Duplicate filenames in kernel source bz2 archive
 

Hello all,

While browsing the the linux kernel source archive "linux-2.6.20.3.tar.bz2" with WinRAR on Windows, i have discovered several duplicate filenames in same folder(s):

/linux-2.6.20.3/include/linux/netfilter/xt_dccp.h

is present two times for the same name, not to mention a "xt_DSCP.h" (note the capitals)

There are many other duplicate filenames.

What does this all mean? Maybe i am not supposed to browse/extract the kernel sources with WinRAR (which DOES support bz2 and tar as far as I can see), or with a Windows box, on an NTFS volume?

I am confused. I will be compiling it on a Linux box, granted, but I wanted to browse the sources before i dive in to it..

UPD: Resolved. there were no two xt_dscp.h. My eyes went bad on me: it was a xt_dscp.h and xt_dccp.h. Concerning capital letter distinction, as in xt_dscp.h vs xt_DSCP.h which DO exist, i am guessing Linux filesystems are case sensitive with filenames, while NTFS / WinRAR is NOT.

Correct, filenames are case sensitive. And unless I'm mistaken, they are also case sensitive on Windows.
The filesystem in question should make no difference, however I would not personally put 100% faith in the extracted archive if extracted with Win-anything. Fine for having a browse around, but please do extract it with a Linux tool when you are ready to work on it ;)

You are mistaken. Windows preserves case, but is not case-sensitive. It will store upper, lower, and mixed-case names as just as you entered them, but FOO.TXT, Foo.Txt, and foo.txt all refer to the same file.

It is possible to configure some (newer?) versions of Windows to use case-sensitive paths, but this has never been the default.

Is that right? WOW <scratches head> as long as I've used Windoze, I think I've never made that distinction :) ; learn something new every day!!
'Course, it's been like 3 months now since I've sat in front of a Windoze machine since switching to Linux, so maybe I've started forgetting already ;)
Thanks for making that clear, AdaHacker.
Sasha

Thanks guys.

Yes, indeed, it seems that either NTFS is case-aware, as opposed to case-sensitive (it is not), or that the extraction software - WinRAR, which i think is fairly compliant and stable piece of work, ignores case upon extracting files (which I dont think it does, since it merely follows the file catalog rules in the archive file).

And, yes, i AM going to use Linux to extract kernel sources when compiling :-)