This is the mail archive of the
mailing list for the Cygwin project.
Re: Tcl/tk dll's
- From: Charles Wilson <cygwin at cwilson dot fastmail dot fm>
- To: cygwin at cygwin dot com
- Date: Thu, 09 Feb 2012 16:55:10 -0500
- Subject: Re: Tcl/tk dll's
- References: <4F32FC8A.email@example.com> <1328753679.2032.12.camel@YAAKOV04>
- Reply-to: Charles Wilson <cygwin at cwilson dot fastmail dot fm>
On 2/8/2012 9:14 PM, Yaakov (Cygwin/X) wrote:
> On Wed, 2012-02-08 at 23:51 +0100, marco atzeri wrote:
>> curiosity, any reason why the tcl/tk dll's are
>> not using the cyg prefix ?
> In fact, there is. The point of the "cyg" prefix is to avoid possible
> mismatches with MinGW DLLs using the "lib" prefix. In this case,
> however, forcing a "cyg" prefix with Tcl's unique build system would be
> a lot of work for no gain, since Tcl already has its own way of
> distinguishing between the two (tclXY.dll vs. libtclX.Y.dll). Note that
> python and R (in Ports) do the same thing for the same reason.
Sadly, for the (in progress) cygwin-fork-that-shall-not-be-named version
of tcl/tk, I have to do this, because to be a good citizen I have to
avoid conflicting with cygwin's DLLs (libfooX.Y.dll) and 'native' DLLs
(fooXY.dll). Thus, I've already developed all the necessary patches to
make this work -- and the changes, incl. those to the in-tcl build
system for TEA extensions, are not that difficult.
However, if cygwin's tcl/tk packages were to *now* adopt such a patch,
then it would cause no end of back-compatibility issues and Yet Another
Recompile Of All Clients. And that's a Heaping Plateful of Pain, for no
So, for once, I actually agree with Yaakov on one of these issues --
which is why I never mentioned the existence of my patch prior to the
Big Tcl/Tk Transition.
Problem reports: http://cygwin.com/problems.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple