From: lhall AT rfk DOT com (Larry Hall) Subject: Re: Answer to your text == binary question 22 Jan 1999 11:24:39 -0800 Message-ID: <3.0.5.32.19990122135357.01609d70.cygnus.cygwin32.developers@pop.ma.ultranet.com> References: <19990122124858 DOT A27735 AT cygnus DOT com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: Christopher Faylor , Geoffrey Noer Cc: cygwin32-developers AT cygnus DOT com At 12:48 PM 1/22/99 -0500, Christopher Faylor wrote: >On Fri, Jan 22, 1999 at 11:53:09AM -0500, Larry Hall (RFK Partners, Inc) wrote: >>OK, I think my concern was that the "automount" would happen >>"on-the-fly" in certain cases (something similar to when the user >>addresses a partition with the current "drive" syntax of "///"). >>In these cases, it may not be possible to change the mode before the >>user performs the desired operation (at least for the first one). For >>example, the user can currently do something like "cp //d/foo ." My >>understanding of what Geoffrey was saying is that this will change to >>"cp d:/foo ." (or something similar... I lost his original message >>now!:-() with a mount being created for d:/ in /cygdrive. > >This changes the binmode type for all auto-mounted drives: > > mount -b --change-automount-root /cygdrive > >If you don't do this, the default is text-mode. > >>If this is the case, there needs to be a way to specify the user's >>default preference for the mode of this mount so that even the "cp" >>command gets it right. If after the fact, the mount still exists (and >>it sounds like it would), it would then be possible to change the mode >>for subsequent uses but this is both cumbersome and, more importantly, >>doesn't help the "cp". Reading through your responses, I was struck by >>the idea that you were suggesting that the things mounted under the >>"automount" root would assume the mode of the root. If this is the >>case, it is like the "/" idea that Earnie and Corinna liked, although I >>like it better since it relates specifically to the mounts which are >>"automounted" (where "/" may have been remounted for a different >>purpose and may not reflect the user's default mode preference for >>"automounts"). Assuming that this "automount" root can be created >>and/or set apriori and that its mode would be the default for all >>things under it, I think that addresses my concern. If it was not your >>intention to imply this Chris, forgive me for putting words in your >>mouth! It may still be a good idea though!;-) Did I miss a point of >>yours relative to the long-winded discussion of my concern? > >I don't think so. I was proposing that binmode is controlled by >two separate mechanisms. > >1) For disk drives the setting of the mount associated with a file > controls the default behavior. > >2) For everything else, the default is binary but this is overridden > by CYGWIN=nobinmode. > >That's it. > >The point of confusion is that 1) applies to the new 'automount' method >as well "standard" mounts. I think that is entirely consistent, >however. > >-chris > OK, I guess I'm having trouble seeing a problem with this idea!;-) Larry