Re: [remove-or-fix] x11/ion

2021-09-14 Thread Kurt Mosiejczuk
On Tue, Sep 14, 2021 at 07:44:52PM +0200, Stefan Hagen wrote:

> I found the cause and also why it's only happening here. It has nothing 
> to do with luajit. This was an accidental correlation.

> Setting MAKE_JOBS>1 breaks it. When not using MAKE_JOBS, it works every
> time.

> According to README.internals MAKE_JOBS is used by dpb, so I wonder why
> this didn't come up in bulk builds.

dpb won't use more than one job unless DPB_PROPERTIES has parallel or 
parallel2.

> Let's not touch it any further, besides the mandoc bit you committed
> already.

> Best regards,
> Stefan

--Kurt



Re: [remove-or-fix] x11/ion

2021-09-14 Thread Stefan Hagen
Stuart Henderson wrote:
> On 2021/09/14 14:06, Stefan Hagen wrote:
> > Stuart Henderson wrote:
> > > On 2021/09/14 07:43, Stefan Hagen wrote:
> > > > X11/ion is currently broken.
> > > 
> > > Hmm? It does build, though it does either need a BUILD_DEPENDS on groff
> > > or your change for mandoc to produce consistent results (i.e. including
> > > the manpage contents in welcome.txt etc).
> 
> I've committed an alternative for the nroff fix. (Generally prefer to
> use patch rather than sed unless there are a large number of files).
> 
> > > $ ls -l */ion-*
> > > -rw-r--r--  1 root  wheel  760147 Sep 13 07:16 aarch64/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  846233 Sep 13 07:03 amd64/ion-20090110p4.tgz
> > > -rw-r--r--  2 root  wheel  756225 Jun  2 08:13 arm/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  760814 Sep  6 07:23 i386/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  869767 Aug 22 07:40 mips64/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  843277 Jun 25 07:55 mips64el/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  849572 Sep  9 09:59 powerpc/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  835022 Sep 10 07:12 
> > > powerpc64/ion-20090110p4.tgz
> > > -rw-r--r--  1 root  wheel  810101 Sep 11 09:30 sparc64/ion-20090110p4.tgz
> > > 
> > > What problem are you seeing with the various exports.h files? Is there
> > > something out of the ordinary with your lua installation?
> > 
> > You're right. Once luajit is installed, it doesn't compile anymore.
> > If I deinstall luajit, it works again.
> > 
> > With the diff it works in both cases.
> 
> Still strange as I don't see that here, and don't recall ever seeing a
> problem in bulk builds with this. Have just tested with both in-tree
> luajit and a local build of luajit 2.1. And I don't see any mechanism in
> system.mk that ought to trigger the problem by having luajit installed
> really..
> 
> I'm not seeing a good reason to remove it at this point, even with a
> successor.
> 
> > If we want to keep it, I'll look closer to find the root cause for the
> > behavior change.
> 
> > AFAIR it worked even with luajit a few months ago.
> > 
> > This is the error I'm seeing:
> > [...]
> > Scanning mplex.c for exports.
> 
> The command line running libextl-mkexports might give clues. Here's the
> full build log from mine. Typical invocations are like this:
> 
> `which lua51` ../libextl/libextl-mkexports -module mod_tiling -o exports.c -h 
> exports.h \
> main.c tiling.c placement.c split.c split-stdisp.c splitfloat.c panehandle.c 
> ops.c

I found the cause and also why it's only happening here. It has nothing 
to do with luajit. This was an accidental correlation.

Setting MAKE_JOBS>1 breaks it. When not using MAKE_JOBS, it works every
time.

According to README.internals MAKE_JOBS is used by dpb, so I wonder why
this didn't come up in bulk builds.

Let's not touch it any further, besides the mandoc bit you committed
already.

Best regards,
Stefan



Re: [remove-or-fix] x11/ion

2021-09-14 Thread Stuart Henderson
On 2021/09/14 14:06, Stefan Hagen wrote:
> Stuart Henderson wrote:
> > On 2021/09/14 07:43, Stefan Hagen wrote:
> > > X11/ion is currently broken.
> > 
> > Hmm? It does build, though it does either need a BUILD_DEPENDS on groff
> > or your change for mandoc to produce consistent results (i.e. including
> > the manpage contents in welcome.txt etc).

I've committed an alternative for the nroff fix. (Generally prefer to
use patch rather than sed unless there are a large number of files).

> > $ ls -l */ion-*
> > -rw-r--r--  1 root  wheel  760147 Sep 13 07:16 aarch64/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  846233 Sep 13 07:03 amd64/ion-20090110p4.tgz
> > -rw-r--r--  2 root  wheel  756225 Jun  2 08:13 arm/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  760814 Sep  6 07:23 i386/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  869767 Aug 22 07:40 mips64/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  843277 Jun 25 07:55 mips64el/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  849572 Sep  9 09:59 powerpc/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  835022 Sep 10 07:12 powerpc64/ion-20090110p4.tgz
> > -rw-r--r--  1 root  wheel  810101 Sep 11 09:30 sparc64/ion-20090110p4.tgz
> > 
> > What problem are you seeing with the various exports.h files? Is there
> > something out of the ordinary with your lua installation?
> 
> You're right. Once luajit is installed, it doesn't compile anymore.
> If I deinstall luajit, it works again.
> 
> With the diff it works in both cases.

Still strange as I don't see that here, and don't recall ever seeing a
problem in bulk builds with this. Have just tested with both in-tree
luajit and a local build of luajit 2.1. And I don't see any mechanism in
system.mk that ought to trigger the problem by having luajit installed
really..

I'm not seeing a good reason to remove it at this point, even with a
successor.

> If we want to keep it, I'll look closer to find the root cause for the
> behavior change.

> AFAIR it worked even with luajit a few months ago.
> 
> This is the error I'm seeing:
> [...]
> Scanning mplex.c for exports.

The command line running libextl-mkexports might give clues. Here's the
full build log from mine. Typical invocations are like this:

`which lua51` ../libextl/libextl-mkexports -module mod_tiling -o exports.c -h 
exports.h \
main.c tiling.c placement.c split.c split-stdisp.c splitfloat.c panehandle.c 
ops.c




+++ Tue Sep 14 13:20:57 BST 2021
===>  Checking files for ion-20090110p5
`/y/Download/ftp/pub/OpenBSD/distfiles/ion-3-20090110.tar.gz' is up to date.
>> (SHA256) ion-3-20090110.tar.gz: OK
===>  Extracting for ion-20090110p5
===>  Patching for ion-20090110p5
===>   Applying OpenBSD patch patch-etc_cfg_ioncore_lua
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--
|$OpenBSD: patch-etc_cfg_ioncore_lua,v 1.8 2013/04/29 12:40:49 gonzalo Exp $
|--- etc/cfg_ioncore.lua.orig   Fri Jan  9 20:19:49 2009
|+++ etc/cfg_ioncore.luaMon Apr  8 11:20:13 2013
--
Patching file etc/cfg_ioncore.lua using Plan A...
Hunk #1 succeeded at 65.
Hunk #2 succeeded at 137.
Hunk #3 succeeded at 156.
done
===>   Applying OpenBSD patch patch-man_Makefile
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--
|$OpenBSD: patch-man_Makefile,v 1.1 2021/09/14 12:16:15 sthen Exp $
|
|Index: man/Makefile
|--- man/Makefile.orig
|+++ man/Makefile
--
Patching file man/Makefile using Plan A...
Hunk #1 succeeded at 25.
done
===>   Applying OpenBSD patch patch-system_mk
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--
|$OpenBSD: patch-system_mk,v 1.12 2014/03/23 18:17:35 kili Exp $
|--- system.mk.orig Fri Jan  9 23:20:27 2009
|+++ system.mk  Tue Mar 18 11:41:59 2014
--
Patching file system.mk using Plan A...
Hunk #1 succeeded at 8.
Hunk #2 succeeded at 16.
Hunk #3 succeeded at 54.
Hunk #4 succeeded at 64.
Hunk #5 succeeded at 82.
Hunk #6 succeeded at 115.
Hunk #7 succeeded at 131.
Hunk #8 succeeded at 153.
done
===>   Applying OpenBSD patch patch-utils_ion-statusd_statusd_mail_lua
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--
|$OpenBSD: patch-utils_ion-statusd_statusd_mail_lua,v 1.1 2013/04/29 12:40:49 
gonzalo Exp $
|--- utils/ion-statusd/statusd_mail.lua.origMon Apr  8 11:59:40 2013
|+++ utils/ion-statusd/statusd_mail.lua Mon Apr  8 12:00:34 2013
--
Patching file utils/ion-statusd/statusd_mail.lua using Plan A...
Hunk #1 succeeded at 12.
done
===>  Compiler link: clang -> ccache /usr/bin/clang
===>  Compiler link: clang++ -> ccache /usr/bin/clang++
===>  Compiler link: cc -> ccache /usr/bin/cc
===>  Compiler link: c++ -> ccache /usr/bin/c++
===>  Generating configure for ion-20090110p5
===>  Configuring for ion-20090110p5
/usr/bin/perl 

Re: [remove-or-fix] x11/ion

2021-09-14 Thread Stefan Hagen
Stuart Henderson wrote:
> On 2021/09/14 07:43, Stefan Hagen wrote:
> > X11/ion is currently broken.
> 
> Hmm? It does build, though it does either need a BUILD_DEPENDS on groff
> or your change for mandoc to produce consistent results (i.e. including
> the manpage contents in welcome.txt etc).
> 
> $ ls -l */ion-*
> -rw-r--r--  1 root  wheel  760147 Sep 13 07:16 aarch64/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  846233 Sep 13 07:03 amd64/ion-20090110p4.tgz
> -rw-r--r--  2 root  wheel  756225 Jun  2 08:13 arm/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  760814 Sep  6 07:23 i386/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  869767 Aug 22 07:40 mips64/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  843277 Jun 25 07:55 mips64el/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  849572 Sep  9 09:59 powerpc/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  835022 Sep 10 07:12 powerpc64/ion-20090110p4.tgz
> -rw-r--r--  1 root  wheel  810101 Sep 11 09:30 sparc64/ion-20090110p4.tgz
> 
> What problem are you seeing with the various exports.h files? Is there
> something out of the ordinary with your lua installation?

You're right. Once luajit is installed, it doesn't compile anymore.
If I deinstall luajit, it works again.

With the diff it works in both cases.

If we want to keep it, I'll look closer to find the root cause for the
behavior change.

AFAIR it worked even with luajit a few months ago.

This is the error I'm seeing:
[...]
Scanning mplex.c for exports.
cc -O2 -pipe  -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND -I/usr/local/include 
-I/usr/local/include/lua-5.1   -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND 
-I/usr/local/include -I/usr/local/include/lua-5.1 -I/usr/X11R6/include -I.. 
-I.. -I..   -c sizehint.c -o sizehint.o
cc -O2 -pipe  -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND -I/usr/local/include 
-I/usr/local/include/lua-5.1   -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND 
-I/usr/local/include -I/usr/local/include/lua-5.1 -I/usr/X11R6/include -I.. 
-I.. -I..   -c window.c -o window.o
cc -O2 -pipe  -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND -I/usr/local/include 
-I/usr/local/include/lua-5.1   -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND 
-I/usr/local/include -I/usr/local/include/lua-5.1 -I/usr/X11R6/include -I.. 
-I.. -I..   -c ioncore.c -o ioncore.o
cc -O2 -pipe  -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND -I/usr/local/include 
-I/usr/local/include/lua-5.1   -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND 
-I/usr/local/include -I/usr/local/include/lua-5.1 -I/usr/X11R6/include -I.. 
-I.. -I..   -c xic.c -o xic.o
cc -O2 -pipe  -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND -I/usr/local/include 
-I/usr/local/include/lua-5.1   -DCF_XFREE86_TEXTPROP_BUG_WORKAROUND 
-I/usr/local/include -I/usr/local/include/lua-5.1 -I/usr/X11R6/include -I.. 
-I.. -I..   -c selection.c -o selection.o
ioncore.c:66:10: fatal error: 'exports.h' file not found
#include "exports.h"
 ^~~
1 error generated.
gmake[1]: *** [../build/rules.mk:189: ioncore.o] Error 1
gmake[1]: *** Waiting for unfinished jobs
Scanning bindmaps.c for exports.
Scanning gr.c for exports.
Scanning infowin.c for exports.
Scanning activity.c for exports.
Scanning netwm.c for exports.
Scanning frame-draw.c for exports.
Scanning kbresize.c for exports.
Scanning rectangle.c for exports.
Scanning xwindow.c for exports.
Scanning presize.c for exports.
Scanning extlrx.c for exports.
Scanning pholder.c for exports.
Scanning mplexpholder.c for exports.
Scanning llist.c for exports.
Scanning basicpholder.c for exports.
Scanning sizepolicy.c for exports.
Scanning stacking.c for exports.
Scanning group.c for exports.
Scanning grouppholder.c for exports.
Scanning group-cw.c for exports.
Scanning navi.c for exports.
Scanning group-ws.c for exports.
Scanning float-placement.c for exports.
Scanning framedpholder.c for exports.
Scanning return.c for exports.
Scanning detach.c for exports.
Scanning screen-notify.c for exports.
Scanning ../libmainloop/select.c for exports.
Scanning ../libmainloop/defer.c for exports.
Scanning ../libmainloop/signal.c for exports.
Scanning ../libmainloop/hooks.c for exports.
Scanning ../libmainloop/exec.c for exports.
gmake[1]: Leaving directory 
'/usr/ports/pobj/ion-20090110/ion-3-20090110/ioncore'
gmake: *** [build/rules.mk:242: subdirs] Error 2
*** Error 2 in . (/usr/ports/infrastructure/mk/bsd.port.mk:2956 
'/usr/ports/pobj/ion-20090110/.build_done': @cd /usr/ports/pobj/ion-20090110...)
*** Error 2 in /usr/ports/x11/ion 
(/usr/ports/infrastructure/mk/bsd.port.mk:2602 'all': @lock=ion-20090110p4;  
export _LOCKS_HELD=" ion-2009...)
ion#


Best regards,
Stefan



Re: [remove-or-fix] x11/ion

2021-09-14 Thread Stuart Henderson
On 2021/09/14 07:43, Stefan Hagen wrote:
> X11/ion is currently broken.

Hmm? It does build, though it does either need a BUILD_DEPENDS on groff
or your change for mandoc to produce consistent results (i.e. including
the manpage contents in welcome.txt etc).

$ ls -l */ion-*
-rw-r--r--  1 root  wheel  760147 Sep 13 07:16 aarch64/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  846233 Sep 13 07:03 amd64/ion-20090110p4.tgz
-rw-r--r--  2 root  wheel  756225 Jun  2 08:13 arm/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  760814 Sep  6 07:23 i386/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  869767 Aug 22 07:40 mips64/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  843277 Jun 25 07:55 mips64el/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  849572 Sep  9 09:59 powerpc/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  835022 Sep 10 07:12 powerpc64/ion-20090110p4.tgz
-rw-r--r--  1 root  wheel  810101 Sep 11 09:30 sparc64/ion-20090110p4.tgz

What problem are you seeing with the various exports.h files? Is there
something out of the ordinary with your lua installation?



[remove-or-fix] x11/ion

2021-09-13 Thread Stefan Hagen
Hi,

X11/ion is currently broken. With x11/notion, we have the successor of 
ion imported. Ion itself is unmaintained upstream since 2009.

Can we remove it?

If someone wants to keep it, the below diff makes it build again.
(compiles, but untested)

It was also missing nroff.

Best regards,
Stefan

Index: x11/ion/Makefile
===
RCS file: /cvs/ports/x11/ion/Makefile,v
retrieving revision 1.67
diff -u -p -u -p -r1.67 Makefile
--- x11/ion/Makefile12 Jul 2019 20:51:14 -  1.67
+++ x11/ion/Makefile14 Sep 2021 05:39:09 -
@@ -6,7 +6,7 @@ V=  20090110
 DISTNAME=  ion-3-${V}
 PKGNAME=   ion-${V}
 CATEGORIES=x11
-REVISION=  4
+REVISION=  5
 FIX_EXTRACT_PERMISSIONS=Yes
 
 HOMEPAGE=  http://modeemi.cs.tut.fi/~tuomov/ion/
@@ -29,5 +29,8 @@ NO_TEST=  Yes
 
 pre-configure:
${SUBST_CMD} ${WRKSRC}/system.mk
+   find ${WRKSRC} -name "*.c" \
+   -exec sed -i 's/#include "exports\.h"//g' "{}" +
+   sed -i 's/NROFF=.*/NROFF=mandoc/g' ${WRKSRC}/man/Makefile
 
 .include