Re: Problems updating math/maxima
On 2023/05/30 13:38, Jonathan Drews wrote:
> On Tue, May 30, 2023 at 03:00:13PM +0200, Ingo Feinerer wrote:
> > Am 2023-05-30 03:13, schrieb George Koehler:
> > > ok gkoehler@ (with the PLIST fix)
> >
> > Also OK from my side (feinerer@). Thank you all for your efforts.
> >
> > I do not use Maxima any longer, so I would like to drop MAINTAINER.
> >
>
> I, Jonathan, would be willing to take tis port over.
>
> > > https://marc.info/?l=openbsd-ports&m=168501263910495&w=2 has the diff.
> > > After I applied the diff, I edited pkg/PLIST to change
> > > "${MODPY_BIN_SUFFIX}" back to "-2".
> > >
> > > On Thu, 25 May 2023 11:32:44 -0600
> > > Jonathan Drews wrote:
> > >
> > > > Stuart and Theo:
> > > >
> > > > I compiled maxima with the Python entries removed and it
> > > > installed and runs ok. Here is the Makefile based on Suarts diff. I
> > > > did retain the references to tcl8.6.
> > > > ...
> > > > MODTCL_VERSION= 8.6
> > > > ...
> > > > pre-build:
> > > > ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
> > >
> > > I didn't need the MODTCL_VERSION line nor the pre-build target, so
> > > I didn't add them to my Makefile.
> > >
> > > If your ports tree is from cvs (https://www.openbsd.org/anoncvs.html)
> > > then you can create a diff like,
> > >
> > > $ cd /usr/ports/math/maxima
> > > $ cvs -q diff > ~/my.diff
> > >
> > > This works if the update doesn't add or remove any files. After the
> > > update to 5.46.0 is in cvs, we will be able to create diffs to
> > > add the sbcl flavor.
>
> Here is what is going on. I applied Stuarts diff of May 25th and
> Maxima-ecl built (make build was successful). I am going to make the
> package and also do make test.
>
> Next I will make build the maxima-sbcl version and test it.
>
> Finally, I will combine the two into one Makefile. It will require a
> nested .if ... .endif as there is already two flavors to deal with.
> Namely:
>
> .if !${FLAVOR:Mno_x11}
> MODULES+= x11/tk
> MODTK_VERSION= 8.6
> RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
> math/gnuplot
> BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
> CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
> .endif
It wouldn't be nested, it would be separate to the no_x11 flavour.
You'd want to add a block along these lines.
FLAVORS=no_x11 sbcl
.if ${FLAVOR:Msbcl}
BUILD_DEPENDS +=lang/sbcl
RUN_DEPENDS += lang/sbcl ## is that needed?
CONFIGURE_ARGS += --enable-sbcl
.else
LIB_DEPENDS += lang/ecl
WANTLIB += ecl
CONFIGURE_ARGS += --enable-ecl
.endif
(and add a note about flavours to pkg/DESCR).
> Alternatively we could split this into two ports called maxima-ecl and
> maxima-sbcl.
>
> Let me know what you think.
>
> --
> Kind regards,
>
> Jonathan
>
Re: Problems updating math/maxima
On Tue, May 30, 2023 at 03:00:13PM +0200, Ingo Feinerer wrote:
> Am 2023-05-30 03:13, schrieb George Koehler:
> > ok gkoehler@ (with the PLIST fix)
>
> Also OK from my side (feinerer@). Thank you all for your efforts.
>
> I do not use Maxima any longer, so I would like to drop MAINTAINER.
>
I, Jonathan, would be willing to take tis port over.
> > https://marc.info/?l=openbsd-ports&m=168501263910495&w=2 has the diff.
> > After I applied the diff, I edited pkg/PLIST to change
> > "${MODPY_BIN_SUFFIX}" back to "-2".
> >
> > On Thu, 25 May 2023 11:32:44 -0600
> > Jonathan Drews wrote:
> >
> > > Stuart and Theo:
> > >
> > > I compiled maxima with the Python entries removed and it
> > > installed and runs ok. Here is the Makefile based on Suarts diff. I
> > > did retain the references to tcl8.6.
> > > ...
> > > MODTCL_VERSION= 8.6
> > > ...
> > > pre-build:
> > > ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
> >
> > I didn't need the MODTCL_VERSION line nor the pre-build target, so
> > I didn't add them to my Makefile.
> >
> > If your ports tree is from cvs (https://www.openbsd.org/anoncvs.html)
> > then you can create a diff like,
> >
> > $ cd /usr/ports/math/maxima
> > $ cvs -q diff > ~/my.diff
> >
> > This works if the update doesn't add or remove any files. After the
> > update to 5.46.0 is in cvs, we will be able to create diffs to
> > add the sbcl flavor.
Here is what is going on. I applied Stuarts diff of May 25th and
Maxima-ecl built (make build was successful). I am going to make the
package and also do make test.
Next I will make build the maxima-sbcl version and test it.
Finally, I will combine the two into one Makefile. It will require a
nested .if ... .endif as there is already two flavors to deal with.
Namely:
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
MODTK_VERSION= 8.6
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
.endif
Alternatively we could split this into two ports called maxima-ecl and
maxima-sbcl.
Let me know what you think.
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
Am 2023-05-30 03:13, schrieb George Koehler:
For the update to maxima 5.46.0, I'm happy with Stuart's diff (from
May 25), after fixing the MODPY_BIN_SUFFIX in PLIST. I rarely use
maxima, but I can run simple commands like "diff(5 * x**2, x);".
ok gkoehler@ (with the PLIST fix)
Also OK from my side (feinerer@). Thank you all for your efforts.
I do not use Maxima any longer, so I would like to drop MAINTAINER.
https://marc.info/?l=openbsd-ports&m=168501263910495&w=2 has the diff.
After I applied the diff, I edited pkg/PLIST to change
"${MODPY_BIN_SUFFIX}" back to "-2".
On Thu, 25 May 2023 11:32:44 -0600
Jonathan Drews wrote:
Stuart and Theo:
I compiled maxima with the Python entries removed and it
installed and runs ok. Here is the Makefile based on Suarts diff. I
did retain the references to tcl8.6.
...
MODTCL_VERSION= 8.6
...
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
I didn't need the MODTCL_VERSION line nor the pre-build target, so
I didn't add them to my Makefile.
If your ports tree is from cvs (https://www.openbsd.org/anoncvs.html)
then you can create a diff like,
$ cd /usr/ports/math/maxima
$ cvs -q diff > ~/my.diff
This works if the update doesn't add or remove any files. After the
update to 5.46.0 is in cvs, we will be able to create diffs to
add the sbcl flavor.
Re: Problems updating math/maxima
On Mon, 29 May 2023 20:30:16 -0600 Jonathan Drews wrote: > Thanks for checking this. I used the ports.tat.gz because that is > what the Porters Handbook showed. Can I update the downloaded tarball > version using cvs? I suppose I can try. $ cd /usr/ports $ cvs -qd anoncvs@:/cvs up -A You need cvs -d because ports.tar.gz doesn't know which anoncvs server to use. Fill in with a server from https://www.openbsd.org/anoncvs.html > Can we have a maxima port similar to Tcl? Wherein there are > subdirectories for ECL and SBCL? Maxima wants FLAVORS, not subdirectories. Tcl is different because Tcl has 2 versions (8.5 and 8.6). Maxima has only 1 version. For example, lang/sbcl has FLAVORS sbcl-- and sbcl--threads. Maxima might have maxima-- and maxima--sbcl. Then we might twice build the same version of Maxima, once with --enable-ecl, once with --enable-sbcl.
Re: Problems updating math/maxima
On Mon, May 29, 2023 at 09:13:09PM -0400, George Koehler wrote:
> For the update to maxima 5.46.0, I'm happy with Stuart's diff (from
> May 25), after fixing the MODPY_BIN_SUFFIX in PLIST. I rarely use
> maxima, but I can run simple commands like "diff(5 * x**2, x);".
>
> ok gkoehler@ (with the PLIST fix)
>
George:
Thanks for checking this. I used the ports.tat.gz because that is
what the Porters Handbook showed. Can I update the downloaded tarball
version using cvs? I suppose I can try.
The Maxima-5.47.0 release is due any week now. Look:
https://tinyurl.com/5frvanz7
It contains bug fixes.
So I would not put a lot of effort into this. In the meantime I can
figure out how to down load ports from cvs.
Can we have a maxima port similar to Tcl? Wherein there are
subdirectories for ECL and SBCL?
> https://marc.info/?l=openbsd-ports&m=168501263910495&w=2 has the diff.
> After I applied the diff, I edited pkg/PLIST to change
> "${MODPY_BIN_SUFFIX}" back to "-2".
>
> On Thu, 25 May 2023 11:32:44 -0600
> Jonathan Drews wrote:
>
> > Stuart and Theo:
> >
> > I compiled maxima with the Python entries removed and it
> > installed and runs ok. Here is the Makefile based on Suarts diff. I
> > did retain the references to tcl8.6.
> > ...
> > MODTCL_VERSION= 8.6
> > ...
> > pre-build:
> > ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
>
> I didn't need the MODTCL_VERSION line nor the pre-build target, so
> I didn't add them to my Makefile.
>
> If your ports tree is from cvs (https://www.openbsd.org/anoncvs.html)
> then you can create a diff like,
>
> $ cd /usr/ports/math/maxima
> $ cvs -q diff > ~/my.diff
>
> This works if the update doesn't add or remove any files. After the
> update to 5.46.0 is in cvs, we will be able to create diffs to
> add the sbcl flavor.
Re: Problems updating math/maxima
For the update to maxima 5.46.0, I'm happy with Stuart's diff (from
May 25), after fixing the MODPY_BIN_SUFFIX in PLIST. I rarely use
maxima, but I can run simple commands like "diff(5 * x**2, x);".
ok gkoehler@ (with the PLIST fix)
https://marc.info/?l=openbsd-ports&m=168501263910495&w=2 has the diff.
After I applied the diff, I edited pkg/PLIST to change
"${MODPY_BIN_SUFFIX}" back to "-2".
On Thu, 25 May 2023 11:32:44 -0600
Jonathan Drews wrote:
> Stuart and Theo:
>
> I compiled maxima with the Python entries removed and it
> installed and runs ok. Here is the Makefile based on Suarts diff. I
> did retain the references to tcl8.6.
> ...
> MODTCL_VERSION= 8.6
> ...
> pre-build:
> ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
I didn't need the MODTCL_VERSION line nor the pre-build target, so
I didn't add them to my Makefile.
If your ports tree is from cvs (https://www.openbsd.org/anoncvs.html)
then you can create a diff like,
$ cd /usr/ports/math/maxima
$ cvs -q diff > ~/my.diff
This works if the update doesn't add or remove any files. After the
update to 5.46.0 is in cvs, we will be able to create diffs to
add the sbcl flavor.
Re: Problems updating math/maxima
On 2023/05/27 02:35, Jonathan Drews wrote: > Thank you George: > > Good catch. We will ignore this update.. > > On Fri, May 26, 2023, at 23:34, George Koehler wrote: > > On Fri, 26 May 2023 06:21:35 -0600 > > Jonathan Drews wrote: > > > > > Dima was right. Also one of the replies I got from the maxima discuss > > > list: > > > "My experience is that sbcl is way faster than ecl in practical use. > > > "The Testsuite (at least +the share one) loads many packages and does > > > try not to do any lengthy work with them. Loading +packages is not > > > extremely fast in sbcl. The only drawback of sbcl is that sbcl runs > > > out of +memory in places ECL doesn't." > > > > lang/sbcl is "ONLY_FOR_ARCHS = amd64 i386 powerpc". I need ecl for > > powerpc64, where I can't run sbcl, but I can run ecl and maxima. > > > > sbcl has "USE_WXNEEDED = Yes", ecl doesn't. Some users might prefer > > sbcl for speed; others might prefer ecl for W^X. > > > > --George > > This situation is well suited to flavours.
Re: Problems updating math/maxima
Thank you George: Good catch. We will ignore this update.. On Fri, May 26, 2023, at 23:34, George Koehler wrote: > On Fri, 26 May 2023 06:21:35 -0600 > Jonathan Drews wrote: > > > Dima was right. Also one of the replies I got from the maxima discuss > > list: > > "My experience is that sbcl is way faster than ecl in practical use. > > "The Testsuite (at least +the share one) loads many packages and does > > try not to do any lengthy work with them. Loading +packages is not > > extremely fast in sbcl. The only drawback of sbcl is that sbcl runs > > out of +memory in places ECL doesn't." > > lang/sbcl is "ONLY_FOR_ARCHS = amd64 i386 powerpc". I need ecl for > powerpc64, where I can't run sbcl, but I can run ecl and maxima. > > sbcl has "USE_WXNEEDED = Yes", ecl doesn't. Some users might prefer > sbcl for speed; others might prefer ecl for W^X. > > --George >
Re: Problems updating math/maxima
On Fri, 26 May 2023 06:21:35 -0600 Jonathan Drews wrote: > Dima was right. Also one of the replies I got from the maxima discuss > list: > "My experience is that sbcl is way faster than ecl in practical use. > "The Testsuite (at least +the share one) loads many packages and does > try not to do any lengthy work with them. Loading +packages is not > extremely fast in sbcl. The only drawback of sbcl is that sbcl runs > out of +memory in places ECL doesn't." lang/sbcl is "ONLY_FOR_ARCHS = amd64 i386 powerpc". I need ecl for powerpc64, where I can't run sbcl, but I can run ecl and maxima. sbcl has "USE_WXNEEDED = Yes", ecl doesn't. Some users might prefer sbcl for speed; others might prefer ecl for W^X. --George
Re: Problems updating math/maxima
On Fri, May 26, 2023 at 12:23:18PM +0100, Stuart Henderson wrote:
> On 2023/05/25 18:57, Jonathan Drews wrote:
> > On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote:
> > >
> > > As well, less urgently,
> > > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl.
> > > (sbcl typically gives considerably better performance, and it's available
> > > in ports,
>
> > I wrote the maxima discuss list and they affirm the performance boost
> > that Dima mentions by using sbcl. However I get this make build error:
> >
> > Missing library for sbcl>=0.0
> > Fatal error
> > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2385
>
> sbcl is a standalone compiler binary + runtime files, there is no
> library so you can't use LIB_DEPENDS+WANTLIB for it. would be
> BUILD_DEPENDS and maybe also RUN_DEPENDS if needed.
>
Stuart and others:
I got maxima running with sbcl. Here is the result of make test:
# make test > MakeTestResults.txt 2>&1
Regression tests for maxima-5.46.0
Making check in admin
gmake[1]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/admin'
gmake[1]: Nothing to be done for 'check'.
gmake[1]: Leaving directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/admin'
Making check in crosscompile-windows
gmake[1]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/crosscompile-windows'
gmake[1]: Nothing to be done for 'check'.
gmake[1]: Leaving directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/crosscompile-windows'
Making check in src
gmake[1]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/src'
if test share-subdirs_autogenerated.lisp -nt share-subdirs.lisp; then
cp share-subdirs_autogenerated.lisp share-subdirs.lisp; fi
gmake[1]: Leaving directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/src'
Making check in lisp-utils
gmake[1]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/lisp-utils'
gmake[1]: Nothing to be done for 'check'.
gmake[1]: Leaving directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/lisp-utils'
Making check in tests
gmake[1]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/tests'
gmake check-TESTS
gmake[2]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/tests'
gmake[3]: Entering directory
'/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/tests'
*** Signal SIGINT in . (/usr/ports/pobj/maxima-5.46.0/.test_done)
*** Error 1 in /usr/ports/math/maxima
(/usr/ports/infrastructure/mk/bsd.port.mk:2600 'test':
@lock=maxima-5.46.0; export _LOCKS_HELD=" maxi...)
I got two errors:
Here is my diff:
$ diff -eu Makefile.org Makefile > Makefile.diff
--- Makefile.orgFri May 26 18:02:01 2023
+++ MakefileFri May 26 18:04:37 2023
@@ -1,6 +1,6 @@
COMMENT= GPL computer algebra system based on DOE
Macsyma
-VERSION= 5.44.0
+VERSION= 5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES=math
@@ -10,31 +10,33 @@
# GPLv2+
PERMIT_PACKAGE=Yes
-WANTLIB += c ecl ffi gc gmp m pthread
+WANTLIB += c ffi gc gmp m pthread
MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
USE_GMAKE= yes
-# Python is not needed for building but configure fails without it.
-# Instead of patching just add the lang/python dependency.
-# http://sourceforge.net/p/maxima/mailman/message/34462954/
-MODULES = lang/python
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
-MODPY_RUNDEP = No
+MODTCL_VERSION=8.6
+#MODPY_RUNDEP =No
-BUILD_DEPENDS =print/texinfo
-LIB_DEPENDS= lang/ecl
+BUILD_DEPENDS =print/texinfo \
+ lang/sbcl
+# LIB_DEPENDS =lang/sbcl
RUN_DEPENDS = devel/desktop-file-utils \
misc/rlwrap \
- misc/shared-mime-info
+ misc/shared-mime-info \
+ lang/sbcl
+
+pre-build:
+ ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
+
SUBST_VARS=VERSION
MAKE_FLAGS = MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
-CONFIGURE_ARGS+= --enable-ecl
+CONFIGURE_ARGS+= --enable-sbcl
CONFIGURE_STYLE= gnu
FLAVORS= no_x11
@@ -42,11 +44,12 @@
Re: Problems updating math/maxima
On Fri, May 26, 2023 at 03:32:57PM +0100, Stuart Henderson wrote:
> On 2023/05/26 14:23, Dima Pasechnik wrote:
> > Is there any particular reason there is no modern makeinfo in OpenBSD?
> > (the one that has texi2any executable, that's version 6.8 or newer)
>
> That is available in the texinfo package ('g' prefix in the binary
> names). It's a little old (6.5), no particular reason it's outdated afaik,
> just nobody has done the work to update and test.
it's too old to build Maxima's docs - that's what I get if I try using it, and
not version 6.8:
make[3]: Entering directory
'/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info'
./build_html.sh
./build_html.sh:
abs_srcdir=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info,
pwd=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info, language=en,
languagedir=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info,
texiinitfile=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info/texi2html.init
./build_html.sh: execute category program: "sbcl" --script make-categories.lisp
maxima.texi:5: warning: @chapter should only appear at the beginning of a line
maxima.texi:5: warning: @chapter should not appear in @node
./Category-Help.texi:33: @ref reference to nonexistent node `Item:
Introduction/deffn/quad_control'
./include-maxima.texi:226: @menu reference to nonexistent node `Documentation
Categories'
Can't use an undefined value as an ARRAY reference at
/usr/local/share/texinfo/Texinfo/Structuring.pm line 932.
maxima.texi:5: warning: @chapter should only appear at the beginning of a line
maxima.texi:5: warning: @chapter should not appear in @node
./Category-Help.texi:33: @ref reference to nonexistent node `Item:
Introduction/deffn/quad_control'
./include-maxima.texi:226: @menu reference to nonexistent node `Documentation
Categories'
Can't use an undefined value as an ARRAY reference at
/usr/local/share/texinfo/Texinfo/Structuring.pm line 932.
./build_html.sh
./build_html.sh:
abs_srcdir=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info,
pwd=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info, language=en,
languagedir=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info,
texiinitfile=/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info/texi2html.init
./build_html.sh: execute category program: "sbcl" --script make-categories.lisp
maxima.texi:5: warning: @chapter should only appear at the beginning of a line
maxima.texi:5: warning: @chapter should not appear in @node
./Category-Help.texi:33: @ref reference to nonexistent node `Item:
Introduction/deffn/quad_control'
./include-maxima.texi:226: @menu reference to nonexistent node `Documentation
Categories'
Can't use an undefined value as an ARRAY reference at
/usr/local/share/texinfo/Texinfo/Structuring.pm line 932.
maxima.texi:5: warning: @chapter should only appear at the beginning of a line
maxima.texi:5: warning: @chapter should not appear in @node
./Category-Help.texi:33: @ref reference to nonexistent node `Item:
Introduction/deffn/quad_control'
./include-maxima.texi:226: @menu reference to nonexistent node `Documentation
Categories'
Can't use an undefined value as an ARRAY reference at
/usr/local/share/texinfo/Texinfo/Structuring.pm line 932.
/usr/bin/env perl ./create_index maxima_singlepage.html
Can't open maxima_singlepage.html at ./create_index line 14.
make[3]: *** [Makefile:1649: contents.hhc] Error 2
make[3]: Leaving directory
'/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info'
make[2]: *** [Makefile:1217: all-recursive] Error 1
make[2]: Leaving directory
'/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc/info'
make[1]: *** [Makefile:390: all-recursive] Error 1
make[1]: Leaving directory
'/home/dima/sagetrac-mirror/upstream/maxima-5.46.0/doc'
make: *** [Makefile:444: all-recursive] Error 1
I'd be happy to try to submit an upgrade to the texinfo port.
Cheers,
Dima
Re: Problems updating math/maxima
On 2023/05/26 14:23, Dima Pasechnik wrote:
> Is there any particular reason there is no modern makeinfo in OpenBSD?
> (the one that has texi2any executable, that's version 6.8 or newer)
That is available in the texinfo package ('g' prefix in the binary
names). It's a little old (6.5), no particular reason it's outdated afaik,
just nobody has done the work to update and test.
The version in base is a much older one from before the license changed
to GPLv3.
Re: Problems updating math/maxima
On Fri, May 26, 2023 at 12:23:18PM +0100, Stuart Henderson wrote: > On 2023/05/25 18:57, Jonathan Drews wrote: > > On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote: > > > > > > As well, less urgently, > > > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > > > (sbcl typically gives considerably better performance, and it's available > > > in ports, > > > I wrote the maxima discuss list and they affirm the performance boost > > that Dima mentions by using sbcl. However I get this make build error: > > > > Missing library for sbcl>=0.0 > > Fatal error > > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2385 > > sbcl is a standalone compiler binary + runtime files, there is no > library so you can't use LIB_DEPENDS+WANTLIB for it. would be > BUILD_DEPENDS and maybe also RUN_DEPENDS if needed. For what it is worth, I am able to build Maxima 5.46.0 with OpenBSD 7.3's sbcl the usual Maxima's way: ./configure --enable-sbcl && make - but only if I have modern makeinfo, version 6.8 or better. (same happens, actually, with ecl instead of sbcl, as well). With the standard ancient OpenBSD's makeinfo (part of texinfo), it fails at building docs (yes, one can of course use info and html docs shipped with Maxima instead) Is there any particular reason there is no modern makeinfo in OpenBSD? (the one that has texi2any executable, that's version 6.8 or newer) If it was the case it'd allow to drop several text2 ports, as they'd be all provided by texinfo. Dima
Re: Problems updating math/maxima
On Fri, May 26, 2023 at 12:23:18PM +0100, Stuart Henderson wrote: > On 2023/05/25 18:57, Jonathan Drews wrote: > > On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote: > > > > > > As well, less urgently, > > > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > > > (sbcl typically gives considerably better performance, and it's available > > > in ports, > > > I wrote the maxima discuss list and they affirm the performance boost > > that Dima mentions by using sbcl. However I get this make build error: > > > > Missing library for sbcl>=0.0 > > Fatal error > > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2385 > > sbcl is a standalone compiler binary + runtime files, there is no > library so you can't use LIB_DEPENDS+WANTLIB for it. would be > BUILD_DEPENDS and maybe also RUN_DEPENDS if needed. > Thanks so much Stuart: I made the additions to BUILD_DEPENDS and RUN_DEPENDS and maxima-5.46.0 successfully built. There is a significant improvement in performance: Compute a Taylor series to 825 terms using ecl: taylor(exp(-x/5)*sin(x), x,0,825)$ takes 20 seconds Compute a Taylor series to 825 terms using sbcl: taylor(exp(-x/5)*sin(x), x,0,825)$ takes 13 seconds Dima was right. Also one of the replies I got from the maxima discuss list: "My experience is that sbcl is way faster than ecl in practical use. "The Testsuite (at least +the share one) loads many packages and does try not to do any lengthy work with them. Loading +packages is not extremely fast in sbcl. The only drawback of sbcl is that sbcl runs out of +memory in places ECL doesn't." I am getting a lot of warnings when I do make build. I am going to capture thise in a file. They flew by to fast to read. > ecl is an embeddable lisp implementation that can run as a library > linked to by other programs. > > > > My Makefile: > > generally a cvs (or git) diff would be preferred instead of sending the > Makefile (please write it to a file and read it into your editor or send > it as an attachment so that tabs/wrapping are maintained). > I will figure out how to do this and do it from now on. -- Kind regards, Jonathan
Re: Problems updating math/maxima
On 2023/05/25 18:57, Jonathan Drews wrote: > On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote: > > > > As well, less urgently, > > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > > (sbcl typically gives considerably better performance, and it's available > > in ports, > I wrote the maxima discuss list and they affirm the performance boost > that Dima mentions by using sbcl. However I get this make build error: > > Missing library for sbcl>=0.0 > Fatal error > *** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2385 sbcl is a standalone compiler binary + runtime files, there is no library so you can't use LIB_DEPENDS+WANTLIB for it. would be BUILD_DEPENDS and maybe also RUN_DEPENDS if needed. ecl is an embeddable lisp implementation that can run as a library linked to by other programs. > My Makefile: generally a cvs (or git) diff would be preferred instead of sending the Makefile (please write it to a file and read it into your editor or send it as an attachment so that tabs/wrapping are maintained).
Re: Problems updating math/maxima
On Fri, May 26, 2023 at 11:38:00AM +0100, Stuart Henderson wrote: > Let's stick to one change at a time, and give the maintainer some > time to respond to the update diff. Good advice. I will do that. -- Kind regards, Jonathan :w
Re: Problems updating math/maxima
Let's stick to one change at a time, and give the maintainer some time to respond to the update diff. On 2023/05/25 15:23, Dima Pasechnik wrote: > On Thu, May 25, 2023 at 05:30:43AM -0600, Jonathan Drews wrote: > > On Thu, May 25, 2023 at 12:02:16PM +0100, Stuart Henderson wrote: > > > Here's an updated diff. The MODTCL_WISH_ADJ from the previous diff > > > is not actually needed, and Python is no longer needed. > > > > > > I don't use this software but the update looks good to me now. > > Can you also add building maxima.fas and installing it into > /usr/local/lib/ecl/ ? > Then one can use Maxima from any application which can load libecl.so > (e.g. from an application written in C). > If we ever port SageMath (sagemath.org) to OpenBSD (there is a port to > FreeBSD) > then maxima.fas will be necessary. > > As well, less urgently, > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > (sbcl typically gives considerably better performance, and it's available in > ports, > but embedding into applications is a different story, harder than with ecl). > > Best, > Dima > > > > > > > > Stuart: > > > > Thank you so much for all your help on this port. I will make the > > changes you have given here and test the port. > > > > > > --- > > Kind regards, > > Jonathan > > >
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote:
>
> As well, less urgently,
> shouldn't there be a way to get Maxima compiled with sbcl, only only ecl.
> (sbcl typically gives considerably better performance, and it's available in
> ports,
I wrote the maxima discuss list and they affirm the performance boost
that Dima mentions by using sbcl. However I get this make build error:
Missing library for sbcl>=0.0
Fatal error
*** Error 1 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2385
My Makefile:
COMMENT=GPL computer algebra system based on DOE
Macsyma
VERSION=5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES= math
HOMEPAGE= http://maxima.sourceforge.net/
MAINTAINER= Ingo Feinerer
# GPLv2+
PERMIT_PACKAGE= Yes
WANTLIB += c sbcl ffi gc gmp m pthread
MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
USE_GMAKE= yes
MODTCL_VERSION= 8.6
#MODPY_RUNDEP = No
BUILD_DEPENDS = print/texinfo
LIB_DEPENDS=lang/sbcl
RUN_DEPENDS = devel/desktop-file-utils \
misc/rlwrap \
misc/shared-mime-info
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
SUBST_VARS= VERSION
MAKE_FLAGS =MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
CONFIGURE_ARGS+=--enable-sbcl
CONFIGURE_STYLE=gnu
FLAVORS=no_x11
FLAVOR?=
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
MODTK_VERSION = 8.6
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
.endif
.include
What am I missing?
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 12:11:05PM -0600, Jonathan Drews wrote: > On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote: > > > > Can you also add building maxima.fas and installing it into > > /usr/local/lib/ecl/ ? > > Then one can use Maxima from any application which can load libecl.so > > (e.g. from an application written in C). > > If we ever port SageMath (sagemath.org) to OpenBSD (there is a port to > > FreeBSD) > > then maxima.fas will be necessary. > > > You are going to have to enlighten me as to what the *.fas extension > means. I did a web search and could only find this page which did not Well, see the ECL manual, Sect. 3.1.1.2: https://ecl.common-lisp.dev/static/manual/System-building.html#Native-FASL It's a basically a dynamic shared library. On Linux I can do $ ldd local/lib/ecl/maxima.fas linux-vdso.so.1 (0x7ffd17c88000) libecl.so.21.2 => /home/dimpase/work/software/sage/local/lib/libecl.so.21.2 (0x7cb5df721000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x7cb5df5c5000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x7cb5df3f1000) libgmp.so.10 => /lib/x86_64-linux-gnu/libgmp.so.10 (0x7cb5df37) libgc.so.1 => /lib/x86_64-linux-gnu/libgc.so.1 (0x7cb5df2f8000) libffi.so.7 => /lib/x86_64-linux-gnu/libffi.so.7 (0x7cb5df2ec000) libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x7cb5df2c8000) libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x7cb5df2c2000) /lib64/ld-linux-x86-64.so.2 (0x7cb5e06d2000) > give me much information. I looked at Sages web page and did not see > anything. Here is what I did find: > > https://archlinux.org/packages/extra/x86_64/maxima-fas/ > > Concerning xmaxima.fas, it's built! Look: > > $ cd /usr/ports/pobj/maxima-5.46.0/ > $ find . -type f -name "maxima.fas" > ./maxima-5.46.0/src/binary-ecl/maxima.fas > > there is a whole bunch of *.fas files in > /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/src/binary-ecl/ I think you don't need any of these, except maxima.fas > > I just have to figure out how to install them. I looked for configure > options but I did not see any. I don't think there is anything provided by Maxima build system in this respect (it's highly lisp-system dependent, after all). One has to copy maxima.fas into /usr/local/lib/ecl/ I suppose OpenBSD ports build system has a way to do this, so that it's correctly gets uninstalled, too. > Lete me test the basic funcionality of > maxima (diff, integrate, taylor etc.). Most of those functions have to > be tested. Afterwards, I will look into isntalling the *.fas files in > /usr/local/lib/ecl/. > > > As well, less urgently, > > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > > (sbcl typically gives considerably better performance, and it's available > > in ports, > > but embedding into applications is a different story, harder than with ecl). > > I suppose you are refering to Steel Bank Common Lisp: > > http://www.sbcl.org/ > > I will have to ask about this on the Maxima mail list. Give me some > time. There is a port: /usr/ports/lang/sbcl - yes, there is a port, as I wrote. I think for Maxima developers sbcl is the primary lisp system, and many Linux distros by default ship Maxima built with sbcl. Cheers, Dima
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 03:23:23PM +0200, Dima Pasechnik wrote: > > Can you also add building maxima.fas and installing it into > /usr/local/lib/ecl/ ? > Then one can use Maxima from any application which can load libecl.so > (e.g. from an application written in C). > If we ever port SageMath (sagemath.org) to OpenBSD (there is a port to > FreeBSD) > then maxima.fas will be necessary. > Hello Dima: You are going to have to enlighten me as to what the *.fas extension means. I did a web search and could only find this page which did not give me much information. I looked at Sages web page and did not see anything. Here is what I did find: https://archlinux.org/packages/extra/x86_64/maxima-fas/ Concerning xmaxima.fas, it's built! Look: $ cd /usr/ports/pobj/maxima-5.46.0/ $ find . -type f -name "maxima.fas" ./maxima-5.46.0/src/binary-ecl/maxima.fas there is a whole bunch of *.fas files in /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/src/binary-ecl/ I just have to figure out how to install them. I looked for configure options but I did not see any. Lete me test the basic funcionality of maxima (diff, integrate, taylor etc.). Most of those functions have to be tested. Afterwards, I will look into isntalling the *.fas files in /usr/local/lib/ecl/. > As well, less urgently, > shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. > (sbcl typically gives considerably better performance, and it's available in > ports, > but embedding into applications is a different story, harder than with ecl). I suppose you are refering to Steel Bank Common Lisp: http://www.sbcl.org/ I will have to ask about this on the Maxima mail list. Give me some time. There is a port: /usr/ports/lang/sbcl -- Kind regards, Jonathan
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 12:28:21PM +0100, Stuart Henderson wrote:
> On 2023/05/25 13:21, Theo Buehler wrote:
> > > Index: pkg/PLIST
> > > ===
> > > RCS file: /cvs/ports/math/maxima/pkg/PLIST,v
> > > retrieving revision 1.25
> > > diff -u -p -r1.25 PLIST
> > > --- pkg/PLIST 11 Mar 2022 19:36:19 - 1.25
> > > +++ pkg/PLIST 25 May 2023 11:00:08 -
> > > @@ -9,8 +9,9 @@ bin/xmaxima
> > > info/maxima-index.lisp
> > > info/maxima.info
> > > info/maxima.info-1
> > > -info/maxima.info-2
> > > +info/maxima.info${MODPY_BIN_SUFFIX}
> >
> > That will probably break packaging after the removal of the Python
> > dependency.
> >
>
> good catch :)
>
Stuart and Theo:
I compiled maxima with the Python entries removed and it
installed and runs ok. Here is the Makefile based on Suarts diff. I
did retain the references to tcl8.6.
COMMENT=GPL computer algebra system based on DOE
Macsyma
VERSION=5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES= math
HOMEPAGE= http://maxima.sourceforge.net/
MAINTAINER= Ingo Feinerer
# GPLv2+
PERMIT_PACKAGE= Yes
WANTLIB += c ecl ffi gc gmp m pthread
MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
USE_GMAKE= yes
MODTCL_VERSION= 8.6
BUILD_DEPENDS = print/texinfo
LIB_DEPENDS=lang/ecl
RUN_DEPENDS = devel/desktop-file-utils \
misc/rlwrap \
misc/shared-mime-info
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
SUBST_VARS= VERSION
MAKE_FLAGS =MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
CONFIGURE_ARGS+=--enable-ecl
CONFIGURE_STYLE=gnu
FLAVORS=no_x11
FLAVOR?=
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
MODTK_VERSION = 8.6
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
.endif
.include
Let me do some more testing
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 05:30:43AM -0600, Jonathan Drews wrote: > On Thu, May 25, 2023 at 12:02:16PM +0100, Stuart Henderson wrote: > > Here's an updated diff. The MODTCL_WISH_ADJ from the previous diff > > is not actually needed, and Python is no longer needed. > > > > I don't use this software but the update looks good to me now. Can you also add building maxima.fas and installing it into /usr/local/lib/ecl/ ? Then one can use Maxima from any application which can load libecl.so (e.g. from an application written in C). If we ever port SageMath (sagemath.org) to OpenBSD (there is a port to FreeBSD) then maxima.fas will be necessary. As well, less urgently, shouldn't there be a way to get Maxima compiled with sbcl, only only ecl. (sbcl typically gives considerably better performance, and it's available in ports, but embedding into applications is a different story, harder than with ecl). Best, Dima > > > > Stuart: > > Thank you so much for all your help on this port. I will make the > changes you have given here and test the port. > > > --- > Kind regards, > Jonathan >
Re: Problems updating math/maxima
On Thu, May 25, 2023 at 12:02:16PM +0100, Stuart Henderson wrote: > Here's an updated diff. The MODTCL_WISH_ADJ from the previous diff > is not actually needed, and Python is no longer needed. > > I don't use this software but the update looks good to me now. > Stuart: Thank you so much for all your help on this port. I will make the changes you have given here and test the port. --- Kind regards, Jonathan
Re: Problems updating math/maxima
On 2023/05/25 13:21, Theo Buehler wrote:
> > Index: pkg/PLIST
> > ===
> > RCS file: /cvs/ports/math/maxima/pkg/PLIST,v
> > retrieving revision 1.25
> > diff -u -p -r1.25 PLIST
> > --- pkg/PLIST 11 Mar 2022 19:36:19 - 1.25
> > +++ pkg/PLIST 25 May 2023 11:00:08 -
> > @@ -9,8 +9,9 @@ bin/xmaxima
> > info/maxima-index.lisp
> > info/maxima.info
> > info/maxima.info-1
> > -info/maxima.info-2
> > +info/maxima.info${MODPY_BIN_SUFFIX}
>
> That will probably break packaging after the removal of the Python dependency.
>
good catch :)
Re: Problems updating math/maxima
> Index: pkg/PLIST
> ===
> RCS file: /cvs/ports/math/maxima/pkg/PLIST,v
> retrieving revision 1.25
> diff -u -p -r1.25 PLIST
> --- pkg/PLIST 11 Mar 2022 19:36:19 - 1.25
> +++ pkg/PLIST 25 May 2023 11:00:08 -
> @@ -9,8 +9,9 @@ bin/xmaxima
> info/maxima-index.lisp
> info/maxima.info
> info/maxima.info-1
> -info/maxima.info-2
> +info/maxima.info${MODPY_BIN_SUFFIX}
That will probably break packaging after the removal of the Python dependency.
Re: Problems updating math/maxima
Here's an updated diff. The MODTCL_WISH_ADJ from the previous diff
is not actually needed, and Python is no longer needed.
I don't use this software but the update looks good to me now.
Index: Makefile
===
RCS file: /cvs/ports/math/maxima/Makefile,v
retrieving revision 1.50
diff -u -p -r1.50 Makefile
--- Makefile11 Mar 2022 19:36:19 - 1.50
+++ Makefile25 May 2023 11:00:08 -
@@ -1,6 +1,6 @@
COMMENT= GPL computer algebra system based on DOE Macsyma
-VERSION= 5.44.0
+VERSION= 5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES=math
@@ -16,13 +16,6 @@ MASTER_SITES=${MASTER_SITE_SOURCEFORGE
USE_GMAKE= yes
-# Python is not needed for building but configure fails without it.
-# Instead of patching just add the lang/python dependency.
-# http://sourceforge.net/p/maxima/mailman/message/34462954/
-MODULES = lang/python
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
-MODPY_RUNDEP = No
-
BUILD_DEPENDS =print/texinfo
LIB_DEPENDS= lang/ecl
RUN_DEPENDS = devel/desktop-file-utils \
@@ -42,7 +35,7 @@ FLAVOR?=
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
+MODTK_VERSION =8.6
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS += ${MODTK_BUILD_DEPENDS}
Index: distinfo
===
RCS file: /cvs/ports/math/maxima/distinfo,v
retrieving revision 1.24
diff -u -p -r1.24 distinfo
--- distinfo9 Jun 2020 19:56:59 - 1.24
+++ distinfo25 May 2023 11:00:08 -
@@ -1,2 +1,2 @@
-SHA256 (maxima-5.44.0.tar.gz) = 2T9eSMTa+PCF1gnLPHsL3zQsZn/QTPdQyEZCaHTJ0uw=
-SIZE (maxima-5.44.0.tar.gz) = 34976354
+SHA256 (maxima-5.46.0.tar.gz) = c5Dwa0jaZckDPosvYpuXi5AFZFSlQCLbfecOIiWqiwc=
+SIZE (maxima-5.46.0.tar.gz) = 47492457
Index: pkg/PLIST
===
RCS file: /cvs/ports/math/maxima/pkg/PLIST,v
retrieving revision 1.25
diff -u -p -r1.25 PLIST
--- pkg/PLIST 11 Mar 2022 19:36:19 - 1.25
+++ pkg/PLIST 25 May 2023 11:00:08 -
@@ -9,8 +9,9 @@ bin/xmaxima
info/maxima-index.lisp
info/maxima.info
info/maxima.info-1
-info/maxima.info-2
+info/maxima.info${MODPY_BIN_SUFFIX}
info/maxima.info-3
+@info info/symplectic_ode.info
@info info/xmaxima.info
lib/maxima/
lib/maxima/${VERSION}/
@@ -41,6 +42,7 @@ share/emacs/site-lisp/imaxima.el
share/emacs/site-lisp/imaxima.lisp
share/emacs/site-lisp/maxima-font-lock.el
share/emacs/site-lisp/maxima.el
+share/emacs/site-lisp/mylatex.ltx.el
share/emacs/site-lisp/setup-imaxima-imath.el
share/emacs/site-lisp/smart-complete.el
share/emacs/site-lisp/sshell.el
@@ -67,152 +69,152 @@ share/maxima/${VERSION}/doc/contributors
share/maxima/${VERSION}/doc/html/
share/maxima/${VERSION}/doc/html/contents.hhc
share/maxima/${VERSION}/doc/html/figures/
-share/maxima/${VERSION}/doc/html/figures/bodegain1.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain2.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain3.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain4.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain5.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain6.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain7.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain8.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase1.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase2.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase3.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase4.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase5.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase6.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase7.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase8.gif
-share/maxima/${VERSION}/doc/html/figures/bodephase9.gif
+share/maxima/${VERSION}/doc/html/figures/bodegain1.png
+share/maxima/${VERSION}/doc/html/figures/bodegain2.png
+share/maxima/${VERSION}/doc/html/figures/bodegain3.png
+share/maxima/${VERSION}/doc/html/figures/bodegain4.png
+share/maxima/${VERSION}/doc/html/figures/bodegain5.png
+share/maxima/${VERSION}/doc/html/figures/bodegain6.png
+share/maxima/${VERSION}/doc/html/figures/bodegain7.png
+share/maxima/${VERSION}/doc/html/figures/bodegain8.png
+share/maxima/${VERSION}/doc/html/figures/bodephase1.png
+share/maxima/${VERSION}/doc/html/figures/bodephase2.png
+share/maxima/${VERSION}/doc/html/figures/bodephase3.png
+share/maxima/${VERSION}/doc/html/figures/bodephase4.png
+share/maxima/${VERSION}/doc/html/figures/bodephase5.png
+share/maxima/${VERSION}/doc/html/figures/bodephase6.png
+share/maxima/${VERSION}/doc/html/figures/bodephase7.png
+share/maxima/${VERSION}/doc/html/figures/bodephase8.png
+share/maxima/${VERSION}/d
Re: Problems updating math/maxima
Thank you so much Stuart: I downloaded maxima-5.46.0.tar.gz and untarred it in /usr. I followed the instructions at: https://www.openbsd.org/faq/ports/ports.html#PortsFetch I'll blow away /usr/ports and reinstall. Again thanks so much On Wed, May 24, 2023, at 07:37, Stuart Henderson wrote: > On 2023/05/24 07:18, Jonathan Drews wrote: > > I almost forgot. I captured this from the beginning of the # make build: > > > > make install > > ===> maxima-5.46.0 depends on: desktop-file-utils-* -> > > desktop-file-utils-0.26 > > ===> maxima-5.46.0 depends on: rlwrap-* -> rlwrap-0.45.2 > > ===> maxima-5.46.0 depends on: shared-mime-info-* -> shared-mime-info-2.2 > > ===> maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6 > > ===> maxima-5.46.0 depends on: gnuplot-* -> gnuplot-5.2.7p1 > > ===> Verifying specs: c ecl ffi gc gmp m pthread > > ===> found c.97.0 ecl.8.0 ffi.2.0 gc.5.0 gmp.11.0 m.10.1 pthread.27.0 > > ===> Installing maxima-5.46.0 from /usr/ports/packages/amd64/all/ > > > > Do you see that line : > > maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6 > > I guess you are working on a copy of the port under mystuff or similar > rather than the main tree. I suggest not doing that, just use mystuff >
Re: Problems updating math/maxima
On 2023/05/24 07:18, Jonathan Drews wrote:
> I almost forgot. I captured this from the beginning of the # make build:
>
> make install
> ===> maxima-5.46.0 depends on: desktop-file-utils-* -> desktop-file-utils-0.26
> ===> maxima-5.46.0 depends on: rlwrap-* -> rlwrap-0.45.2
> ===> maxima-5.46.0 depends on: shared-mime-info-* -> shared-mime-info-2.2
> ===> maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6
> ===> maxima-5.46.0 depends on: gnuplot-* -> gnuplot-5.2.7p1
> ===> Verifying specs: c ecl ffi gc gmp m pthread
> ===> found c.97.0 ecl.8.0 ffi.2.0 gc.5.0 gmp.11.0 m.10.1 pthread.27.0
> ===> Installing maxima-5.46.0 from /usr/ports/packages/amd64/all/
>
> Do you see that line :
> maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6
I guess you are working on a copy of the port under mystuff or similar
rather than the main tree. I suggest not doing that, just use mystuff
for new things, otherwise you need to override the default PORTSDIR_PATH
and then take care that you remove files from mystuff when you're no
longer working on them.
It does get set correctly for me, and the exec wish line is set to the
expected wish8.6.
> What is that doing in there when I have overridden it with:
> RUN_DEPENDS = devel/desktop-file-utils \
> misc/rlwrap \
> misc/shared-mime-info \
> lang/tcl/8.6
That's already setup by tcl.port.mk/tk.port.mk and added in the port
Makefile (MODTCL_RUN_DEPENDS etc) so you don't need to duplicate it.
This worked for me, though perhaps the pre-build is not actually needed
if the other problem resulting in it picking up 8.5 is fixed (not tested,
build/install are pretty slow on this).
Index: Makefile
===
RCS file: /cvs/ports/math/maxima/Makefile,v
retrieving revision 1.50
diff -u -p -r1.50 Makefile
--- Makefile11 Mar 2022 19:36:19 - 1.50
+++ Makefile24 May 2023 13:29:11 -
@@ -1,6 +1,6 @@
COMMENT= GPL computer algebra system based on DOE Macsyma
-VERSION= 5.44.0
+VERSION= 5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES=math
@@ -20,6 +20,7 @@ USE_GMAKE=yes
# Instead of patching just add the lang/python dependency.
# http://sourceforge.net/p/maxima/mailman/message/34462954/
MODULES = lang/python
+# XXX can this be removed i.e. move to py3?
MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
MODPY_RUNDEP = No
@@ -29,6 +30,9 @@ RUN_DEPENDS = devel/desktop-file-utils
misc/rlwrap \
misc/shared-mime-info
+pre-build:
+ ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
+
SUBST_VARS=VERSION
MAKE_FLAGS = MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
@@ -42,11 +46,12 @@ FLAVOR?=
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
-MODPY_VERSION =${MODPY_DEFAULT_VERSION_2}
+MODTK_VERSION =8.6
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS += ${MODTK_BUILD_DEPENDS}
CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
.endif
+
.include
Index: distinfo
===
RCS file: /cvs/ports/math/maxima/distinfo,v
retrieving revision 1.24
diff -u -p -r1.24 distinfo
--- distinfo9 Jun 2020 19:56:59 - 1.24
+++ distinfo24 May 2023 13:29:11 -
@@ -1,2 +1,2 @@
-SHA256 (maxima-5.44.0.tar.gz) = 2T9eSMTa+PCF1gnLPHsL3zQsZn/QTPdQyEZCaHTJ0uw=
-SIZE (maxima-5.44.0.tar.gz) = 34976354
+SHA256 (maxima-5.46.0.tar.gz) = c5Dwa0jaZckDPosvYpuXi5AFZFSlQCLbfecOIiWqiwc=
+SIZE (maxima-5.46.0.tar.gz) = 47492457
Index: pkg/PLIST
===
RCS file: /cvs/ports/math/maxima/pkg/PLIST,v
retrieving revision 1.25
diff -u -p -r1.25 PLIST
--- pkg/PLIST 11 Mar 2022 19:36:19 - 1.25
+++ pkg/PLIST 24 May 2023 13:29:11 -
@@ -9,8 +9,9 @@ bin/xmaxima
info/maxima-index.lisp
info/maxima.info
info/maxima.info-1
-info/maxima.info-2
+info/maxima.info${MODPY_BIN_SUFFIX}
info/maxima.info-3
+@info info/symplectic_ode.info
@info info/xmaxima.info
lib/maxima/
lib/maxima/${VERSION}/
@@ -41,6 +42,7 @@ share/emacs/site-lisp/imaxima.el
share/emacs/site-lisp/imaxima.lisp
share/emacs/site-lisp/maxima-font-lock.el
share/emacs/site-lisp/maxima.el
+share/emacs/site-lisp/mylatex.ltx.el
share/emacs/site-lisp/setup-imaxima-imath.el
share/emacs/site-lisp/smart-complete.el
share/emacs/site-lisp/sshell.el
@@ -67,152 +69,152 @@ share/maxima/${VERSION}/doc/contributors
share/maxima/${VERSION}/doc/html/
share/maxima/${VERSION}/doc/html/contents.hhc
share/maxima/${VERSION}/doc/html/figures/
-share/maxima/${VERSION}/doc/html/figures/bodegain1.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain2.gif
-share/maxima/${VERSION}/doc/html/figures/bodegain3.gif
-share/maxima/
Re: Problems updating math/maxima
I almost forgot. I captured this from the beginning of the # make build: make install ===> maxima-5.46.0 depends on: desktop-file-utils-* -> desktop-file-utils-0.26 ===> maxima-5.46.0 depends on: rlwrap-* -> rlwrap-0.45.2 ===> maxima-5.46.0 depends on: shared-mime-info-* -> shared-mime-info-2.2 ===> maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6 ===> maxima-5.46.0 depends on: gnuplot-* -> gnuplot-5.2.7p1 ===> Verifying specs: c ecl ffi gc gmp m pthread ===> found c.97.0 ecl.8.0 ffi.2.0 gc.5.0 gmp.11.0 m.10.1 pthread.27.0 ===> Installing maxima-5.46.0 from /usr/ports/packages/amd64/all/ Do you see that line : maxima-5.46.0 depends on: tcl-* -> tcl-8.5.19p6 What is that doing in there when I have overridden it with: RUN_DEPENDS = devel/desktop-file-utils \ misc/rlwrap \ misc/shared-mime-info \ lang/tcl/8.6 On Wed, May 24, 2023, at 07:03, Jonathan Drews wrote: > Stuart and others: > > Here is the thing. The variables are set! Look: > # make show=MODTCL_RUN_DEPENDS > tcl->=8.6,<8.7:lang/tcl/8.6 > # make show=MODTK_RUN_DEPENDS > tk->=8.6,<8.7:x11/tk/8.6 tcl->=8.6,<8.7:lang/tcl/8.6 > > and the only change is: > # diff /usr/local/bin/xmaxima > /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima > 3c3 > < exec /usr/local/bin/wish8.5 "$0" -- "$@" > --- > > exec /usr/local/bin/wish8.6 "$0" -- "$@" > > I sent the tarball to Stuart only and not to the list. > > On Wed, May 24, 2023, at 06:40, Stuart Henderson wrote: > > On 2023/05/24 13:40, Stuart Henderson wrote: > > > please send a tar.gz of the whole port. > > > >
Re: Problems updating math/maxima
Stuart and others: Here is the thing. The variables are set! Look: # make show=MODTCL_RUN_DEPENDS tcl->=8.6,<8.7:lang/tcl/8.6 # make show=MODTK_RUN_DEPENDS tk->=8.6,<8.7:x11/tk/8.6 tcl->=8.6,<8.7:lang/tcl/8.6 and the only change is: # diff /usr/local/bin/xmaxima /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima 3c3 < exec /usr/local/bin/wish8.5 "$0" -- "$@" --- > exec /usr/local/bin/wish8.6 "$0" -- "$@" I sent the tarball to Stuart only and not to the list. On Wed, May 24, 2023, at 06:40, Stuart Henderson wrote: > On 2023/05/24 13:40, Stuart Henderson wrote: > > please send a tar.gz of the whole port. >
Re: Problems updating math/maxima
On 2023/05/24 13:40, Stuart Henderson wrote:
> please send a tar.gz of the whole port.
...or a git/cvs diff.
>
> On 2023/05/24 06:28, Jonathan Drews wrote:
> >
> >
> > On Tue, May 23, 2023, at 06:03, Jonathan Drews wrote:
> > > On Tue, May 23, 2023 at 03:41:05AM -0600, Jonathan Drews wrote:
> > > > On Mon, May 22, 2023 at 07:43:20PM +0100, Stuart Henderson wrote:
> > > > > On 2023/05/22 09:11, Jonathan Drews wrote:
> > > > > > My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> > > > > > The port: math/maxima
> > > > > >
> > > > > >
> > > > > > and the xmaxima shell script says requires tk8.6:
> > > > > >
> > > > > > #!/bin/sh
> > > > > > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > > > > > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> > > > > >
> > > > > > package require Tk 8.6
> > > > >
> > > > > you have the wrong wish version in the exec line. this is usually
> > > > > fixed
> > > > > up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
> > > > > stage. maybe along the lines of
> > > > >
> > > > > pre-build:
> > > > > ${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
> > >
> > Folks this is my latest Makefile:
> >
> > COMMENT=GPL computer algebra system based on DOE Macsyma
> >
> > VERSION=5.46.0
> > DISTNAME= maxima-${VERSION}
> > CATEGORIES= math
> >
> > HOMEPAGE= http://maxima.sourceforge.net/
> > MAINTAINER= Ingo Feinerer
> >
> > # GPLv2+
> > PERMIT_PACKAGE= Yes
> >
> > WANTLIB += c ecl ffi gc gmp m pthread
> >
> > MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
> >
> > USE_GMAKE= yes
> >
> > # Python is not needed for building but configure fails without it.
> > # Instead of patching just add the lang/python dependency.
> > # http://sourceforge.net/p/maxima/mailman/message/34462954/
> > MODULES = lang/python
> > MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
> > MODPY_RUNDEP = No
> >
> > BUILD_DEPENDS = print/texinfo \
> > lang/tcl/8.6
> > LIB_DEPENDS=lang/ecl \
> > lang/tcl/8.6
> > RUN_DEPENDS = devel/desktop-file-utils \
> > misc/rlwrap \
> > misc/shared-mime-info \
> > lang/tcl/8.6
> >
> > pre-build:
> > ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
> >
> > SUBST_VARS= VERSION
> >
> > MAKE_FLAGS =MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
> > TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
> >
> > CONFIGURE_ARGS+=--enable-ecl
> > CONFIGURE_STYLE=gnu
> >
> > FLAVORS=no_x11
> > FLAVOR?=
> >
> > .if !${FLAVOR:Mno_x11}
> > MODULES+= x11/tk
> > MODTK_VERSION = 8.6
> > MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
> > RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
> > math/gnuplot
> > BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
> > CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
> > .endif
> >
> > It is still not replacing wish 8.5 with wish8.6:
> > # cat /usr/local/bin/xmaxima | head
> > #!/bin/sh
> > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> >
> > package require Tk 8.6
> >
> > BUT, the file in
> > /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima
> > *IS* getting replaced.
> >
> > # cat
> > /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima |
> > head
> > #!/bin/sh
> > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > exec /usr/local/bin/wish8.6 "$0" -- "$@"
> >
> > package require Tk 8.6
> >
> > What am I missing here?
> >
> > --
> > Kind regards,
> > Jonathan
> >
Re: Problems updating math/maxima
please send a tar.gz of the whole port.
On 2023/05/24 06:28, Jonathan Drews wrote:
>
>
> On Tue, May 23, 2023, at 06:03, Jonathan Drews wrote:
> > On Tue, May 23, 2023 at 03:41:05AM -0600, Jonathan Drews wrote:
> > > On Mon, May 22, 2023 at 07:43:20PM +0100, Stuart Henderson wrote:
> > > > On 2023/05/22 09:11, Jonathan Drews wrote:
> > > > > My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> > > > > The port: math/maxima
> > > > >
> > > > >
> > > > > and the xmaxima shell script says requires tk8.6:
> > > > >
> > > > > #!/bin/sh
> > > > > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > > > > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> > > > >
> > > > > package require Tk 8.6
> > > >
> > > > you have the wrong wish version in the exec line. this is usually fixed
> > > > up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
> > > > stage. maybe along the lines of
> > > >
> > > > pre-build:
> > > > ${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
> >
> Folks this is my latest Makefile:
>
> COMMENT=GPL computer algebra system based on DOE Macsyma
>
> VERSION=5.46.0
> DISTNAME= maxima-${VERSION}
> CATEGORIES= math
>
> HOMEPAGE= http://maxima.sourceforge.net/
> MAINTAINER= Ingo Feinerer
>
> # GPLv2+
> PERMIT_PACKAGE= Yes
>
> WANTLIB += c ecl ffi gc gmp m pthread
>
> MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
>
> USE_GMAKE= yes
>
> # Python is not needed for building but configure fails without it.
> # Instead of patching just add the lang/python dependency.
> # http://sourceforge.net/p/maxima/mailman/message/34462954/
> MODULES = lang/python
> MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
> MODPY_RUNDEP = No
>
> BUILD_DEPENDS = print/texinfo \
> lang/tcl/8.6
> LIB_DEPENDS=lang/ecl \
> lang/tcl/8.6
> RUN_DEPENDS = devel/desktop-file-utils \
> misc/rlwrap \
> misc/shared-mime-info \
> lang/tcl/8.6
>
> pre-build:
> ${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
>
> SUBST_VARS= VERSION
>
> MAKE_FLAGS =MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
> TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
>
> CONFIGURE_ARGS+=--enable-ecl
> CONFIGURE_STYLE=gnu
>
> FLAVORS=no_x11
> FLAVOR?=
>
> .if !${FLAVOR:Mno_x11}
> MODULES+= x11/tk
> MODTK_VERSION = 8.6
> MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
> RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
> math/gnuplot
> BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
> CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
> .endif
>
> It is still not replacing wish 8.5 with wish8.6:
> # cat /usr/local/bin/xmaxima | head
> #!/bin/sh
> # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> exec /usr/local/bin/wish8.5 "$0" -- "$@"
>
> package require Tk 8.6
>
> BUT, the file in
> /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima
> *IS* getting replaced.
>
> # cat /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima
> | head
> #!/bin/sh
> # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> exec /usr/local/bin/wish8.6 "$0" -- "$@"
>
> package require Tk 8.6
>
> What am I missing here?
>
> --
> Kind regards,
> Jonathan
>
Re: Problems updating math/maxima
On Tue, May 23, 2023, at 06:03, Jonathan Drews wrote:
> On Tue, May 23, 2023 at 03:41:05AM -0600, Jonathan Drews wrote:
> > On Mon, May 22, 2023 at 07:43:20PM +0100, Stuart Henderson wrote:
> > > On 2023/05/22 09:11, Jonathan Drews wrote:
> > > > My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> > > > The port: math/maxima
> > > >
> > > >
> > > > and the xmaxima shell script says requires tk8.6:
> > > >
> > > > #!/bin/sh
> > > > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > > > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> > > >
> > > > package require Tk 8.6
> > >
> > > you have the wrong wish version in the exec line. this is usually fixed
> > > up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
> > > stage. maybe along the lines of
> > >
> > > pre-build:
> > > ${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
>
Folks this is my latest Makefile:
COMMENT=GPL computer algebra system based on DOE Macsyma
VERSION=5.46.0
DISTNAME= maxima-${VERSION}
CATEGORIES= math
HOMEPAGE= http://maxima.sourceforge.net/
MAINTAINER= Ingo Feinerer
# GPLv2+
PERMIT_PACKAGE= Yes
WANTLIB += c ecl ffi gc gmp m pthread
MASTER_SITES= ${MASTER_SITE_SOURCEFORGE:=maxima/}
USE_GMAKE= yes
# Python is not needed for building but configure fails without it.
# Instead of patching just add the lang/python dependency.
# http://sourceforge.net/p/maxima/mailman/message/34462954/
MODULES = lang/python
MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
MODPY_RUNDEP = No
BUILD_DEPENDS = print/texinfo \
lang/tcl/8.6
LIB_DEPENDS=lang/ecl \
lang/tcl/8.6
RUN_DEPENDS = devel/desktop-file-utils \
misc/rlwrap \
misc/shared-mime-info \
lang/tcl/8.6
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/interfaces/xmaxima/xmaxima
SUBST_VARS= VERSION
MAKE_FLAGS =MAKEINFO="${LOCALBASE}/bin/gtexi2any" \
TEXI2DVI="${LOCALBASE}/bin/gtexi2dvi"
CONFIGURE_ARGS+=--enable-ecl
CONFIGURE_STYLE=gnu
FLAVORS=no_x11
FLAVOR?=
.if !${FLAVOR:Mno_x11}
MODULES+= x11/tk
MODTK_VERSION = 8.6
MODPY_VERSION = ${MODPY_DEFAULT_VERSION_2}
RUN_DEPENDS += ${MODTK_RUN_DEPENDS} \
math/gnuplot
BUILD_DEPENDS +=${MODTK_BUILD_DEPENDS}
CONFIGURE_ARGS += --with-wish=${MODTK_BIN}
.endif
It is still not replacing wish 8.5 with wish8.6:
# cat /usr/local/bin/xmaxima | head
#!/bin/sh
# -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
exec /usr/local/bin/wish8.5 "$0" -- "$@"
package require Tk 8.6
BUT, the file in
/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima
*IS* getting replaced.
# cat /usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima |
head
#!/bin/sh
# -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
exec /usr/local/bin/wish8.6 "$0" -- "$@"
package require Tk 8.6
What am I missing here?
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
On Tue, May 23, 2023 at 03:41:05AM -0600, Jonathan Drews wrote:
> On Mon, May 22, 2023 at 07:43:20PM +0100, Stuart Henderson wrote:
> > On 2023/05/22 09:11, Jonathan Drews wrote:
> > > My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> > > The port: math/maxima
> > >
> > >
> > > and the xmaxima shell script says requires tk8.6:
> > >
> > > #!/bin/sh
> > > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> > >
> > > package require Tk 8.6
> >
> > you have the wrong wish version in the exec line. this is usually fixed
> > up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
> > stage. maybe along the lines of
> >
> > pre-build:
> > ${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
> >
>
> Stuart and others:
>
> I added the following to the end of my Makefile:
>
> pre-build:
> ${MODTCL_WISH_ADJ} ${WRKSRC}/wish8.6
I made a correction:
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/xmaxima
and the correct wish shows up in :
/usr/ports/pobj/maxima-5.46.0/maxima-5.46.0/interfaces/xmaxima/xmaxima
#!/bin/sh
# -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
exec /usr/local/bin/wish8.6 "$0" -- "$@"
but in /usr/local/bin/xmaxima I still get wish8.5
cat /usr/local/bin/xmaxima | head
#!/bin/sh
# -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
exec /usr/local/bin/wish8.5 "$0" -- "$@"
package require Tk 8.6
Any advice would be greatly appreciated.
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
On Mon, May 22, 2023 at 07:43:20PM +0100, Stuart Henderson wrote:
> On 2023/05/22 09:11, Jonathan Drews wrote:
> > My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> > The port: math/maxima
> >
> >
> > and the xmaxima shell script says requires tk8.6:
> >
> > #!/bin/sh
> > # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> > exec /usr/local/bin/wish8.5 "$0" -- "$@"
> >
> > package require Tk 8.6
>
> you have the wrong wish version in the exec line. this is usually fixed
> up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
> stage. maybe along the lines of
>
> pre-build:
> ${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
>
Stuart and others:
I added the following to the end of my Makefile:
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/wish8.6
but I still get this error.:
$ xmaxima
Error in startup script: version conflict for package "Tk": have
8.5.19, need 8.6
while executing
"package require Tk 8.6"
(file "/usr/local/bin/xmaxima" line 5)
If I change line 5 in /usr/local/bin/xmaxima:
#!/bin/sh
# -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
exec /usr/local/bin/wish8.5 "$0" -- "$@"
package require Tk 8.6
to
exec /usr/local/bin/wish8.6 "$0" -- "$@"
then xmaxoma works. I am at loss as to what I am missing here.
--
Kind regards,
Jonathan
Re: Problems updating math/maxima
On 2023/05/22 09:11, Jonathan Drews wrote:
> My system: OpenBSD 7.3 GENERIC.MP#1125 amd64
> The port: math/maxima
>
> Hi Folks:
>
> This port has not been updated for a while. I got permission from the
> current maitainer to update it. I am having a little problem after
> making the necessary changes to the Makefile. xmaxima will not
> display. I get the following error when I try to run it:
>
> $ xmaxima
>
> Error in startup script: version conflict for package "Tk": have
> 8.5.19, need 8.6
> while executing
> "package require Tk 8.6"
> (file "/usr/local/bin/xmaxima" line 5)
>
> and the xmaxima shell script says requires tk8.6:
>
> #!/bin/sh
> # -*- mode: tcl; fill-column: 75; tab-width: 8 -*- \
> exec /usr/local/bin/wish8.5 "$0" -- "$@"
>
> package require Tk 8.6
you have the wrong wish version in the exec line. this is usually fixed
up by passing the file to ${MODTCL_WISH_ADJ} in some build/configure
stage. maybe along the lines of
pre-build:
${MODTCL_WISH_ADJ} ${WRKSRC}/path/to/file
