This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: cyg1.7 - DOS character remapping: change request.

Eric Blake wrote:
Rather than complaining, write a patch to prove your point.  Patches speak
much louder than rants on open source projects.  But I won't be the one
writing the patch.
I already supplied code in the first email. It's a matter of
using those constants instead of the ones being used.

Corinna Vinschen wrote:
According to Linda Walsh on 11/28/2009 3:24 AM:

Any other standards group I know of is going UTF-8.  All of the
linux distributions I know are going UTF-8.  I'd like to see Cygwin
go that way too.

I don't understand this one. What on earth are you think we're doing? Do you really understand the sense of the mapping?
	You are trying to allow use of the commonly usable
'7 banned chars' (under DOS) that are occasionally used
on Unix.  No?  The UTF-8 is referring to the fact
that the substitute chars I'm referring to would be
viewable on platforms using UTF-8 -- as the unicode
values in UTF-16 are directly convertible to UTF-8
platforms (which such wouldn't be the case to platforms
using not Unicode-based platforms).

But that mapping doesn't make sense.  Instead of mapping valid, but
forbidden characters into a range which doesn't contain valid
characters, the valid characters are then mapped onto other valid
characters.  How are you going to ever map them back?  When is a
FULLWIDTH QUOTATION MARK?  You're covering perfectly valid characters
and make them unusable.  Besides, we have not only to map the few
characters you're talking about, the U+f0XX range is also used to
map invalid UTF-8 chars.
	I'm aware that this would reserve the 'display forms'
of those chars and map them them to their real forms when
interpreted within cygwin.  I don't see this to be a problem.

There are no applications that currently use those
values because cygwin 1.5 didn't support either the real ascii values OR the unicode values. It's a matter of
seeing those file names, produced by 1.7 as semi valid
values when viewed in explorer and when they are viewed
on linux servers. _I_ use those values in filesnames,
and know of no compatibility problems having them
treated as 'real' ascii characters under cygwin --
since I am just using them for 'display' purposes
in file names like like "Music:the group:title 1/3".
In linux and windows, I'm using them as display forms.
I put quotes around the filesnames so whether they are converted
to real ascii forms under cygwin or stored as visual display
forms is of no consequence.

	I don't see a problem.  And the patch is just
changing the hex values used to the ones I supplied in
my original posting.

Am I indicating an understanding of the problem
and implications? Do you feel it's really a problem given
how they are commonly used and how there would be no conflicting applications?


-- Problem reports: FAQ: Documentation: Unsubscribe info:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]