Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

[Wincent Colaiuta]

El 29/8/2007, a las 4:23, Johannes Schindelin escribió:

> On Wed, 29 Aug 2007, Wincent Colaiuta wrote:
>
>> El 28/8/2007, a las 22:45, Johannes Schindelin escribi?:
>>
>>> On Tue, 28 Aug 2007, Wincent Colaiuta wrote:
>>>
>>>> Personally, I don't like HFS+ much at all, but your statement that
>>>> Mac OS X "doesn't even have a case sensitive filesystem" is false.
>>>
>>> It is right that they support it.  But since the _default_ is case
>>> insensitive (but only as long as it is not _reporting_ file  
>>> names), it
>>> is _as bad_ as "doesn't even have a case sensitive filesystem".  No
>>> sophistry helps here.
>>
>> I don't think my correction qualifies as "sophistry" by any  
>> reasonable
>> definition of the word. Andreas claimed that Apple had turned "a
>> perfectly decent unix-clone such as FreeBSD into some defect  
>> monstrosity
>> that doesn't even have a case sensitive filesystem"; I merely  
>> sought to
>> correct his misstatement.
>
> I acknowledged that.
>
> But it does not change the _meaning_ of Andreas' criticism.  They  
> took a
> perfectly sane system, and turned it into a mess.

I don't think it's productive to enter into arguments about whether a  
particular operating system is a "monstrosity", "sane" or "a mess";  
that's why I limited my comment to a correction of a factual  
misstatement by Andreas and left the opinion part unchallenged. We  
could have avoided this thread if instead of "doesn't even have a  
case sensitive filesystem", he'd started his flame with "defaults to  
a using a case-insensitive filesystem".

> Yes, you _can_ change the setting.  No, most don't.  Yes, the  
> effect is...
> you guessed it: the same as if they did not allow case sensitivity  
> at all.

While I agree that most just accept the default, your argument here  
and in your previous email is spurious; it amounts to saying "A and B  
is offered, but because A is the default it is the same as if B were  
not supported at all", which isn't valid logic. If you limited your  
argument to criticizing case-insensitive filesystems then I would  
have to agree with you; I am not a fan of them. Likewise, if you said  
that the default has undesirable consequences I would also concur.  
It's your attempt to extend your claim to an argument for equivalence  
that I can't buy.

> Plenty of people have this setup

Agreed...

> and we have to suffer.

but, don't you think you're exaggerating a bit here? How much  
"suffering" has this really caused you? I gather that you're not even  
a user of that operating system; I am (have been since the first  
developer previews), have always accepted the default filesystem  
choice, don't even like the filesystem (have ranted repeatedly about  
its flakiness on my weblog), but I think I've run into case- 
insensitivity issues two times. If you want to talk about "suffering"  
then ask me about HFS+ data integrity.

Cheers,
Wincent

Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

[David Kastrup]

Wincent Colaiuta <win@wincent.com> writes:

> El 29/8/2007, a las 4:23, Johannes Schindelin escribió:
>
>> On Wed, 29 Aug 2007, Wincent Colaiuta wrote:
>>
>> But it does not change the _meaning_ of Andreas' criticism.  They
>> took a perfectly sane system, and turned it into a mess.
>
> While I agree that most just accept the default, your argument here
> and in your previous email is spurious; it amounts to saying "A and
> B is offered, but because A is the default it is the same as if B
> were not supported at all", which isn't valid logic. If you limited
> your argument to criticizing case-insensitive filesystems then I
> would have to agree with you; I am not a fan of them. Likewise, if
> you said that the default has undesirable consequences I would also
> concur.  It's your attempt to extend your claim to an argument for
> equivalence that I can't buy.

Read up on his posting history.

>> Plenty of people have this setup
>
> Agreed...
>
>> and we have to suffer.
>
> but, don't you think you're exaggerating a bit here? How much
> "suffering" has this really caused you? I gather that you're not
> even a user of that operating system;

He is the main worker on the MSYS port of git as far as I can tell,
while not using Windows as his work platform.  So indeed he gets to
bear _all_ the pain, and no personal gain.  Since I have provided for
Windows/MSYS compatibility in AUCTeX's _installation_ procedure (a
small part of the project which is mostly operating system agnostic),
a much smaller stuff under similar circumstances, I can vouch for the
amount of hassle such a seemingly "everything we need should be there
in principle" Windows accommodation task causes.

And I actually find it extremely distasteful how little involvement in
this project is there from genuine Windows users: most restrict
themselves to complaining.

So at least on this account, I think you are not doing him justice.

> I am (have been since the first developer previews), have always
> accepted the default filesystem choice, don't even like the
> filesystem (have ranted repeatedly about its flakiness on my
> weblog), but I think I've run into case- insensitivity issues two
> times. If you want to talk about "suffering" then ask me about HFS+
> data integrity.

-- 
David Kastrup

Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

[Andreas Ericsson]

Wincent Colaiuta wrote:
> El 29/8/2007, a las 4:23, Johannes Schindelin escribió:
> 
>>
>> I acknowledged that.
>>
>> But it does not change the _meaning_ of Andreas' criticism.  They took a
>> perfectly sane system, and turned it into a mess.
> 
> I don't think it's productive to enter into arguments about whether a 
> particular operating system is a "monstrosity", "sane" or "a mess"; 
> that's why I limited my comment to a correction of a factual 
> misstatement by Andreas and left the opinion part unchallenged. We could 
> have avoided this thread if instead of "doesn't even have a case 
> sensitive filesystem", he'd started his flame with "defaults to a using 
> a case-insensitive filesystem".
> 
>> Yes, you _can_ change the setting.  No, most don't.  Yes, the effect 
>> is...
>> you guessed it: the same as if they did not allow case sensitivity at 
>> all.
> 
> While I agree that most just accept the default, your argument here and 
> in your previous email is spurious; it amounts to saying "A and B is 
> offered, but because A is the default it is the same as if B were not 
> supported at all", which isn't valid logic. If you limited your argument 
> to criticizing case-insensitive filesystems then I would have to agree 
> with you; I am not a fan of them. Likewise, if you said that the default 
> has undesirable consequences I would also concur. It's your attempt to 
> extend your claim to an argument for equivalence that I can't buy.
> 

My original comment was in error. I was unaware that Apple shipped
different filesystems and that there was actually a choice. Me having
sat by a friend who did his installation didn't realize I had the option
of using a case-sensitive filesystem, so I assumed it wasn't there, which
only goes to show that assumption indeed is the mother of all fuckups. Not
wanting a pointless discussion in which everyone agrees in spirit but
nitpicks about factual details, I didn't respond to any of the posts.

Since it seems we all agree that case-insensitive, case-reporting file-
systems are disgusting and it's now thoroughly proven that Apple does
indeed ship a proper filesystem with its OS X releases, can we please
drop this now and move on to trying to improve things? I'll take
universal silence on the subject as "good idea, let's do that"

> 
>> and we have to suffer.
> 
> but, don't you think you're exaggerating a bit here? How much 
> "suffering" has this really caused you?

Plenty enough, I should imagine. As David Kastrup mentioned, Dscho
does extensive work on MSYS/MingW while not using it himself.

Besides that, there are all these timeconsuming threads that some
ignorant dumbass (in this case me) turn into silly pie-fights with
some uninformed remark that is overly tainted by personal opinion.

My apologies. I shall do my research better in the future.

-- 
Andreas Ericsson                   andreas.ericsson@op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231

Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

[Wincent Colaiuta]

El 29/8/2007, a las 11:43, David Kastrup escribió:

>>> Plenty of people have this setup
>>
>> Agreed...
>>
>>> and we have to suffer.
>>
>> but, don't you think you're exaggerating a bit here? How much
>> "suffering" has this really caused you? I gather that you're not
>> even a user of that operating system;
>
> He is the main worker on the MSYS port of git as far as I can tell,
> while not using Windows as his work platform.  So indeed he gets to
> bear _all_ the pain, and no personal gain.

If you look at the thread, his comments in this instance were about  
Mac OS X, not Windows.

As for Dscho's Window's-related pain, I do sympathize.

Cheers,
Wincent

Re: git: cannot rename foo to Foo on a case-insensitive filesystem (e.g. on Windows)

[Nix]

On 29 Aug 2007, Linus Torvalds verbalised:
> And if you use that algorithm, them 'ß' and 'ss' are equal in a 
> case-insensitive filesystem. 
>
> Does anybody do that? I don't know. But I sure as hell wouldn't be 
> surprised.

Oh yes, they do. Guess what I spent the last month fixing in a huge heap
of proprietary software. Of course the problem is not truly fixable
because of the potentially unbounded set of context-dependent stuff such
as `Masse'.

(I hear some people got a summer holiday...)