[extropy-chat] when will computers improve?

[Eugen Leitl]

On Sat, Dec 20, 2003 at 12:55:19PM -0700, Alan Eliasen wrote:
> 
>    Actually, Cygwin is on the APE.  But it's one of the larger causes of

Good! It's nice to pull the latest version, though, as it is in more or less
vigorous development.

> my pain, when you look at its portability issues.  :)  One of the
> biggest problems is the way they do symbolic links--they actually set
> the "system" file attribute on each file that's a symbolic link and the
> operating system has to see that and act appropriately.  Admittedly,
> it's half a Windows problem, but their solution doesn't work for anyone

It is a Windows problem. Cygwin can't implement real links on a file system
that doesn't support them.

> but themselves.

I agree, cygwin is an ugly hack. However, the host is an even uglier hack,
and you're just glad to find anything to make that pain go away.
 
>    Needless to say, there ain't no "system" bit on a CD.  To make this
> run off a read-only medium, I have to selectively fix thousands of
> symbolic links by actually copying the file (if I fix too many, it's
> just wasted space as the same program might be called by 10 names.)

I have no idea how Knoppix feels on a FAT32 system.
 
>    The fact that the cygwin1.dll file isn't well-versioned and
> poorly-licensed makes it difficult to coexist with multiple versions of
> cygwin or redistribute anything linked with cygwin.

Yes, a standard case of DLL hell. I deal with it (because people pay me for
this pain, but not nearly enough) by building stuff on a consistent box, and
just package cygwin1.dll with the system.
 
>    It also stinks that mountpoints are crammed into the registry (why?)
> and aren't on a per-process basis; it makes it impossible to have
> different processes using different versions of cygwin or even different
> mount points.  It makes it hard to have two versions of the APE running
> well at the same time.  (Actually, everything works *but* cygwin.)

Aargh! Good thing I don't cygwin very often.
 
>    Cygwin corrupts environment variables, too, (notably any paths) so
> you can't switch into a bash shell and switch back out to cmd.exe
> without that whole session geting hosed.

Aargh^2!
 
>    But I'll admit, cygwin is outstanding when you're trying to compile
> and run a UN*X-only application under windows.  I'm sure you've seen all
> of these problems.

Thankfully, not much of these. I tend to avoid them by avoiding to deal with
Redmondware. Thankfully, this is getting easier and easier these days. Though
some customers have still to see the light.
 
> > I'd rather prefer support offered by the Debian team, than you, no offence.
> > And I do carry a very useful system around everywhere I go: my iBook.
> 
>    Good, because I don't offer support, other than the fact that I'm a
> nice guy and I'll answer questions if I can.  :)  I prefer the Debian
> team too.  They're smarter than me.

It is a classical tradeoff. Lemmingware sucks, but no man is an island. I
tend to stick to the systems I can handle. Redhat/Fedora is a tolerable
tradeoff between support and flakiness on the desktop, Debian does it for the
server. I haven't been touching *BSD because I'm trying to get away from IT
altogether, so it doesn't make sense spending time learning something I won't
have to deal with anyway.
 
>    I certainly prefer Linux to running Windows; it's just that I also
> choose to ship product for Windows, and some of the tools I have only
> run under Windows, (e.g. packagers for some handheld devices) and most

Yeah, sometimes I have to shovel manure, too.

> of my work has made me have to touch Windows at least half the time.
> And occasionally I have to drive around and steal cars and shoot people
> in Grand Theft Auto III.  :)

Luckily, I don't game. There's simply no option to Redmond there.

-- Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20031220/2a8fa8df/attachment.bin>