On Wed, 2002-05-22 at 21:44, Bernd Petrovitsch wrote: > Rodrigo Barbosa <rodrigob@tisbrasil.com.br> wrote: > >On Wed, May 22, 2002 at 07:43:22PM +0200, petter wahlman wrote: > [...] > >> Linux should IMO be as adaptible and flexible as possible, and _forcing_ > >> filesystem case is wrong. > > No - see below point 2. > > >Errr. Standard answer: code it and submit :-) > > 1) ACK - actually that's the open-source standard answer. > 2) And - if you really want to code and submit it - you are > opening a can of worms : The case-insensitivity bases on the fact > that you know for 2 (or even more) characters that they are equal > though the numerical representation (independent of any encoding) > is the same, namely upper and lower case letters. > This relies actually on the language used, so e.g. in German we > have umlauts which also have a upper and lower case version > (similar to other letters). > a) AFAICS (and AFAIK) there is no i18n and l10n support in the > kernel (which is required to do this properly). > So you have to put it in yourself (one more reason why I believe > that the patch will not be accepted by the core folks[0]). That is a very good point. A case insensitive FS could not have filenames with different names and identical byte by byte values. (strcasecmp would threat them as identical). Still, this could be mentioned in the mount man page or something. This also means that JFS and other filesystems wich has the possiblilty of ignoring case is not unicode safe when this option is turned on(?). > This leaves open the problem at which level the user is able to > specify the language of one given filename - only at the mount point ? No. Let the people that want/need the (and understand the potential consequences) feature use it. > b) You just ignore not-English languages : Well, the not-English > part of the world (or significant parts of it) will probably ignore > your patch too. I don't use an english character set, but a simple casecmp sollution would work for me. FYI I hope that even a patch that 'perfectly' ignores case would be used by as few people as possible. I am logging off for a few days now. I do not belive we will agree on this issue anyway ;) > > Bernd > > [0] : Yes, I know that Win* has apparently such stuff in the kernel. And > so what? > -- > Bernd Petrovitsch Email : bernd@gams.at > g.a.m.s gmbh Fax : +43 1 205255-900 > Prinz-Eugen-Straße 8 A-1040 Vienna/Austria/Europe > LUGA : http://www.luga.at > > > -- > Kernelnewbies: Help each other learn about the Linux kernel. > Archive: http://mail.nl.linux.org/kernelnewbies/ > FAQ: http://kernelnewbies.org/faq/ > -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
