Re: OSUser status and improvements

I agree - the problem I always encounter is that as soon as you start
changing it a little, you end up knee deep in shit, mired in a swamp of
ugly, tightly coupled code. (How's that for a bileous sentence?)

IMHO the key here is _very small steps_.

First I'd advise:
- getting rid of that horrendous EntityAccessor class
- make User and Group into interfaces with separate implementations

That solves a lot of the problems with only minor changes, from there we ca=
n
look at step 2.

M

On 13/2/04 4:58 AM, "Hani Suleiman" ([email protected]) penned the words:

> Hey now!
>=20
> Throughout the years, we've had at least 4-5 different attempts at this
> rewrite. Someone gets excited, says they'll do it, and then realises
> that it isn't quite as trivial as it seems.
>=20
> I'm all for ANY rewrite (oh, I and I try to get CVS patches in when
> sent, if I'm late or ignore one, yell at me, don't just give up!).
> However, the only two requirements I (and I suspect anyone else) using
> osuser have are:
>=20
> - API should be at least somewhat similar, or be able to migrate to
> with a search/replace
>=20
> - Providers idea should remain. Exact API can change a bit, but again,
> clear migration guide if so, with simple/obvious API changes. It'd be
> stupid to throw away all the existing providers, given that a not
> insignificant amount of effort has been put into them (some more than
> others).
>=20
> - Should not try to be 'cool' by using a bunch of technology for the
> sake of it. Sure, using pico and suchlike might save you code, but it
> all has to be optional. osuser jar should not require anything other
> than oscore and propertyset.
>=20
> Other than that, knock yourself out.
>=20
> On Feb 12, 2004, at 11:42 AM, Konstantin Priblouda wrote:
>=20
>>=20
>> --- David Smiley <[email protected]> wrote:
>>> Hi Scott.
>>>=20
>>> Last year I developed what I was calling OSUser 2
>>> but unfortunately, the
>>> OSUser community rejected it.  The main reason was
>>> that any providers
>>> would have to be re-written because my interfaces
>>> were different.  The
>>> client side use was almost the same though.  I
>>> didn't consider that
>>> deficiency to be a deal killer because OSUser isn't
>>> even 1.0, and I felt
>>> the current incarnation of OSUser is too broken
>>> underneath to work
>>> with.  My opinion was the minority.  If you want to
>>> enhance OSUser as I
>>> have done, you will come up against resistance too.
>>> I have since given up.
>>=20
>>=20
>> I'm personally unhappy with osuser too. Especially
>> with current state of project management.
>>=20
>> I patched version for me so it works properly, but
>> submitting patches did not worked.... And now it's
>> unclear where is actual CVS repository, who is in
>> charge etc...
>>=20
>> Anyway, I started to play with picocontainer, and
>> incorporate it into my application.
>>=20
>> It looks that using pico container it's possible to
>> invalidate major advantage of osuser - configuration
>> system ( I'm killing my configuration code at amazing
>> rates now. It's dissapearing without trace, and
>> everything still works. )
>>=20
>> So, If your osuser2 can be refactored to be IoC model
>> 3
>> compliant it would be really cool.
>>=20
>> And picoable user/group/property/entitlement
>> management
>> would be really cool addition for pico/nano community.
>>=20
>>=20
>> I can not promiss you 100% of my development time (
>> I' project leader, and our deadlines are nearing...
>> and passing by with a wooshing sound... )  - but
>> you can cound on sugnificant help from me.
>>=20
>> If you like we can move this disscussion off the list,
>> to avoid embarassing OSUser people.
>>=20
>> regards,
>>=20
>> =3D=3D=3D=3D=3D
>> ----[ Konstantin Pribluda ( ko5tik ) ]----------------
>> Zu Verst=E4rkung meines Teams suche ich ab Sofort einen
>> Softwareentwickler[In] f=FCr die Festanstellung.
>> Arbeitsort: Mainz
>> Skills:  Programieren, Kentnisse in OpenSource-Bereich
>> ----[ http://www.pribluda.de ]------------------------
>>=20
>> __________________________________
>> Do you Yahoo!?
>> Yahoo! Finance: Get your refund fast by filing online.
>> http://taxes.yahoo.com/filing.html
>>=20
>>=20
>> -------------------------------------------------------
>> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
>> Build and deploy apps & Web services for Linux with
>> a free DVD software kit from IBM. Click Now!
>> http://ads.osdn.com/?ad_id=3D1356&alloc_id=3D3438&op=3Dclick
>> _______________________________________________
>> Opensymphony-developers mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/opensymphony-developers
>>=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id438&op=CCk
> _______________________________________________
> Opensymphony-developers mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/opensymphony-developers



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click