-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2010-05-08 02:56, Linda Walsh wrote:
Patrick Shanahan wrote:
...
The cli is populated by humans. Not machines. Backwards compatibility is not as paramount. Bill Gates wanted to keep same interfacesin Windows because he thought his customers were incapable of adapting. I certainly DON't see as being the case among unix users, depsite some people's feelings that CLI operations should remain case sensitive.
It is poor design to rely on case as a differentiator between different functions. If you were to program constantly re-using variable names that vary only in 'case', you'd be cursed by any maintainers. Why? because it's difficult to maintain. Why is the CLI-interface any different?
C is that way, too.
If it was in any program -- someone would say "re-write the bad portions". Why shouldn't we strive to make CLI operations the best possible as well?
If you/we want case sensitivity/or not, it a global ENV var -- that might allow both sides to have what they want without having to type something 'extra' in at a command prompt.
The filesystem is case-sensitive, and that can not change, it is not possible to let users choose. It would have to be a universal change, and would break lot of things. Thus bash has to be case sensitive. Options/parameters to commands are also case sensitive. And thus rpm archives are also case sensitive. It is zypper which is different.
Being the person bitten by this instance of the bug (because I'm not used to differing case being accepted by two different package managers on linux, I am stating what would have made the situation acceptable to me -- as someone who as used unix and/or linux for over 20 years. A simple notification that it had ignored my case to make a match would be fine with me.
Agreed, it should not have ignored case and should have reported that it acted contrary to expected.
I didn't say it shouldn't have ignored case -- but that it should have *told* me it was doing so (i.e. doing something other than traditional behavior).
I feel that giving warning of non-tradition behavior is enough -- I don't feel that it has to adhere to traditional behavior. Why wouldn't the warning be sufficient (like either of us are going to do anything about this other than hash-at it here...;) )
Yes, zypper should have warned. It is an interesting feature that it searches ignoring case, but it should at least warn the user that if found a match with different case - as you say you were bitten by this. As zypper is a slow command, finding a package with a similar name that can be the one you wanted is an interesting feature. If warned. Then switches could be added to modify the behaviour. And devs have posted on this list, they are going to do something about this. The person who posted wasn't sure what exactly to do, and it will not be for 11.3, anyway. - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAkvlLg0ACgkQU92UU+smfQU8oACfUYVk7UVU2yMWaXAx2zUbDG4l VRUAnjXnOZy9GghDeFw6HdpZZwDlrW1e =tMIN -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org