Linux-Advocacy Digest #372, Volume #35 Mon, 18 Jun 01 18:13:07 EDT
Contents:
Re: Why homosexuals are no threat to heterosexuals ("Aaron R. Kulkis")
Re: Linux inheriting "DLL Hell" (Craig Kelley)
Re: PC power switch wont shut down Windows (Chris Street)
Re: Linux inheriting "DLL Hell" (Craig Kelley)
Re: Linux inheriting "DLL Hell" (Craig Kelley)
Re: The beginning of the end for microsoft ("Bill Todd")
Re: Microsoft and open source ("Erik Funkenbusch")
Re: So how many applications can Windows run on the IA-64? ("Erik Funkenbusch")
Re: Linux inheriting "DLL Hell" ("Ayende Rahien")
Re: Linux inheriting "DLL Hell" ("Ayende Rahien")
Re: More microsoft innovation ("Ayende Rahien")
Re: More microsoft innovation ("Ayende Rahien")
Re: Linux inheriting "DLL Hell" ("Ayende Rahien")
----------------------------------------------------------------------------
From: "Aaron R. Kulkis" <[EMAIL PROTECTED]>
Crossposted-To: soc.men,soc.singles,alt.fan.rush-limbaugh
Subject: Re: Why homosexuals are no threat to heterosexuals
Date: Mon, 18 Jun 2001 17:57:42 -0400
drsquare wrote:
>
> On Mon, 18 Jun 2001 10:14:04 -0500, in comp.os.linux.advocacy,
> (SSunbird <[EMAIL PROTECTED]>) wrote:
>
> >Rich Soyack wrote:
> >
> >> Really? Methods used to study the transmission of other STDs don't work
> >> with AIDS? Why is that?
> >
> >what methods are those?
> >
> >> Which acts carry with it the most chances of getting AIDS?
> >
> >acts of stupidity
>
> Well, with statements like that, you must be an absolute AIDS guru.
As having anal sex with a homosexual man:
a) smart
B) STUPID
Is doing illicit intravenous drugs:
a) smart
B) STUPID
no more questions, your honor
--
Aaron R. Kulkis
Unix Systems Engineer
DNRC Minister of all I survey
ICQ # 3056642
L: This seems to have reduced my spam. Maybe if everyone does it we
can defeat the email search bots. [EMAIL PROTECTED] [EMAIL PROTECTED]
[EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
[EMAIL PROTECTED]
K: Truth in advertising:
Left Wing Extremists Charles Schumer and Donna Shalala,
Black Seperatist Anti-Semite Louis Farrakhan,
Special Interest Sierra Club,
Anarchist Members of the ACLU
Left Wing Corporate Extremist Ted Turner
The Drunken Woman Killer Ted Kennedy
Grass Roots Pro-Gun movement,
J: Other knee_jerk reactionaries: billh, david casey, redc1c4,
The retarded sisters: Raunchy (rauni) and Anencephielle (Enielle),
also known as old hags who've hit the wall....
I: Loren Petrich's 2-week stubborn refusal to respond to the
challenge to describe even one philosophical difference
between himself and the communists demonstrates that, in fact,
Loren Petrich is a COMMUNIST ***hole
H: "Having found not one single carbon monoxide leak on the entire
premises, it is my belief, and Willard concurs, that the reason
you folks feel listless and disoriented is simply because
you are lazy, stupid people"
G: Knackos...you're a retard.
F: Unit_4's "Kook hunt" reminds me of "Jimmy Baker's" harangues against
adultery while concurrently committing adultery with Tammy Hahn.
E: Jet is not worthy of the time to compose a response until
her behavior improves.
D: Jet Silverman now follows me from newgroup to newsgroup
...despite (C) above.
C: Jet Silverman claims to have killfiled me.
B: Jet Silverman plays the fool and spews out nonsense as a
method of sidetracking discussions which are headed in a
direction that she doesn't like.
A: The wise man is mocked by fools.
------------------------------
From: Craig Kelley <[EMAIL PROTECTED]>
Subject: Re: Linux inheriting "DLL Hell"
Date: 18 Jun 2001 15:58:12 -0600
"Seán Ó Donnchadha" <[EMAIL PROTECTED]> writes:
> "Peter Köhlmann" <[EMAIL PROTECTED]>:
> > >
> > > Not only do you not know anything about Windows (see above), but you're
> > > also sadly mistaken to think that versioning solves the problem. It
> > > doesn't. It doesn't even come close. The simple example is as follows.
> You
> > > install libfoo-1.1. You install appfoo, which requires libfoo-1.1 and
> > > inadvertently relies on a bug therein. You then get appbar, which
> requires
> > > libfoo-1.2. You install libfoo-1.2, which fixes the aforementioned bug.
> > > You install appbar and everything seems fine. Then you run appfoo, and
> it
> > > breaks. Don't tell me it's appfoo's author's fault, because blame is
> > > irrelevant. This kind of shit happens in the real world. The point is
> that
> > > versioning is not a silver bullet by any stretch.
> >
> > Why does appfoo break?
>
> Appfoo breaks because it relied on a bug in libfoo-1.1 which was fixed in
> libfoo-1.2. And it doesn't even have to be a bug; it could just be some
> underspecified behavior.
>
> > Waht happened to libfoo-1.1 ?
>
> It's still there, but no longer used, since both programs load libfoo
> through the libfoo.so.1 symbolic link, which now points at libfoo.so.1.2.
>
> > How come this problem you
> > are describing here does *not* happen that way on linux?
>
> Your question is based on an incorrect assumption. You obviously don't
> understand the problem...
Rather, in your specific example it breaks; but there's nothing
stopping appfoo from specifically linking to libfoo-1.1 explicitly.
This isn't generally done, though.
--
It won't be long before the CPU is a card in a slot on your ATX videoboard
Craig Kelley -- [EMAIL PROTECTED]
http://www.isu.edu/~kellcrai finger [EMAIL PROTECTED] for PGP block
------------------------------
From: [EMAIL PROTECTED] (Chris Street)
Crossposted-To: comp.os.ms-windows.nt.advocacy,alt.destroy.microsoft
Subject: Re: PC power switch wont shut down Windows
Reply-To: [EMAIL PROTECTED]
Date: Mon, 18 Jun 2001 21:53:41 GMT
On 18 Jun 2001 20:39:30 GMT, [EMAIL PROTECTED] (Hans) wrote:
>In article <[EMAIL PROTECTED]>, Chris Street wrote:
>>My win2000 SQL and Exchange server, running on a lowly AMD K62/350 VIA
>>board shows an up time of 72 days at the moment.......
>
>did MS finally have the courage to add an 'uptime' function?
>
>Hans
It's in Performance monitor - System - System Up Time. Certainly in my
W2k and NT4 servers. I'm fairly sure it goes all the way back to NT
3.1 but having decommed my 3.51 server I cannot tell.
79.84% of all statistics are made up on the spot.
The other 42% are made up later on.
In Warwick - looking at flat fields and that includes the castle.
------------------------------
From: Craig Kelley <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux inheriting "DLL Hell"
Date: 18 Jun 2001 15:59:56 -0600
"Seán Ó Donnchadha" <[EMAIL PROTECTED]> writes:
> "drsquare" <[EMAIL PROTECTED]>:
> > >>
> > >>
> > >> Your point?
> > >
> > >My point is that Unix/Linux is just as susceptible to shared library
> > >conflicts as Windows.
> >
> > Stop lying.
> >
>
> So were the folks on lwn.net lying when they wrote about GnuCash's DLL Hell?
They said DLL Hell to mean 'lots of libraries', which it doesn't. DLL
Hell happened a lot under Windows in a single library (MFC42.DLL comes
to mind; but I haven't really used Windows in years, so that may date
me).
--
It won't be long before the CPU is a card in a slot on your ATX videoboard
Craig Kelley -- [EMAIL PROTECTED]
http://www.isu.edu/~kellcrai finger [EMAIL PROTECTED] for PGP block
------------------------------
From: Craig Kelley <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux inheriting "DLL Hell"
Date: 18 Jun 2001 16:04:21 -0600
[EMAIL PROTECTED] (The Ghost In The Machine) writes:
> In comp.os.linux.advocacy, Bob Hauck
> <[EMAIL PROTECTED]>
> wrote
> on Sun, 17 Jun 2001 13:44:06 GMT
> <[EMAIL PROTECTED]>:
> >On Sun, 17 Jun 2001 00:50:34 -0700, GreyCloud <[EMAIL PROTECTED]> wrote:
> >
> >> Got a question then: Under Solaris 8 I have gcc and the libs. If the
> >> lib name is xyz.so.n.n is that a static or shared lib?
> >
> >Shared. The convention is that .so is shared, .a is static.
>
> Pedant point. *.a is an archive library whose modules (.o files)
> are linked statically into an executable; the whole .a file may
> not be needed. ('man ar' for details on 'ar', a command which
> helps manage these archives; note that these archives are not
> limited to .o files, but are most often used therewith.)
>
> Also, HP/UX uses .sl. Just to confuse us engineering types. :-)
MacOS X uses .dylib :o)
--
It won't be long before the CPU is a card in a slot on your ATX videoboard
Craig Kelley -- [EMAIL PROTECTED]
http://www.isu.edu/~kellcrai finger [EMAIL PROTECTED] for PGP block
------------------------------
From: "Bill Todd" <[EMAIL PROTECTED]>
Subject: Re: The beginning of the end for microsoft
Date: Mon, 18 Jun 2001 18:07:25 -0400
"Peter Köhlmann" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Bill Todd wrote:
>
> >
> > "Peter Seebach" <[EMAIL PROTECTED]> wrote in message
> > news:3b2e4b03$0$325$[EMAIL PROTECTED]...
> >> In article <9gc8qb$lbt$[EMAIL PROTECTED]>, Bill Todd
<[EMAIL PROTECTED]>
> > wrote:
> >> >Windows actually does a semi-decent job of boot-time configuration, to
> > the
> >> >point that an image of your old system will quite possibly boot on a
> >> >different system - after which you can perform the upgrade
installation.
> >>
> >> I have heard many, many, horror stories about this not working. I've
> > heard
> >> very few stories of Mac disks not booting on different Macs (if they
were
> >> at all supported), and none at all about Unixes (once again, if the
> > hardware
> >> was anything like supported). By contrast, people in a nearby thread
> >> were commenting that it's fairly typical to have to reinstall Windows
to
> >> change SCSI controllers.
> >
> > While I'm not a great fan of Windows, the above has a definite odor of
FUD
> > to it. Why not let the people with *direct* experience comment, so that
> > one can properly ascertain what real problems may of may not have
> > occurred, and why?
> >
> You are telling BS.
Perhaps your command of English (or rather the lack of it) is the cause of
your misunderstanding. My comment had nothing to do with whether or not
such things occurred, only with the desirability of having first-hand rather
than hearsay (and unattributed hearsay at that) evidence of them.
> I have experienced exactly this 3 times. Each time Windows had to be
> reinstalled completely from scratch, because it simply did not work.
Fine: *that's* first-hand evidence.
> And just today a friend of mine called. He had switched the MoBo from
> a PentiumII 350 MHz to an Athlon 800MHz.
> He reinstalled from scratch after 3 hours of trying to get that damn thing
> to work. Even the CD-Rom was no longer present. No network card.
> Kind of hard to put in new drivers then.
Again, your understanding of the discussion seems inadequate. My suggestion
was that Windows stood a good chance of *booting up minimally* on a
configuration other than that on which it had originally been installed, to
the point where an upgrade installation could then be performed to get
everything working reasonably well, not that one could *ever* tweak an
existing system image to work well on new hardware *without* performing such
an upgrade installation.
It does surprise me a bit that your friend couldn't get minimal CD-ROM
function after booting, but then again since Athlons may not even have
existed when he got his P-II 350 system (and the Windows software on it) the
fact that that software may have had motherboard problems isn't all that
hard to imagine.
> Those MS-developers who did that job deserve to be grilled alive after
> beeing flocked. They proved to be even more braindead than Windows
> already is.
>
> Peter
>
> --
> Microsoft is not the the answer.
> Microsoft is the question.
> The answer is NO
Clearly the sentiment of a truly objective observer. It's people like you
who cause those of us who don't particularly like Microsoft's software or a
lot of its business tactics to wind up defending it out of a sense of
fairness.
- bill
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.unix.bsd.freebsd.misc
Subject: Re: Microsoft and open source
Date: Mon, 18 Jun 2001 16:51:56 -0500
"." <[EMAIL PROTECTED]> wrote in message
news:9glngt$rfl$[EMAIL PROTECTED]...
> In comp.os.linux.advocacy Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
> > "Jason Bowen" <[EMAIL PROTECTED]> wrote in message
> > news:9gl6ci$fab$[EMAIL PROTECTED]...
> >> http://public.wsj.com/news/hmc/sb992819157437237260.htm
>
> > The author seems only slightly aware that there is a major difference
> > between BSD licensing and GPL licensing (in fact, the word GPL doesn't
even
> > exist in the article). The author incorrectly says that MS has never
made a
> > distinction between BSD and other Open Source licensing, which is false.
MS
> > has many times lauded the BSD license.
>
> Now eric funkybreath knows more than the wall street journal.
It's not the WSJ, it's an author published by the journal.
> > While it's true that Craig Mundie was a bit vague in his statements, it
> > seems pretty clear he was only talking about GPL'd code when he said
"Open
> > Source" because he specifically referred to the Licensing used in Linux.
I
> > think the FreeBSD camp is being a little sensative about the matter.
>
> I think that you would oppose god himself if he said something bad about
> microsoft.
If what were said was true, I would have no problem with it.
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: So how many applications can Windows run on the IA-64?
Date: Mon, 18 Jun 2001 16:54:08 -0500
"Brian Langenberger" <[EMAIL PROTECTED]> wrote in message
news:9glils$dqh$[EMAIL PROTECTED]...
> Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
>
> <more snip>
>
> :> I can go to a software store and find 1500 MS-Windows apps that I think
> :> are pretty worthless.
> :>
> :> But you'll have a hard time convincing me that "less apps == better".
>
> : You are basing your decision of "worthless" as "value to you". My
decision
> : of worthless is "value to more than a few people".
>
> What, exactly, would any particular distro gain by shipping with *less*
> applications? Isn't that a bit like a software store advertising:
> "We only sell the top 20 Windows apps and pass the lack of choices on
> to you!" Since the vast bulk of a Linux distro's apps cost the same
> (0$) the packager has little to gain by shipping a "Best-Of" distro
> and risk alienating a set of users who rather like LittleKnonwnApp-1.0
Of course, however using the sheer number of apps as some kind of major
advantage is misleading. Which has more value to more people? 1000
free/shareware programs that nobody would buy, or a dozen apps of the Office
calibre?
> And besides, if it wasn't for people trying out LittleKnownApp-1.0
> and liking it, it would never evolve into WellKnownApp-2.0
I guess that's why so many of those little known apps stay little known
apps.
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux inheriting "DLL Hell"
Date: Tue, 19 Jun 2001 00:38:51 +0200
"Craig Kelley" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Oh? Just like Adaptec Easy-CD Creator 4 users were SOL under Windows
> 2000 perhaps? Where's the outrage? People had to go and buy version
> 5.
EZCD is a POS of an application, but it works on 2K.
Can you believe that it has a major (as in "you aren't leaving, sleeping,
eating, *OR* breathing until you get it fixed") bug that not only never got
fixed, but also *got carried on to version 5*?
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux inheriting "DLL Hell"
Date: Tue, 19 Jun 2001 00:42:43 +0200
"Stephen S. Edwards II" <[EMAIL PROTECTED]> wrote in message
news:bzoX6.80$[EMAIL PROTECTED]...
>
> If a current system shared lib gets replaced by an older
> version, and breaks several applications, it's possible
> to fix the problem simply by installing the latest
> Service Pack to make sure everything is up to date, or
> more conveniently, simply extract the contents of the
> Service Pack, and copy the libs over manually.
That is a bad way to do this, actually.
What you should do is to have WFP active, and create a private DLLs for the
applications that need other versions that what you've as shared.
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy
Subject: Re: More microsoft innovation
Date: Tue, 19 Jun 2001 00:59:54 +0200
"Peter Hayes" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Mon, 18 Jun 2001 15:30:45 GMT, Macman <[EMAIL PROTECTED]> wrote:
>
> > In article <[EMAIL PROTECTED]>,
> > Sandman <[EMAIL PROTECTED]> wrote:
>
> > > Nowhere, these are all my logic assumptions. Perhaps I should have
stated
> > > that more clearly. I know almost nothing about the Smart Tags
technology,
> > > but I assume, and you have to agree with me I think, that my IE needs
to
> > > connect to a MS server to get the information the Smart Tags provide,
> > > right? It's not logical to assume that all the information is already
in
> > > the browser, right?
> >
> > Actually, you _seem_ to be wrong. As far as I've heard, the Smart Tags
> > are there when the browser has shipped. I haven't seen any reports that
> > the browser updates its smart tags.
>
> Surely the SmartTags would have to be updated on a regular basis,
otherwise
> they pretty soon become out of date. Plus, of course, Microsoft will be
> marketing SmartTags to all and sundry; it could prove quite a money
spinner
> as corporations try to out SmartTag each other.
>
> So they'd need a mechanism for getting them to your browser.
They are unlikely to need to be updated on a regular basis.
Currently, the MS implementation is for a list of Companies & Universities,
when it sees a company, (a relatively short list, so far) it offers to send
you to a site where you can watch info about this company (stocks, news,
homepage, etc).
When it sees a university, it sends you to different parts of the university
homepage, depending on where you want to go. (allimony, homepage, stuff like
that).
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy
Subject: Re: More microsoft innovation
Date: Tue, 19 Jun 2001 01:01:25 +0200
"Daniel Johnson" <[EMAIL PROTECTED]> wrote in message
news:TPoX6.5672$[EMAIL PROTECTED]...
> "Sandman" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > In article <[EMAIL PROTECTED]>, macman
> > <[EMAIL PROTECTED]> wrote:
> [snip]
> > > That's EXACTLY the kind of thing I consider evil about Smart tags. Not
> > > only do they deface the web site, it could entirely change the overall
> > > meaning.
> >
> > And furthermore, as I stated in another reply in this thread, Smart Tags
> > makes my IE do a connection to a MS server for -every- page I surf (or
so
> > I've understood it) to collect data about the SmartTags on the page I am
> > surfing on. I imagined a connection like that could look like
> > "http://www.mysite.com/index.html Chrysler Apple Donuts Linux" that gets
> > sent to and MS server. Now, by that request, amongst the others my IE
has
> > done, MS has actually a total record of what sites I have surfed and
what
> > they contained.
>
> What's your source for this?
>
> No description of SmartTags I've yet seen
> has implied it. It's the sort of thing that MS bashers
> will believe if you tell it to them, but I'd like
> to know if you have any reason to believe it
> yourself.
Um, I hate to be on his side at this point of the arguement, but he is at
least partially correct.
It is certainly possible to make SmartTags fetch information or report
across the web.
MS' implementation doesn't do it, just to get the record clean. But it's
*possible* to do this, as you can link SmartTags to COM objects, (persumebly
to run complex queries).
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux inheriting "DLL Hell"
Date: Tue, 19 Jun 2001 01:08:51 +0200
"The Ghost In The Machine" <[EMAIL PROTECTED]> wrote in
message news:[EMAIL PROTECTED]...
> In comp.os.linux.advocacy, Ayende Rahien
> <[EMAIL PROTECTED]>
> wrote
> on Sat, 16 Jun 2001 23:15:09 +0200
> <9ggfsp$ib$[EMAIL PROTECTED]>:
> >
> >"The Ghost In The Machine" <[EMAIL PROTECTED]> wrote in
> >message news:[EMAIL PROTECTED]...
> >>
> >> After all, Linux is essentially a large, distributed development
effort;
> >> not everyone is going to have the same version of everything.
> >>
> >> The only way to solve that problem is to rebuild everything from
> >> scratch -- an idea which I for one would love to do someday on
> >> my own personal system -- so that everything refers to the same version
> >> of everything else, and everything is consistent. More or less.
>
> I should point out at this point that I forgot to add "IMO" in there. :-)
>
> >
> >One way to do it is to have a directory where all the shared libraries
are
> >stored, which no one has write/delete access to.
> >The only way to add/remove libraries is via system API, which would check
> >for versioning.
>
> And the system API does precisely what checking?
Versioning checking. The idea is to make what is now suggestion into
mandatory.
> I'll admit it's
> a good idea in principle (one shouldn't remove files that are actually
> in use), but I'm curious as to the details, as it can get very
> messy if the system reboots and the file has been deleted by a simple
'rm'.
That is why I said that no one should've write/delete access to it.
I was thinking about NT's model here, where you can set it up so only the
system account has write/delete access, I don't think that Linux can do it.
(root can do everything, etc).
> (Remember that Unix and Linux only reclaim disk space if no one has the
file
> actually open. New apps will refer to it by name -- a problem since
> the name's already been removed.)
You mean that you can actually delete an open file?
Why?
> >It would require package management, I think, so when you install
> >application foo, which depend on libarary bar3.1.3, the package
management
> >ask the system whatever it has this library's version, and if not,
install
> >it.
>
> Bear in mind that the package will depend on *other packages*, not
> other libraries (which are within a package of their own) [*]. With
> intelligent package design, this isn't a problem -- but if someone
> starts to get stupid, things could get interesting. :-) For example,
> a package defining two libraries A.so and B.so may be upgraded, and
> other packages may only really require A.so, not B.so, so the old
> version of B.so can be reclaimed, except that it's part of the first
> package so it cannot be removed by the package management system (since
> it doesn't know library dependencies, only package ones). This
> particular flaw, fortunately, only wastes disk space... :-)
>
> However, with package management, it does make things a little easier.
Okay, here is how it works on Windows:
There is a place in the registry (which I can't recall right now), where you
register your DLLs and set a counter to one.
If two applications need the same DLL, the second application increment a
counter for this file.
That way, if you uninstall the first application, it knows not to delete
this file while uninstalling.
Only when the counter reach zero, you are allowed to delete it.
The idea is of something like this:
RegisterLib(char * path_to_lib);
UnRegisterLib(char * path_to_lib);
LoadLib(char * lib_name,VER * lib_version);
On install, the installer would call RegisterLib, and the system would check
the versions, if it already have this dll's version, it increment the
counter and does nothing.
If not, it copies it, and set a new counter.
UnRegisterLib decrement the counter and delete it if no one elses is using
it.
LoadLib lets you request the library by version, as well as by name.
Maybe using GUID will be better, though. And some way (non brute force,
please) to track which application uses which library.
> Of course, one could get crude about it and simply scan for all
> thingies that look like executables and shared system libraries
> ('man file'), then run 'ldd' on each of them, and gather the
> resulting list of dependencies into one useful list. Any DLL
> not on that list can be safely removed -- in theory, anyway.
>
> Unix (and Linux) allows for that sort of bodgery routinely; I can't
> say it's the most elegant solution, though.
I believe you can do this on Windows as well, there is a dependency part on
the EXE that the system reads before it execute it, IIRC.
> >When you uninstall an application, the package management tells the
system
> >that it no longer need this version of the application
> >The system keeps tracks on how many applications uses a library & library
> >version, when no application uses this library, it's deleted.
>
> The only drawback there of course is that: suppose an application
> developer wants that library? Hopefully there's a more or less
> elegant method of indicating whether someone's explicitly installed
> that package, or not. (Or maybe a simple flag will do; that flag would
> indicate whether that package is eligible for removal/purge should
> no other system package be using it.)
Well, he could tell the package mangement to add it.
That should solve this problem, and even if other software is added/removed
from the system, the library would remain until he remember to unregister
his need for this library.
> [*] it is possible that there could be library dependencies, however;
> this occasionally becomes an issue with non-free products or
> products limited by now-obsolete crypto restrictions.
In this case, I think that what should happen is that there would have to be
someway to track dependecies.
IE, a big table where you look at a library and say, here thos
application/libraries depend on this library, and this library depend on the
following applications/libraries.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
