remember correctly, the CVS system has some bizarre workings=
=20
> > that
> > > obliges you to do this. it has something to do with keeping the local=
=20
> > (or
> > > remote) "image" of the file structure updated... maybe you should=20
> > check CVS=20
> > > docs... apart from that I can't remeber very well.=20
> > >=20
> >=20
> > for all that i know of cvs (and i use quite a bit) it doesn't make sens=
e=20
> > to uptade after you commit, because you've just uploaded all your info,=
 so=20
> > even if you update, cvs isn't going to download anything, cause nothing=
's=20
> > changed. but our program is still going to file in everything, even tho=
ugh=20
> > the image is what it's supposed t be...
>=20
>=20
> well. i really don't remember. you should probably check the code for tha=
t=20
> message (the executeCVSUpdate) perhaps it has to do more with the things=
=20
> it does (smalltalk-wise) than CVS. maybe the Birmann can help us out here=
! I=20
> don't have the code here, so that does not make is easier!...
>=20

well, the only thing that executeCVSUpdate does, smalltalk wise, is file in=
=20
all the .st(s) in the rootDirectory, but that doesn't make much sense since=
=20
you're the one commiting your coding state, so your image should be all=20
good...
the Birmann is the one that told me to write to the list...., we're waiting=
=20
for his contact. good thing you wrote though, or i would be here sitting an=
d=20
waiting....

the only case in wich i think you have to update and commit is when there's=
=20
> > somebody coding at the same time as you, and commited something after y=
ou=20
> > updated, but i think we should resolve this by ending the beginCodingSe=
ssion=20
> > endCodingSession routines, and do the whole sync business, where you ca=
n=20
> > see exactly what you should commit and what you have to update. i think=
 it=20
> > makes more sense.=20
> >=20
> > > Q. why isn't CVSMediator a singleton?=20
> > > Q. can't remember. why would it be a Singleton?=20
> > >=20
> >=20
> > to me it seems that you don't need more than one object in your system=
=20
> > doing the mediation and executing the shell cvs commands. in the curren=
t=20
> > situation what happens is that we instantiate a mediator for every=20
> > interaction (update, commit (wich updates at the end, so instantiates=
=20
> > another mediator)) and it's up to garbage collection to clean it up.=20
> > please correct me if i'm wrong.
>=20
>=20
> if you see it like that, than probably you are right. like I said, I don'=
t=20
> remember, so we probably didn't even think of that... change it if you wa=
nt=20
> to and see fit. we can always change it back if we need!
>=20

ok=20

> roberto.=20
> >=20
> > nano
>=20
>=20
> roberto=20
>=20

nano=20

>=20
> > > On 6/23/05, Fernando Freire < th...@gm...> wrote:=20
> > > >=20
> > > > ent=E3o, as quest=F5es eram as seguintes:
> > > > porque que no final do m=E9todo CVSMediator>>executeCVSCommit a gen=
te=20
> > > > manda a mensagem executeCVSUpdate?
> > > > porque CVSMediator n=E3o =E9 singleton?
> > > > acho que era isso por enquanto...
> > > >=20
> > > >=20
> > > > -------------------------------------------------------=20
> > > > SF.Net <http://SF.Net> email is sponsored by: Discover Easy Linux=
=20
> > Migration Strategies=20
> > > > from IBM. Find simple to follow Roadmaps, straightforward articles,=
=20
> > > > informative Webcasts and more! Get everything you need to get up to=
=20
> > > > speed, fast.
> > > http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick=20
> > <http://ads.osdn.com/?ad_idt77&alloc_id%16492&opclick>
> > > > _______________________________________________=20
> > > > Ignite-devel mailing list
> > > > Ign...@li...=20
> > > > https://lists.sourceforge.net/lists/listinfo/ignite-devel
> > > >=20
> > >=20
> > >=20
> >=20
>=20
>

Fwd: [Ignite-devel] desenvolvimento

[Roberto W. <rhw...@gm...>]

ok...

On 6/23/05, Fernando Freire <th...@gm...> wrote:
>=20
> opa!
>=20
> On 6/23/05, Roberto Winter <rhw...@gm...> wrote:
> > hello.=20
> > shouldn't this be in english?=20
> >=20
>=20
> whatever
>=20
> > Q. why do we executeCVSUpdate on the CVSMediator>>executeCVSCommit=20
> message.=20
> > A. if I remember correctly, the CVS system has some bizarre workings=20
> that
> > obliges you to do this. it has something to do with keeping the local=
=20
> (or
> > remote) "image" of the file structure updated... maybe you should check=
=20
> CVS=20
> > docs... apart from that I can't remeber very well.=20
> >=20
>=20
> for all that i know of cvs (and i use quite a bit) it doesn't make sense=
=20
> to uptade after you commit, because you've just uploaded all your info, s=
o=20
> even if you update, cvs isn't going to download anything, cause nothing's=
=20
> changed. but our program is still going to file in everything, even thoug=
h=20
> the image is what it's supposed t be...


well. i really don't remember. you should probably check the code for that=
=20
message (the executeCVSUpdate) perhaps it has to do more with the things it=
=20
does (smalltalk-wise) than CVS. maybe the Birmann can help us out here! I=
=20
don't have the code here, so that does not make is easier!...

the only case in wich i think you have to update and commit is when there's=
=20
> somebody coding at the same time as you, and commited something after you=
=20
> updated, but i think we should resolve this by ending the beginCodingSess=
ion=20
> endCodingSession routines, and do the whole sync business, where you can=
=20
> see exactly what you should commit and what you have to update. i think i=
t=20
> makes more sense.=20
>=20
> > Q. why isn't CVSMediator a singleton?=20
> > Q. can't remember. why would it be a Singleton?=20
> >=20
>=20
> to me it seems that you don't need more than one object in your system=20
> doing the mediation and executing the shell cvs commands. in the current=
=20
> situation what happens is that we instantiate a mediator for every=20
> interaction (update, commit (wich updates at the end, so instantiates=20
> another mediator)) and it's up to garbage collection to clean it up.=20
> please correct me if i'm wrong.


if you see it like that, than probably you are right. like I said, I don't=
=20
remember, so we probably didn't even think of that... change it if you want=
=20
to and see fit. we can always change it back if we need!

> roberto.=20
>=20
> nano


roberto=20

>=20
> > On 6/23/05, Fernando Freire < th...@gm...> wrote:=20
> > >=20
> > > ent=E3o, as quest=F5es eram as seguintes:
> > > porque que no final do m=E9todo CVSMediator>>executeCVSCommit a gente=
=20
> > > manda a mensagem executeCVSUpdate?
> > > porque CVSMediator n=E3o =E9 singleton?
> > > acho que era isso por enquanto...
> > >=20
> > >=20
> > > -------------------------------------------------------=20
> > > SF.Net <http://SF.Net> email is sponsored by: Discover Easy Linux=20
> Migration Strategies=20
> > > from IBM. Find simple to follow Roadmaps, straightforward articles,=
=20
> > > informative Webcasts and more! Get everything you need to get up to=
=20
> > > speed, fast.
> > http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick=20
> <http://ads.osdn.com/?ad_idt77&alloc_id%16492&opclick>
> > > _______________________________________________=20
> > > Ignite-devel mailing list
> > > Ign...@li...=20
> > > https://lists.sourceforge.net/lists/listinfo/ignite-devel
> > >=20
> >=20
> >=20
>

[Ignite-devel] desenvolvimento

[Fernando F. <th...@gm...>]

opa!

On 6/23/05, Roberto Winter <rhw...@gm...> wrote:
> hello.=20
> shouldn't this be in english?
>=20

whatever

> Q. why do we executeCVSUpdate on the CVSMediator>>executeCVSCommit=20
message.=20
> A. if I remember correctly, the CVS system has some bizarre workings that
> obliges you to do this. it has something to do with keeping the local (or
> remote) "image" of the file structure updated... maybe you should check=
=20
CVS=20
> docs... apart from that I can't remeber very well.=20
>=20

for all that i know of cvs (and i use quite a bit) it doesn't make sense to=
=20
uptade after you commit, because you've just uploaded all your info, so eve=
n=20
if you update, cvs isn't going to download anything, cause nothing's=20
changed. but our program is still going to file in everything, even though=
=20
the image is what it's supposed t be...
the only case in wich i think you have to update and commit is when there's=
=20
somebody coding at the same time as you, and commited something after you=
=20
updated, but i think we should resolve this by ending the beginCodingSessio=
n=20
endCodingSession routines, and do the whole sync business, where you can se=
e=20
exactly what you should commit and what you have to update. i think it make=
s=20
more sense.=20

> Q. why isn't CVSMediator a singleton?=20
> Q. can't remember. why would it be a Singleton?=20
>=20

to me it seems that you don't need more than one object in your system doin=
g=20
the mediation and executing the shell cvs commands. in the current situatio=
n=20
what happens is that we instantiate a mediator for every interaction=20
(update, commit (wich updates at the end, so instantiates another mediator)=
)=20
and it's up to garbage collection to clean it up.=20
please correct me if i'm wrong.

> roberto.=20

nano

>=20
> On 6/23/05, Fernando Freire <th...@gm...> wrote:=20
> >=20
> > ent=E3o, as quest=F5es eram as seguintes:
> > porque que no final do m=E9todo CVSMediator>>executeCVSCommit a gente=
=20
> > manda a mensagem executeCVSUpdate?
> > porque CVSMediator n=E3o =E9 singleton?
> > acho que era isso por enquanto...
> >=20
> >=20
> > -------------------------------------------------------=20
> > SF.Net <http://SF.Net> email is sponsored by: Discover Easy Linux=20
Migration Strategies=20
> > from IBM. Find simple to follow Roadmaps, straightforward articles,
> > informative Webcasts and more! Get everything you need to get up to=20
> > speed, fast.
> http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick<http://ads.osdn.com/=
?ad_idt77&alloc_id%16492&opclick>
> > _______________________________________________=20
> > Ignite-devel mailing list
> > Ign...@li...=20
> > https://lists.sourceforge.net/lists/listinfo/ignite-devel
> >=20
>=20
>

Re: [Ignite-devel] desenvolvimento

[Roberto W. <rhw...@gm...>]

hello.
shouldn't this be in english?
 Q. why do we *executeCVSUpdate* on the *CVSMediator>>executeCVSCommit*mess=
age.
A. if I remember correctly, the CVS system has some bizarre workings that=
=20
obliges you to do this. it has something to do with keeping the local (or=
=20
remote) "image" of the file structure updated... maybe you should check CVS=
=20
docs... apart from that I can't remeber very well.
 Q. why isn't CVSMediator a singleton?
Q. can't remember. why would it be a Singleton?
 roberto.
 On 6/23/05, Fernando Freire <th...@gm...> wrote:=20
>=20
> ent=E3o, as quest=F5es eram as seguintes:
> porque que no final do m=E9todo CVSMediator>>executeCVSCommit a gente
> manda a mensagem executeCVSUpdate?
> porque CVSMediator n=E3o =E9 singleton?
> acho que era isso por enquanto...
>=20
>=20
> -------------------------------------------------------
> SF.Net <http://SF.Net> email is sponsored by: Discover Easy Linux=20
> Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id=16492&opclick
> _______________________________________________
> Ignite-devel mailing list
> Ign...@li...
> https://lists.sourceforge.net/lists/listinfo/ignite-devel
>

[Ignite-devel] desenvolvimento

[Fernando F. <th...@gm...>]

ent=E3o, as quest=F5es eram as seguintes:
porque que no final do m=E9todo CVSMediator>>executeCVSCommit a gente
manda a mensagem executeCVSUpdate?
porque CVSMediator n=E3o =E9 singleton?
acho que era isso por enquanto...

[Ignite-devel] Task Manager no longer used

[Ricardo B. <ric...@gm...>]

The interface is not very good... takes to much time to
read/add/update, etc....  (not to mention the damn emails)

besides we have the ToDoBrowser INSIDE VW.... talk about efficiency!!!

[Ignite-devel] test

[Ricardo B. <ric...@gm...>]

just testing the list