Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Date: Tue, 3 Apr 2001 18:06:24 -0400 (EDT) From: Jan Vicherek To: Glen Coakley cc: cygwin AT cygwin DOT com Subject: RE: How to (dynamically) control Unix/Dos PATH-like variable tran slation In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi all, Here is the problem I have, and why do I think that the variable translation would be really useful: Products get installed on NT/W2k. These products use environment variables (ENVVARs), that point to files and directories, to determine context from which they are being called. I am constructing some platform-independent makefiles, which 1) invoke the installed programs and 2) pass ENVVARs containing pathnames to these to communicate the context and 3) pass cmd-line args containing pathnames to the invoked pgms. The makefile references the same variables. They do not contain all relative paths, but ofter contain absolute paths (c:\prod\app\cfg\...). It semms to me quite natural, that the ENVVARs (and cmd line params) get translated when crossing the Windows-Cygwin boundaries. It doesn't seem to me easy / easily maintainable to translate a variable and parse and translate cmd args to windows path using "cygpath" every time I make an invocation. But I find "cygpath" quite useful otherwise. I would welcome any suggestions on how to solve this ! Would anybody have a suggestion for better solution than translating the ENVVARs and cmd line args ? Thanks, Jan On Mon, 2 Apr 2001, Glen Coakley wrote: > > I also question the usefulness of this. There are special cases where it is > useful, like when you are passing paths from a windows program to a Cygwin > one (I created a hack recently to allow me to right-click on a directory and > get a Cygwin Bash shell in that directory.) But for those cases, I would > think the 'cygpath' utility would be sufficient. > > ________________________________ > Glen Coakley, Sr. Software Engineer > MQSoftware Inc., (763) 543-4845 > Have you ever wonder what happens when you run "rm -rf / " but been afraid > to try it? > > > > -----Original Message----- > > From: Christopher Faylor [mailto:cgf AT redhat DOT com] > > Sent: Sunday, April 01, 2001 9:23 PM > > To: cygwin AT cygwin DOT com > > Cc: honza AT ied DOT com > > Subject: Re: How to (dynamically) control Unix/Dos PATH-like variable > > translation > > > > > > On Sun, Apr 01, 2001 at 10:00:57PM -0400, Jan Vicherek wrote: > > >On Sun, 1 Apr 2001, Christopher Faylor wrote: > > >> On Sun, Apr 01, 2001 at 12:53:28AM -0500, Jan Vicherek wrote: > > >> >On Fri, 30 Mar 2001, Christopher Faylor wrote: > > >> >How can I control which variables get translated Dos->Unix when > > >> >starting a CYGWIN binary from Windows and which variables get > > >> >translated Unix->Dos when starting a DOS/Windows binary > > from CYGWIN ? > > >> > > > >> > > >http://www.cygwin.com/cygwin-ug-net/using.html#USING-PATHNAMES says > > >> >that HOME, PATH, and LD_LIBRARY_PATH are converted, but > > it doesn't say > > >> >how to control any additional ones in runtime. > > >> > > >> The main reason for this is that it isn't possible. Don't > > know why you think > > >> that this list is flexible. It isn't. > > > > > >I didn't see what would preclude the code which says "now translate > > >PATH variable" from one format to another from saying "now translate > > >PATH variable and all other variables listed in variable > > >CYGWIN_TRANSLATABLE" > > > > > >Is there a reason why we couldn't have such "CYGWIN_TRANSLATABLE" > > >variable, which would list names (or patterns) of other > > variables that > > >are to be translated in fashion similair to "PATH" variable ? > > > > Maybe you are misinterpreting what I said. It isn't possible > > because it > > isn't implemented. My observation of "Don't know why..." was > > basically > > aimed at the fact that you seemed to be under the impression that this > > was some kind of undocumented feature. > > > > >What efforts do you estimate it would take to whip up such code ? > > > > > >Do you think such feature wouldn't be very useful ? Why ? > > > > I'm not sure what estimates you're looking for. It would > > probably take > > me a couple of hours of programming to do this. > > > > Would it be useful? Since this is the first time that I can > > recall that > > anyone has asked for this, I think it would be only marginally useful. > > If you are controlling the environment to such a state that > > you can add > > a CYGWIN_TRANSLATABLE variable, then just set the environment > > variables > > using posix paths. I believe that Cygwin already handles the > > ones that > > should be handled. > > > > However, as usual, if someone wants to contribute a patch, I will > > consider it's inclusion in the sources. > > > > In case anyone is interested in submitting a patch, please first check > > out the Contributing link at http://cygwin.com/ . Read it thoroughly > > and, when it is time to provide the patch, make sure that you follow > > all of the rules with regard to the ChangeLog entry, patch format, > > and coding styles. > > > > cgf -- Want to unsubscribe from this list? Check out: http://cygwin.com/ml/#unsubscribe-simple