Let's keep this on the list since it might be of general interest:
On May 31, 2006, at 5:46 PM, mal content wrote:
Hello.
On 31/05/06, Hans-Christoph Steiner hans@eds.org wrote:
If you want this bug to be remembered, submit a bug report.
I wanted to do that, but I couldn't work out where to submit a report.
http://puredata.org/dev/bugtracker
Or find it in the nav bar on the left "developers -> bug tracker"
As for configure, its not so bad for most things, but it can be strange. Mostly, its just like writing a shell script. I would be easy to make it test for both tcl84 and tcl8.4, for example. If you actually read the pd/src/configure.in, I think you'll be able to make the change quite easiler.
I had a look in configure.in first of all, and I saw the lines:
if test $foundit == "no"; then AC_CHECK_HEADER(tcl.h,foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.7/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.7";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.6/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.6";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.5/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.5";foundit=yes,) fi ...
Now, I could add another check for tcl in /usr/local (about 30 lines). It's not pretty, but it'd work.
Just add it to the bottom, its already not pretty, so no real harm done.
The problem is that the same check doesn't occur for the library:
AC_CHECK_LIB(tcl8.7, main,, AC_CHECK_LIB(tcl8.6, main,, AC_CHECK_LIB(tcl8.5, main,, AC_CHECK_LIB(tcl8.4, main,, AC_CHECK_LIB(tcl8.3, main,, AC_CHECK_LIB(tcl8.2, main,,
How do I instruct AC_CHECK_LIB to check /usr/local as well?
This is already in one of the Pd-extended patches (packages/patches/ mingw_fixes-0.39-2.patch) and I think its been incorporated into Miller's CVS code. It should check in /usr/local automatically since /usr/local is generally part of the standard path. But on GNU/ Linux, Tcl/Tk is generally installed into /usr/bin.
.hc
________________________________________________________________________ ____
Using ReBirth is like trying to play an 808 with a long stick. -David Zicarelli
On 31/05/06, Hans-Christoph Steiner hans@eds.org wrote:
Let's keep this on the list since it might be of general interest:
Yes, I didn't realise that you'd added the CC: line, I didn't want to put what might have been intended as private mail on the list. Some people get quite explosive about that!
:)
On May 31, 2006, at 5:46 PM, mal content wrote:
Hello.
On 31/05/06, Hans-Christoph Steiner hans@eds.org wrote:
If you want this bug to be remembered, submit a bug report.
I wanted to do that, but I couldn't work out where to submit a report.
http://puredata.org/dev/bugtracker
Or find it in the nav bar on the left "developers -> bug tracker"
As for configure, its not so bad for most things, but it can be strange. Mostly, its just like writing a shell script. I would be easy to make it test for both tcl84 and tcl8.4, for example. If you actually read the pd/src/configure.in, I think you'll be able to make the change quite easiler.
I had a look in configure.in first of all, and I saw the lines:
if test $foundit == "no"; then AC_CHECK_HEADER(tcl.h,foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.7/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.7";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.6/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.6";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.5/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.5";foundit=yes,) fi ...
Now, I could add another check for tcl in /usr/local (about 30 lines). It's not pretty, but it'd work.
Just add it to the bottom, its already not pretty, so no real harm done.
Right, I'll submit a patch to the bug tracker.
The problem is that the same check doesn't occur for the library:
AC_CHECK_LIB(tcl8.7, main,, AC_CHECK_LIB(tcl8.6, main,, AC_CHECK_LIB(tcl8.5, main,, AC_CHECK_LIB(tcl8.4, main,, AC_CHECK_LIB(tcl8.3, main,, AC_CHECK_LIB(tcl8.2, main,,
How do I instruct AC_CHECK_LIB to check /usr/local as well?
This is already in one of the Pd-extended patches (packages/patches/ mingw_fixes-0.39-2.patch) and I think its been incorporated into Miller's CVS code. It should check in /usr/local automatically since /usr/local is generally part of the standard path. But on GNU/ Linux, Tcl/Tk is generally installed into /usr/bin.
I think the problem is actually that the script looks for libtcl8.X instead of libtcl8X. FreeBSD renamed it. I'm not sure who should be responsible for that one (FreeBSD with compatibility symlinks or pd with extra configure code).
thanks MC
On May 31, 2006, at 6:27 PM, mal content wrote:
On 31/05/06, Hans-Christoph Steiner hans@eds.org wrote:
Let's keep this on the list since it might be of general interest:
Yes, I didn't realise that you'd added the CC: line, I didn't want to put what might have been intended as private mail on the list. Some people get quite explosive about that!
:)
On May 31, 2006, at 5:46 PM, mal content wrote:
Hello.
On 31/05/06, Hans-Christoph Steiner hans@eds.org wrote:
If you want this bug to be remembered, submit a bug report.
I wanted to do that, but I couldn't work out where to submit a report.
http://puredata.org/dev/bugtracker
Or find it in the nav bar on the left "developers -> bug tracker"
As for configure, its not so bad for most things, but it can be strange. Mostly, its just like writing a shell script. I
would be
easy to make it test for both tcl84 and tcl8.4, for example.
If you
actually read the pd/src/configure.in, I think you'll be able
to make
the change quite easiler.
I had a look in configure.in first of all, and I saw the lines:
if test $foundit == "no"; then AC_CHECK_HEADER(tcl.h,foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.7/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.7";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.6/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.6";foundit=yes,) fi if test $foundit == "no"; then AC_CHECK_HEADER(tcl8.5/tcl.h, GUIFLAGS="$GUIFLAGS -I/usr/include/tcl8.5";foundit=yes,) fi ...
Now, I could add another check for tcl in /usr/local (about 30
lines).
It's not pretty, but it'd work.
Just add it to the bottom, its already not pretty, so no real harm done.
Right, I'll submit a patch to the bug tracker.
Excellent, just be sure to use "diff -uw" format.
The problem is that the same check doesn't occur for the library:
AC_CHECK_LIB(tcl8.7, main,, AC_CHECK_LIB(tcl8.6, main,, AC_CHECK_LIB(tcl8.5, main,, AC_CHECK_LIB(tcl8.4, main,, AC_CHECK_LIB(tcl8.3, main,, AC_CHECK_LIB(tcl8.2, main,,
How do I instruct AC_CHECK_LIB to check /usr/local as well?
This is already in one of the Pd-extended patches (packages/patches/ mingw_fixes-0.39-2.patch) and I think its been incorporated into Miller's CVS code. It should check in /usr/local automatically since /usr/local is generally part of the standard path. But on GNU/ Linux, Tcl/Tk is generally installed into /usr/bin.
I think the problem is actually that the script looks for libtcl8.X instead of libtcl8X. FreeBSD renamed it. I'm not sure who should be responsible for that one (FreeBSD with compatibility symlinks or pd with extra configure code).
MinGW also uses the names without dots, so I mean that the MinGW fixes include searching for libtcl84
.hc
________________________________________________________________________ ____
¡El pueblo unido jamás será vencido!
-
Hans-Christoph Steiner -
mal content