[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Case Insensitive Filesystems

From: Mark D. Baushke
Subject: Re: Case Insensitive Filesystems
Date: Wed, 04 Jun 2003 10:24:33 -0700

Derek Robert Price <derek@ximbiot.com> writes:

> Larry Jones wrote:
> >I really wish case insensitive file systems
> >would just go away.  :-)
> >
> You've got my vote!  :)
> I fixed an RCS archive corruption problem back in March that occurred
> when case insensitve clients attempted to rename a file (rm/add), but
> I introduced some new spurious error messages & a server abort when it
> probably shouldn't.  At least it's no longer archive corruption, and
> the trigger cases are rare enough, but GAH!  I'll probably have to
> spend another day or two fixing that.  :(
> Mark, adding the tests cases for the case insensitivity problem is
> going to be, at the least, a pain.  It requires a case-insensitive
> client and a case-sensitive server at the same time.  It's possible
> that some forced definitions on a case-sensitive server could produce
> the case-insensitive client, but that's more overhead than I want to
> get into yet.  Maybe after I finish the autotest update.  One of these
> days I'll get back to that.  :(

Yeah, I can understand that. I have been trying to figure out a way to
do a 'replay' of the appropriate client/server protocols, but nothing
brilliant has come to mind for me either.

> Maybe the script to pipe to an invocation of `cvs server' would be
> short enough to make it reasonable to mimic a case-insensitve client
> in sanity.sh, but I'm not sure I have the time for that.

The other idea is to work out what needs to happen in sanity.sh on
the server versus what has to happen on the client and actually have
the manipulations done using $CVS_RSH commands rather than locally.

This would then allow heterogeneous tests to be run between some
platforms configurations.

To be honest, I think this kind of split configuration will be needed if
we want to get the PAM-basd authentication methods really tested

        -- Mark

reply via email to

[Prev in Thread] Current Thread [Next in Thread]