Somehow, I always find it amusing how many FOSS programs assume Windows is 
basically how it was in Win98 and completely fail to do the right thing.

For example, languages in which the call to "open" doesn't let you specify 
how to lock the file, calls to get the owner of the file fail on Windows as 
if there isn't any such concept, the ubiquitous assumption that Windows 
doesn't have hard or soft links, and stuff like
"""
On Windows, HOME and USERPROFILE will be used if set, otherwise a 
combination of HOMEPATH and HOMEDRIVE will be used. An initial ~user is 
handled by stripping the last directory component from the created user path 
derived above.
"""
which in other words means "we haven't bothered to find the API call that 
tells you where a particular user's home directory is in Windows, so we'll 
assume they're all subdirectories of the same parent directory, even tho 
this isn't true in UNIX either."

And some people wonder why FOSS feels klunky on Windows.

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Somehow, I always find it amusing how many FOSS programs assume Windows is 
> basically how it was in Win98 and completely fail to do the right thing.

  Well, you can always demand your money back.

-- 
                                                          - Warp

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> which in other words means "we haven't bothered to find the API call that
> tells you where a particular user's home directory is in Windows, so we'll
> assume they're all subdirectories of the same parent directory, even tho
> this isn't true in UNIX either."

My guess is that that won't change while the portable C stdlibs they use don't
get updated to more modern times to cope with extra features... when in doubt,
use the lowest common denominator.

> And some people wonder why FOSS feels klunky on Windows.

I merrily don't.  Windows always felt klunky to me.

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> Somehow, I always find it amusing how many FOSS programs assume Windows is 
>> basically how it was in Win98 and completely fail to do the right thing.
> 
>   Well, you can always demand your money back.

Sometimes you're extremely predictable.

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New wrote:
> Somehow, I always find it amusing how many FOSS programs assume Windows
> is basically how it was in Win98 and completely fail to do the right thing.
> 

I wouldn't limit this strictly to FOSS. Granted, it's been a while since
I've tried Windows -based freeware and shareware utilities, but on
Windows NT-series this has always been a problem with small utilities.

-Aero

Post a reply to this message

nemesis wrote:
> My guess is that that won't change while the portable C stdlibs they use don't
> get updated to more modern times to cope with extra features... when in doubt,
> use the lowest common denominator.

Yeah, I think that's the basic problem. Even so, you'd think the portable 
libraries would at least try to support the same semantics where they can. I 
hear the "Windows can't delete an open file" so often it's silly. Why 
wouldn't you make your C compiler's "open" function set the flag that says 
"let me delete open files" if that's how it works in POSIX?

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Eero Ahonen wrote:
> Darren New wrote:
>> Somehow, I always find it amusing how many FOSS programs assume Windows
>> is basically how it was in Win98 and completely fail to do the right thing.
>>
> 
> I wouldn't limit this strictly to FOSS. 

True. It mostly annoys me in big programs, tho, like language interpreters 
for example.

As "nemesis" pointed out, it's more a problem in the use of C portability 
layers than anything, which is excusable for small one-off programs. But if 
you're going to port an X-windows based graphics system to Windows, you'd 
think it would be easy to fix some of the other stuff too.

Of course it's FOSS, and if you don't need that, you're not going to 
implement it. I understand that, which is why I find it amusing. Personally, 
I'd be embarrassed to do something like porting Ruby or Python to Windows 
without fixing such a thing.

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Warp wrote:
> > Darren New <dne### [at] sanrrcom> wrote:
> >> Somehow, I always find it amusing how many FOSS programs assume Windows is 
> >> basically how it was in Win98 and completely fail to do the right thing.
> > 
> >   Well, you can always demand your money back.

> Sometimes you're extremely predictable.

  Well, if the argument is valid, why change it?

-- 
                                                          - Warp

Post a reply to this message

Well, if it is open source, why don't you fix it?

(Ducking and taking cover ;-)

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> Warp wrote:
>>> Darren New <dne### [at] sanrrcom> wrote:
>>>> Somehow, I always find it amusing how many FOSS programs assume Windows is 
>>>> basically how it was in Win98 and completely fail to do the right thing.
>>>   Well, you can always demand your money back.
> 
>> Sometimes you're extremely predictable.
> 
>   Well, if the argument is valid, why change it?

You're predictable in that you make that argument even when nobody else is 
arguing. :-)

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

TC wrote:
> Well, if it is open source, why don't you fix it?

I have, in other situations. It's difficult to get involved enough to fix 
all of them. And in part for the same reasons the original authors didnt: 
it's not enough of a pain to me to do that. But it's much more of a pain for 
me to fix it than it is for the original authors to do it right in the first 
place. :-)

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Warp wrote:
> > Darren New <dne### [at] sanrrcom> wrote:
> >> Warp wrote:
> >>> Darren New <dne### [at] sanrrcom> wrote:
> >>>> Somehow, I always find it amusing how many FOSS programs assume Windows is 
> >>>> basically how it was in Win98 and completely fail to do the right thing.
> >>>   Well, you can always demand your money back.
> > 
> >> Sometimes you're extremely predictable.
> > 
> >   Well, if the argument is valid, why change it?

> You're predictable in that you make that argument even when nobody else is 
> arguing. :-)

  I don't think that presenting an argument is the same thing as "arguing"
(ie. to contend in oral disagreement).

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> Warp wrote:
>>> Darren New <dne### [at] sanrrcom> wrote:
>>>> Warp wrote:
>>>>> Darren New <dne### [at] sanrrcom> wrote:
>>>>>> Somehow, I always find it amusing how many FOSS programs assume Windows is 
>>>>>> basically how it was in Win98 and completely fail to do the right thing.
>>>>>   Well, you can always demand your money back.
>>>> Sometimes you're extremely predictable.
>>>   Well, if the argument is valid, why change it?
> 
>> You're predictable in that you make that argument even when nobody else is 
>> arguing. :-)
> 
>   I don't think that presenting an argument is the same thing as "arguing"
> (ie. to contend in oral disagreement).

You're very predictable in that you present that argument every time someone 
says something about FOSS.

In any case, wouldn't you think it rather lame if (for example) Mono only 
read text files that had lines terminated in CRLF? Or if you frequently ran 
across large complex Linux programs that assumed "~fred/x" should expand to 
"/home/fred/x" regardless of what was in /etc/passwd?

They should at least stop putting misinformation in the documentation like 
"you can't delete open files in Windows" or "Windows doesn't have links". 
That's just plain wrong. The minimum would be to phrase it as "we don't open 
files in a mode that allows them to be deleted while open" and "we don't 
support links in Windows".

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> They should at least stop putting misinformation in the documentation like 
> "you can't delete open files in Windows" or "Windows doesn't have links". 
> That's just plain wrong. The minimum would be to phrase it as "we don't open 
> files in a mode that allows them to be deleted while open" and "we don't 
> support links in Windows".

  It's an open source project. You can contribute.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> They should at least stop putting misinformation in the documentation like 
>> "you can't delete open files in Windows" or "Windows doesn't have links". 
>> That's just plain wrong. The minimum would be to phrase it as "we don't open 
>> files in a mode that allows them to be deleted while open" and "we don't 
>> support links in Windows".
> 
>   It's an open source project. You can contribute.

Wow, really?? Thanks, Warp! I didn't know that!

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

> They should at least stop putting misinformation in the documentation like 
> "you can't delete open files in Windows" or "Windows doesn't have links". 
> That's just plain wrong. The minimum would be to phrase it as "we don't 
> open files in a mode that allows them to be deleted while open" and "we 
> don't

Darren, how do you delete an OS-locked file or directory under Windows XP? 
Can you point me to an api-call?

This is NOT meant as a tease, I am really interested in a solution to this 
problem, if you know it.

About the links: yes, windows does indeed support links. I do not dare to 
use them anymore. I tried once, but since people don't expect hard links 
under Windows to some it came as a nasty surprise when they discovered the 
doublettes they thought they were deleting were the actually hard linked 
files...

Post a reply to this message

Darren New wrote:
> Warp wrote:
>> Darren New <dne### [at] sanrrcom> wrote:
>>> They should at least stop putting misinformation in the documentation 
>>> like "you can't delete open files in Windows" or "Windows doesn't 
>>> have links". That's just plain wrong. The minimum would be to phrase 
>>> it as "we don't open files in a mode that allows them to be deleted 
>>> while open" and "we don't support links in Windows".
>>
>>   It's an open source project. You can contribute.
> 
> Wow, really?? Thanks, Warp! I didn't know that!

Wait, wait!  I have an even better idea!

I can go around the internet and correct *everyone* who says something 
wrong! I'm sure that will improve the world and make me loved and respected 
everywhere!

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

TC wrote:
> Darren, how do you delete an OS-locked file or directory under Windows XP? 

You don't lock it.  It's not that you can't delete an *open* file. You can't 
delete an open file that was opened with the "don't let others write to this 
while I have it open" flag.  The program opening the file has to open it in 
a way that lets you delete it, which is why I'm saying I'm surprised that 
(for example) the usual C runtimes don't default to setting this flag, since 
that's the default under UNIX.

The primary problem is that the call to run a program opens the file with 
that lock, in part because it pages out of the file while it's running. The 
file gets locked because to delete a file in Windows you actually have to 
open the file for delete, which requires write privileges. Opening an 
executing file for writing doesn't work under Windows *or* UNIX, but you 
don't have to open the file under UNIX to delete it.

Alternately, you can stick stuff in the registry that'll delete the file (or 
rename it) next time the machine is booted, before all the code that tends 
to hold files open runs, if that's what you mean.

> Can you point me to an api-call?

It's just a normal delete.

> About the links: yes, windows does indeed support links. I do not dare to 
> use them anymore. I tried once, but since people don't expect hard links 
> under Windows to some it came as a nasty surprise when they discovered the 
> doublettes they thought they were deleting were the actually hard linked 
> files... 

Well, yes. That's not a good reason for a programming environment not to use 
them. It's just a good reason to be careful how you expose that to your 
users. :-)

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New wrote:
> Somehow, I always find it amusing how many FOSS programs assume Windows 
> is basically how it was in Win98 and completely fail to do the right thing.

...which does, on the other hand, have the side-effect of making it 
compatible with Win98. ;-)

> which in other words means "we haven't bothered to find the API call 
> that tells you where a particular user's home directory is in Windows, 

Or, more likely, "we just took our C sources and recompiled them on top 
of a POSIX emulator, so we can't do anything that POSIX doesn't support 
/ the emulator doesn't implement".

> And some people wonder why FOSS feels klunky on Windows.

Heh, well, the *really* fun thing is when you have to install GTK+ 
before the program will run... We all know what happens then. ;-)

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Warp wrote:
> > Darren New <dne### [at] sanrrcom> wrote:
> >> They should at least stop putting misinformation in the documentation like 
> >> "you can't delete open files in Windows" or "Windows doesn't have links". 
> >> That's just plain wrong. The minimum would be to phrase it as "we don't open 
> >> files in a mode that allows them to be deleted while open" and "we don't 
> >> support links in Windows".
> > 
> >   It's an open source project. You can contribute.

> Wow, really?? Thanks, Warp! I didn't know that!

  And then you criticize me when I use sarcasm.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> Warp wrote:
>>> Darren New <dne### [at] sanrrcom> wrote:
>>>> They should at least stop putting misinformation in the documentation like 
>>>> "you can't delete open files in Windows" or "Windows doesn't have links". 
>>>> That's just plain wrong. The minimum would be to phrase it as "we don't open 
>>>> files in a mode that allows them to be deleted while open" and "we don't 
>>>> support links in Windows".
>>>   It's an open source project. You can contribute.
> 
>> Wow, really?? Thanks, Warp! I didn't know that!
> 
>   And then you criticize me when I use sarcasm.

No, I criticize you for not making it obvious enough. For the record. :-)

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Orchid XP v8 wrote:
> Darren New wrote:
>> Somehow, I always find it amusing how many FOSS programs assume 
>> Windows is basically how it was in Win98 and completely fail to do the 
>> right thing.
> 
> ...which does, on the other hand, have the side-effect of making it 
> compatible with Win98. ;-)

One might think that, but that tends not to be the case. In part, for 
example, it'll support things like memory-mapped files and etc. as well.

> Or, more likely, "we just took our C sources and recompiled them on top 
> of a POSIX emulator, so we can't do anything that POSIX doesn't support 
> / the emulator doesn't implement".

Yes. Altho I don't think that's really the case. Tcl, for example, has a 
whole X-windows emulation built in. That's why it's (a) somewhat slow, (b) 
tends to look like Motif, and (c) gets used everywhere as a graphics library 
for other languages. :-)

Come to think of it, one of the nice things about Tcl is how much people 
have made extensions for accessing Windows-like stuff.  Considering this, I 
expect I can find Python (for example) extensions to Do The Right Thing.

It really wasn't a complaint, as much as it was a "I wonder why people don't 
fix this when they do it in the first place" sort of thought. Sort of "if 
you're going to try to invent an algorithm to figure out where the home 
directory is, why not just ask the OS". If it was completely unsupported, I 
could see it, but to have it half-ass supported is ... half-ass.

Maybe I'm just getting sick of half-assed technology, tho.

>> And some people wonder why FOSS feels klunky on Windows.
> 
> Heh, well, the *really* fun thing is when you have to install GTK+ 
> before the program will run... We all know what happens then. ;-)

Actually, I don't think I do.

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

>> Or, more likely, "we just took our C sources and recompiled them on 
>> top of a POSIX emulator, so we can't do anything that POSIX doesn't 
>> support / the emulator doesn't implement".
> 
> Yes. Altho I don't think that's really the case.

It is sometimes.

Like, there are several CLI tools which are just recompiled against 
Cygwin to make a "Windows port", and you find that all sorts of strange 
things break when you do that.

> Maybe I'm just getting sick of half-assed technology, tho.

I know *I* am. But it's not like you can do anything about it... I used 
to think you could, but apparently you can't.

>>> And some people wonder why FOSS feels klunky on Windows.
>>
>> Heh, well, the *really* fun thing is when you have to install GTK+ 
>> before the program will run... We all know what happens then. ;-)
> 
> Actually, I don't think I do.

...the program uses the GTK+ look and feel, rather than behaving like a 
native Windows application?

Than again, on Linux it seems each program has its own totally different 
look and feel, so...

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> ...the program uses the GTK+ look and feel, rather than behaving like a
> native Windows application?

I'm pretty sure Gimp on Windows uses the Windows (2000) look and feel.  It
doesn't use its API, though, it instead uses GTK file dialogs and others.

Post a reply to this message

Orchid XP v8 wrote:
>>> Or, more likely, "we just took our C sources and recompiled them on 
>>> top of a POSIX emulator, so we can't do anything that POSIX doesn't 
>>> support / the emulator doesn't implement".
>>
>> Yes. Altho I don't think that's really the case.
> 
> It is sometimes.

Sure. I meant, I don't think it's really the case that the entire Python 
language suite was ported simply by recompiling with cygwin. I might be 
wrong, but it seems like a rather large program for that sort of easy technique.

> I know *I* am. But it's not like you can do anything about it... I used 
> to think you could, but apparently you can't.

You can give up. Or get away from anything even remotely cutting-edge or 
designed for clueful people.

I have noticed my XBox works great. It's very robust and forgiving. (I 
haven't had any hardware problems, mind.) I attribute that to being aimed at 
children.

> ...the program uses the GTK+ look and feel, rather than behaving like a 
> native Windows application?

OK. I guess I don't know what GTK+ looks like enough to recognise that as 
such. Motif, yes. Tcl/Tk and derivatives, yes.

> Than again, on Linux it seems each program has its own totally different 
> look and feel, so...

Sometimes. There's at least three or four common ones.

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Orchid XP v8 wrote:
> >>> Or, more likely, "we just took our C sources and recompiled them on
> >>> top of a POSIX emulator, so we can't do anything that POSIX doesn't
> >>> support / the emulator doesn't implement".
> >>
> >> Yes. Altho I don't think that's really the case.
> >
> > It is sometimes.
>
> Sure. I meant, I don't think it's really the case that the entire Python
> language suite was ported simply by recompiling with cygwin. I might be
> wrong, but it seems like a rather large program for that sort of easy technique.

Certainly not.  But I don't think either it'd be as easy as submitting a patch
with the W32 calls in place.  I'm sure many people may have tried before and it
was ditched because it'd not be fine to have W32 calls right into the
multiplatform core.  Are you sure the MS Windows specific modules don't give you
some handy extra features?  msvcrt, in particular.

> I have noticed my XBox works great. It's very robust and forgiving. (I
> haven't had any hardware problems, mind.) I attribute that to being aimed at
> children.

apart from 3RLs, that is? ;)

> > ...the program uses the GTK+ look and feel, rather than behaving like a
> > native Windows application?
>
> OK. I guess I don't know what GTK+ looks like enough to recognise that as
> such. Motif, yes. Tcl/Tk and derivatives, yes.

Nowadays I'd say it's mostly the industrial or clearlooks look.  The GTK file
dialog is easily recognizable too.

Post a reply to this message

Darren New wrote:
>> About the links: yes, windows does indeed support links. I do not dare 
>> to use them anymore. I tried once, but since people don't expect hard 
>> links under Windows to some it came as a nasty surprise when they 
>> discovered the doublettes they thought they were deleting were the 
>> actually hard linked files... 
> 
> Well, yes. That's not a good reason for a programming environment not to 
> use them. It's just a good reason to be careful how you expose that to 
> your users. :-)
> 

Personally more annoyed by the fact that, at least in XP, you sometimes 
get a message, "You are only deleting the shortcut to the file, not the 
file, are you sure?" Well, since 90% of the shortcuts are programs, 
deleting the original isn't going to do jack for me either, in any case, 
since it won't delete the other 2 GB of shit that got installed with the 
executable. If they cared about the distinction, they should provide an 
"Uninstall" option with that warning... Then again, its a warning for 
morons, so.. probably a bad idea, since some of them are bound to 
uninstall, when all they meant was to delete the shortcut. :p


-- 
void main () {
   If Schrödingers_cat is alive or version > 98 {
     if version = "Vista" {
       call slow_by_half();
       call DRM_everything();
     }
     call functional_code();
   }
   else
     call crash_windows();
}

<A HREF='http://www.daz3d.com/index.php?refid=16130551'>Get 3D Models, 
3D Content, and 3D Software at DAZ3D!</A>

Post a reply to this message

Patrick Elliott <sel### [at] npgcablecom> wrote:
> Then again, its a warning for morons, so..

  I wouldn't say that. It was previously a lot more confusing when you
attempted to delete a shortcut and it simply said "really delete xyz?"
where 'xyz' was the name of the file the shortcut was pointing to, with
no indication whatsoever in the dialog that only the shortcut itself would
be deleted, not the file. It always gave a feeling of playing russian
roulette with a potentially important file.

  Now it explicitly says that what is being deleted is only the shortcut,
not the file being pointed to, which is an improvement.

  An "uninstall" option to shortcuts would actually be a good idea IMO.

-- 
                                                          - Warp

Post a reply to this message

Darren New wrote:
> nemesis wrote:
>> My guess is that that won't change while the portable C stdlibs they use
>> don't get updated to more modern times to cope with extra features...
>> when in doubt, use the lowest common denominator.
> 
> Yeah, I think that's the basic problem. Even so, you'd think the portable
> libraries would at least try to support the same semantics where they can.
> I hear the "Windows can't delete an open file" so often it's silly. Why
> wouldn't you make your C compiler's "open" function set the flag that says
> "let me delete open files" if that's how it works in POSIX?

I'm pretty sure you have to blame Microsoft itself for that. Isn't open() 
defined in msvcrt.dll?

Post a reply to this message

Orchid XP v8 wrote:
> Heh, well, the *really* fun thing is when you have to install GTK+
> before the program will run... We all know what happens then. ;-)

In Linux you have to install GTK+ before the program will run too. The 
difference is that Linux distros have package managers that make it sane.

Post a reply to this message

nemesis wrote:
> Orchid XP v8 <voi### [at] devnull> wrote:
>> ...the program uses the GTK+ look and feel, rather than behaving like a
>> native Windows application?
> 
> I'm pretty sure Gimp on Windows uses the Windows (2000) look and feel.

AFAIK, GTK+ looks like Windows using the same technique as Java Swing: 
implementing the theme from scratch to look like that.

And that is when you notice subtle Windows details that you don't normally 
notice... when Swing gets them wrong.

Post a reply to this message

Nicolas Alvarez wrote:
> I'm pretty sure you have to blame Microsoft itself for that. Isn't open() 
> defined in msvcrt.dll?

Now that you mention it, quite possibly yes. :-)

-- 
Darren New, San Diego CA, USA (PST)
   Human nature dictates that toothpaste tubes spend
   much longer being almost empty than almost full.

Post a reply to this message

>> Heh, well, the *really* fun thing is when you have to install GTK+
>> before the program will run... We all know what happens then. ;-)
> 
> In Linux you have to install GTK+ before the program will run too. The 
> difference is that Linux distros have package managers that make it sane.

More correctly: The difference is that on Linux, *lots* of programs use 
GTK+, not just the GIMP. ;-)

Actually, one of the more perplexing things about Linux is how different 
widget toolkits look and behave completely differently. Things seem to 
have got more sane over there recently, thank God...

Post a reply to this message