Re: WIP Patch for p5-Text-Autoformat
Upstream replied to me: Just cut the offending pod. I've done the same in the lastest CPAN version (which I just uploaded). Thanks for pointing out the problem. Damian -- A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects. -- Robert Heinlein
Re: WIP Patch for p5-Text-Autoformat
Chris Bennett wrote:
This looks harder than p5-SVG patch
# perldoc Autoformat.pm
Got a 0-length file from Autoformat.pm via Pod::Perldoc::ToMan!?
also checked
/usr/ports/pobj/p5-Text-Autoformat-1.12p0/Text-Autoformat-1.12/blib/lib/Text/Autoformat.pm
and
/usr/ports/pobj/p5-Text-Autoformat-1.12p0/Text-Autoformat-1.12/lib/Text/Autoformat.pm
There is a lengthy POD inside that should be working.
I just tried removing all the perl in the module ( up from __END__)
and perldoc works fine
I have found the problem but I'm not sure what step to take since this
existing error nullifies the perl code contained within.
The pod notations of =begin up to =cut comments out what looks like
debugging code?
If I comment out these two, the problem is solved for perldoc, but I'm
not sure that this code should be re-activated this way.
Perhaps this whole STDERR should be commented out?
I found this:
# blockquote($text,$para, $format, $tlen, \...@hang, \%args);
sub blockquote {
my ($dummy, $para, $format, $tlen, $hang, $args) = @_;
=begin other
print STDERR "[", join("|", $para->{raw} =~
/ \A(\s*)# $1 - leading whitespace (quotation)
(["']|``)# $2 - opening quotemark
(.*)# $3 - quotation
(''|\2)# $4 closing quotemark
\s*?\n# trailing whitespace
(\1[ ]+)# $5 - leading whitespace (attribution)
(--|-)# $6 - attribution introducer
([^\n]*?$)# $7 - attribution line 1
((\5[^\n]*?$)*)# $8 - attributions lines 2-N
\s*\Z
/xsm
), "]\n";
=cut
$para->{text} =~
/ \A(\s*)# $1 - leading whitespace (quotation)
(["']|``)# $2 - opening quotemark
(.*)# $3 - quotation
(''|\2)# $4 closing quotemark
\s*?\n# trailing whitespace
(\1[ ]+)# $5 - leading whitespace (attribution)
(--|-)# $6 - attribution introducer
(.*?$)# $7 - attribution line 1
((\5.*?$)*)# $8 - attributions lines 2-N
\s*\Z
/xsm
or return;
#print "[$1][$2][$3][$4][$5][$6][$7]\n";
my $indent = length $1;
my $text = $2.$3.$4;
my $qindent = length $2;
my $aindent = length $5;
my $attribintro = $6;
my $attrib = $7.$8;
$text =~ s/\n/ /g;
$_[0] .=
form {squeeze=>$args->{squeeze}, trim=>1,
fill => $args->{expfill}
},
$format . q{ }x$indent . q{<}x$tlen,
@$hang, $text,
$format . q{ }x($qindent) . q{[}x($tlen-$qindent),
@$hang, $text,
{squeeze=>0},
$format . q{ } x $aindent . q{>> } . q{[}x($tlen-$aindent-3),
@$hang, $attribintro, $attrib;
return 1;
}
--
A human being should be able to change a diaper, plan an invasion,
butcher a hog, conn a ship, design a building, write a sonnet, balance
accounts, build a wall, set a bone, comfort the dying, take orders,
give orders, cooperate, act alone, solve equations, analyze a new
problem, pitch manure, program a computer, cook a tasty meal, fight
efficiently, die gallantly. Specialization is for insects.
-- Robert Heinlein
Re: WIP Patch for p5-SVG
On 2010/04/03 16:32, Chris Bennett wrote:
>
>
> Stuart Henderson wrote:
> >On 2010/04/03 14:23, Chris Bennett wrote:
> >>This seems to work.
> >>
> >>in patches directory of p5-SVG
> >>
> >># cat patch-lib_SVG_Extension.pm
> >
> >$ cvs add patches/patch-lib_SVG_Extension.pm
> >- you can add and remove files, but not directories, against anoncvs
> >
> >then
> >
> >$ cvs di -uNp
> >
> >..and send it, along with the PKGNAME bump in Makefile, to ${MAINTAINER) :)
> >
> >
> That does not work. There isn't a patches directory yet.
ah, ok. When you wrote 'in patches directory of p5-SVG' I was assuming
that it existed ;)
Re: WIP Patch for p5-SVG
Stuart Henderson wrote:
On 2010/04/03 14:23, Chris Bennett wrote:
This seems to work.
in patches directory of p5-SVG
# cat patch-lib_SVG_Extension.pm
$ cvs add patches/patch-lib_SVG_Extension.pm
- you can add and remove files, but not directories, against anoncvs
then
$ cvs di -uNp
..and send it, along with the PKGNAME bump in Makefile, to ${MAINTAINER) :)
That does not work. There isn't a patches directory yet.
# cvs -d$CVSROOT add patches/patch-lib_SVG_Extension.pm
cvs add: cannot open CVS/Entries for reading: No such file or directory
cvs [add aborted]: no repository
--
A human being should be able to change a diaper, plan an invasion,
butcher a hog, conn a ship, design a building, write a sonnet, balance
accounts, build a wall, set a bone, comfort the dying, take orders,
give orders, cooperate, act alone, solve equations, analyze a new
problem, pitch manure, program a computer, cook a tasty meal, fight
efficiently, die gallantly. Specialization is for insects.
-- Robert Heinlein
Re: WIP Patch for p5-SVG
On 2010/04/03 14:23, Chris Bennett wrote:
> This seems to work.
>
> in patches directory of p5-SVG
>
> # cat patch-lib_SVG_Extension.pm
$ cvs add patches/patch-lib_SVG_Extension.pm
- you can add and remove files, but not directories, against anoncvs
then
$ cvs di -uNp
..and send it, along with the PKGNAME bump in Makefile, to ${MAINTAINER) :)
[UPDATE] net/ejabberd 2.1.3
Running with this for over a week now on i386 with erlang R13B04 posted
here before.
I know it at the very least builds and starts with R12 on amd64.
Previous notes about updating from 2.0.x to 2.1.x:
http://marc.info/?l=openbsd-ports&m=125958365731168&w=2
http://marc.info/?l=openbsd-ports&m=126385653027305&w=2
For release notes and fixed bugs since 2.1.2 see
http://www.process-one.net/en/ejabberd/release_notes/release_note_ejabberd_2.1.3/
https://support.process-one.net/browse/EJAB/fixforversion/10453
--
viq
Index: Makefile
===
RCS file: /cvs/ports/net/ejabberd/Makefile,v
retrieving revision 1.15
diff -u -d -r1.15 Makefile
--- Makefile5 Apr 2009 23:48:55 - 1.15
+++ Makefile3 Apr 2010 20:10:59 -
@@ -2,7 +2,7 @@
COMMENT= jabber server written in Erlang
-V= 2.0.5
+V= 2.1.3
DISTNAME= ejabberd-$V
CATEGORIES=net
Index: distinfo
===
RCS file: /cvs/ports/net/ejabberd/distinfo,v
retrieving revision 1.6
diff -u -d -r1.6 distinfo
--- distinfo5 Apr 2009 23:48:55 - 1.6
+++ distinfo3 Apr 2010 20:10:59 -
@@ -1,5 +1,5 @@
-MD5 (ejabberd-2.0.5.tar.gz) = LYW0fflp2vCnjtOxZWLXMQ==
-RMD160 (ejabberd-2.0.5.tar.gz) = XsAhCzmVAExwgYCmagluC5vl8bo=
-SHA1 (ejabberd-2.0.5.tar.gz) = 5bjEt0L9zEOdqUWPlPUwYEq9/dU=
-SHA256 (ejabberd-2.0.5.tar.gz) = N++Q4q+itzpiC/caCW30jV/ejxzWafrIPYwUOhKVGYw=
-SIZE (ejabberd-2.0.5.tar.gz) = 1796737
+MD5 (ejabberd-2.1.3.tar.gz) = 5ch+2lMSpuilPfD5tIRLaQ==
+RMD160 (ejabberd-2.1.3.tar.gz) = xXZyko4w9x7GZjbdXnatuALxewM=
+SHA1 (ejabberd-2.1.3.tar.gz) = d/3xjAuzslz5BF4Mgzuk2C8YBGk=
+SHA256 (ejabberd-2.1.3.tar.gz) = YCxuTnQsTKuFlOnwSfCNC/ISX6hK9JHgzrbW8vOFTvY=
+SIZE (ejabberd-2.1.3.tar.gz) = 2243874
Index: patches/patch-src_Makefile_in
===
RCS file: patches/patch-src_Makefile_in
diff -N patches/patch-src_Makefile_in
--- /dev/null 1 Jan 1970 00:00:00 -
+++ patches/patch-src_Makefile_in 3 Apr 2010 20:10:59 -
@@ -0,0 +1,22 @@
+No need to install release notes.
+
+$OpenBSD$
+--- src/Makefile.in.orig Fri Mar 12 11:40:41 2010
src/Makefile.inSun Mar 14 02:57:29 2010
+@@ -111,7 +111,7 @@ MSGSDIR = $(PRIVDIR)/msgs
+ SPOOLDIR = $(DESTDIR)@localstatedir@/lib/ejabberd
+
+ # /var/lock/ejabberdctl
+-CTLLOCKDIR = $(DESTDIR)@localstatedir@/lock/ejabberdctl
++CTLLOCKDIR = $(DESTDIR)/tmp
+
+ # /var/lib/ejabberd/.erlang.cookie
+ COOKIEFILE = $(SPOOLDIR)/.erlang.cookie
+@@ -249,7 +249,6 @@ install: all
+ install -d $(DOCDIR)
+ install -m 644 ../doc/guide.html $(DOCDIR)
+ install -m 644 ../doc/*.png $(DOCDIR)
+- install -m 644 ../doc/*.txt $(DOCDIR)
+
+ uninstall: uninstall-binary
+
Index: patches/patch-src_ejabberd_cfg_example
===
RCS file: /cvs/ports/net/ejabberd/patches/patch-src_ejabberd_cfg_example,v
retrieving revision 1.2
diff -u -d -r1.2 patch-src_ejabberd_cfg_example
--- patches/patch-src_ejabberd_cfg_example 6 Mar 2009 09:19:13 -
1.2
+++ patches/patch-src_ejabberd_cfg_example 3 Apr 2010 20:10:59 -
@@ -1,8 +1,8 @@
$OpenBSD: patch-src_ejabberd_cfg_example,v 1.2 2009/03/06 09:19:13 jasper Exp $
src/ejabberd.cfg.example.orig Wed Jan 14 10:54:15 2009
-+++ src/ejabberd.cfg.example Thu Feb 26 11:36:23 2009
-@@ -114,7 +114,7 @@
- %% certificate, put the correct path to the
+--- src/ejabberd.cfg.example.orig Fri Mar 12 11:40:41 2010
src/ejabberd.cfg.example Sat Mar 13 12:41:08 2010
+@@ -115,7 +115,7 @@
+ %% certificate, specify the full path to the
%% file and uncomment this line:
%%
- %%{certfile, "/path/to/ssl.pem"}, starttls,
@@ -10,9 +10,9 @@
{access, c2s},
{shaper, c2s_shaper},
-@@ -124,12 +124,12 @@
+@@ -125,12 +125,12 @@
%%
- %% To enable the old SSL connection method in port 5223:
+ %% To enable the old SSL connection method on port 5223:
%%
- %%{5223, ejabberd_c2s, [
- %% {access, c2s},
@@ -29,7 +29,7 @@
{5269, ejabberd_s2s_in, [
{shaper, s2s_shaper},
-@@ -160,18 +160,18 @@
+@@ -172,18 +172,18 @@
%% Allowed values are: true or false.
%% You must specify a certificate file.
%%
Index: patches/patch-src_ejabberdctl_template
===
RCS file: /cvs/ports/net/ejabberd/patches/patch-src_ejabberdctl_template,v
retrieving revision 1.2
diff -u -d -r1.2 patch-src_ejabberdctl_template
--- patches/patch-src_ejabberdctl_template 12 Feb 2009 09:58:11 -
1.2
+++ patches/patch-src_ejabberdctl_template 3 Apr 2010 20:10:59 -
@@ -1,6 +
Re: chromium port update
On Sat, Apr 3, 2010 at 3:10 PM, frantisek holop wrote: > hmm, on Wed, Mar 31, 2010 at 08:07:41PM -0700, Peter Valchev said that >> After a long battle with various issues that cropped up, and thanks to >> huge help from the guy working on the FreeBSD port (linked from my >> page), I have an update to chromium-5.0.539.0 > > thanks for the pacakge/port. > > 2 issues so far: > > 1. massive amount of linux specific debug messages on the terminal > > 2. options -> personal stuff -> import data from another browser > the import dialog window going bonkers, jumping between 2 sizes > forever, impossible to click on anything > > the font rendering is really different from all the others.. > it's way too airy, too much whitespace. Have you guys also seen chrome take forever or just not resolve stuff?
WIP Patch for p5-Text-Autoformat
This looks harder than p5-SVG patch # perldoc Autoformat.pm Got a 0-length file from Autoformat.pm via Pod::Perldoc::ToMan!? also checked /usr/ports/pobj/p5-Text-Autoformat-1.12p0/Text-Autoformat-1.12/blib/lib/Text/Autoformat.pm and /usr/ports/pobj/p5-Text-Autoformat-1.12p0/Text-Autoformat-1.12/lib/Text/Autoformat.pm There is a lengthy POD inside that should be working. I just tried removing all the perl in the module ( up from __END__) and perldoc works fine -- A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects. -- Robert Heinlein
WIP Patch for p5-SVG
This seems to work.
in patches directory of p5-SVG
# cat patch-lib_SVG_Extension.pm
--- lib/SVG/Extension.pm.orig Sat Apr 3 13:41:42 2010
+++ lib/SVG/Extension.pm Sat Apr 3 13:41:55 2010
@@ -23,6 +23,9 @@
%TYPES=map { $_ => 1 } @TYPES;
#-
+=head1 NAME
+
+SVG::Extension
sub new {
return shift->SUPER::new(@_);
It patches the file and produces the needed results, as far as I can tell.
man SVG::Extension
and
perldoc SVG::Extension
work with update
--
A human being should be able to change a diaper, plan an invasion,
butcher a hog, conn a ship, design a building, write a sonnet, balance
accounts, build a wall, set a bone, comfort the dying, take orders,
give orders, cooperate, act alone, solve equations, analyze a new
problem, pitch manure, program a computer, cook a tasty meal, fight
efficiently, die gallantly. Specialization is for insects.
-- Robert Heinlein
net/openvpn_bsdauth, was: mandoc errors in ports
Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200:
> net/openvpn_bsdauth
> openvpn_bsdauth.8:22:16: error: multi-line scope breaks multi-line
> scope of Xo
At least on a short time scale, mandoc(1) will not be able to cope
with block nesting errors like the "Oo Xo Oc Xc" here.
In the case at hand, there is no need for explicit blocks at all,
so i suggest to commit this and also push it upstream:
OK?
diff -Nup /var/empty/patch-openvpn_bsdauth_8 patches/patch-openvpn_bsdauth_8
--- /var/empty/patch-openvpn_bsdauth_8 Thu Jan 1 01:00:00 1970
+++ patches/patch-openvpn_bsdauth_8 Sat Apr 3 21:03:15 2010
@@ -0,0 +1,12 @@
+$OpenBSD$
+--- openvpn_bsdauth.8.orig Sat Apr 3 20:58:49 2010
openvpn_bsdauth.8 Sat Apr 3 21:01:27 2010
+@@ -19,7 +19,7 @@
+ .Nd Authenticate users for OpenVPN
+ .Sh SYNOPSYS
+ .Nm
+-.Oo Xo Ar file Oc Xc
++.Op Ar file
+ .Sh DESCRIPTION
+ .Nm
+ is invoked by OpenVPN to authenticate a user by checking a username and a
Index: Makefile
===
RCS file: /cvs/ports/net/openvpn_bsdauth/Makefile,v
retrieving revision 1.3
diff -u -p -r1.3 Makefile
--- Makefile30 Dec 2009 01:11:05 - 1.3
+++ Makefile3 Apr 2010 19:07:42 -
@@ -3,6 +3,7 @@
COMMENT = BSD Auth helper program for OpenVPN
DISTNAME = openvpn_bsdauth-5
+PKGNAME = ${DISTNAME}p0
CATEGORIES = net
HOMEPAGE = http://www.wormhole.hu/~ice/openvpn_bsdauth/
Re: chromium port update
hmm, on Wed, Mar 31, 2010 at 08:07:41PM -0700, Peter Valchev said that > After a long battle with various issues that cropped up, and thanks to > huge help from the guy working on the FreeBSD port (linked from my > page), I have an update to chromium-5.0.539.0 thanks for the pacakge/port. 2 issues so far: 1. massive amount of linux specific debug messages on the terminal 2. options -> personal stuff -> import data from another browser the import dialog window going bonkers, jumping between 2 sizes forever, impossible to click on anything the font rendering is really different from all the others.. it's way too airy, too much whitespace. -f -- lie: a very poor substitute for the truth
net/bluetooth-tools, was: mandoc errors in ports
> net/bluetooth-tools
> btkey.1:123:2: error: bad child for parent context
This one is rather bad, the nesting of blocks is broken in these manual
pages. I will have a look whether mandoc(1) can be made to cope,
but i recommend pushing the following upstream in any case -
and perhaps commit it for now to prevent the port from getting broken.
OK?
diff -Nup /var/empty/patch-btkey_btkey_1 patches/patch-btkey_btkey_1
--- /var/empty/patch-btkey_btkey_1 Thu Jan 1 01:00:00 1970
+++ patches/patch-btkey_btkey_1 Sat Apr 3 19:25:42 2010
@@ -0,0 +1,13 @@
+$OpenBSD$
+--- btkey/btkey.1.orig Sun Feb 17 16:58:17 2008
btkey/btkey.1 Sat Apr 3 19:23:31 2010
+@@ -120,7 +120,8 @@ Clear key for phone at bt3c0 from file and device
+ .Ex -std
+ .Sh FILES
+ .Bl -tag -compact
+-.Pa /var/db/bthcid.keys
++.It Pa /var/db/bthcid.keys
++Bluetooth Link Key cache.
+ .El
+ .Sh SEE ALSO
+ .Xr btpin 1 ,
diff -Nup /var/empty/patch-btpin_btpin_1 patches/patch-btpin_btpin_1
--- /var/empty/patch-btpin_btpin_1 Thu Jan 1 01:00:00 1970
+++ patches/patch-btpin_btpin_1 Sat Apr 3 19:31:46 2010
@@ -0,0 +1,13 @@
+$OpenBSD$
+--- btpin/btpin.1.orig Sat Apr 3 19:31:35 2010
btpin/btpin.1 Sat Apr 3 19:28:49 2010
+@@ -81,7 +81,8 @@ The default path is
+ .Ex -std
+ .Sh FILES
+ .Bl -tag -compact
+-.Pa /var/run/bthcid
++.It Pa /var/run/bthcid
++The control socket.
+ .El
+ .Sh SEE ALSO
+ .Xr btkey 1 ,
diff -u -p -r1.11 Makefile
--- Makefile22 Nov 2008 14:05:37 - 1.11
+++ Makefile3 Apr 2010 17:35:48 -
@@ -2,6 +2,7 @@
COMMENT= bluetooth network tools
DISTNAME= bluetooth-tools-20081122
+PKGNAME= ${DISTNAME}p0
CATEGORIES=net
MAINTAINER=Uwe Stuehler
Re: mandoc errors in ports
Chris Bennett wrote: Ingo Schwarze wrote: Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200: textproc/sgmlformat instant.1:0:1: error: document has no title/section x11/sclock sclock.1:0:1: error: document has no title/section Are these really man(7) pages without a .TH header macro? Must have a look later... In any case, this issue need not be fatal. When we have no title, why not just use "unknown" and prod on? For mdoc(7), mandoc(1) already does something similar in one place. Kristaps, OK? Since someone is talking about man pages, I have gotten the following for a long time: Rebuilding whatis databases: Couldn't find subject in old manpage /usr/local/man/man3p/SVG::Extension.3p Unknown manpage type /usr/local/man/man3p/SVG::Extension.3p Couldn't find subject in old manpage /usr/local/man/man3p/Text::Autoformat.3p Unknown manpage type /usr/local/man/man3p/Text::Autoformat.3p This was easy to fix. Both lacked a NAME. There is a manpage for SVG::Extension Nothing for Text::Autoformat Should a patch be added to the port of p5-SVG to correct this? -- A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects. -- Robert Heinlein
Re: mandoc errors in ports
Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200:
> misc/omconsole
> omconsole.1:36:5: error: requires line arguments > 0 (has 0)
That's a bad syntax error in the mdoc(7) source code.
Matthieu should really fix it upstream.
Just remove the .Xr having no argument!
Still, mandoc(1) need not die, we can just ignore the empty .Xr macro.
OK?
Remaining:
> net/bluetooth-tools
> btkey.1:123:2: error: bad child for parent context
> net/openvpn_bsdauth
> openvpn_bsdauth.8:22:16: error: multi-line scope breaks multi-line
> scope of Xo
> net/ipcalc
> ipcalc.1:70:2: error: multi-line scope breaks multi-line scope of It
> sysutils/pftop
> pftop.8:146:2: error: multi-line scope breaks multi-line scope of It
> sysutils/login_ldap
> login_ldap.8:283:2: error: multi-line scope breaks multi-line scope of
> It
> sysutils/login_fingerprint
> login_fingerprint.8:72:2: error: multi-line scope breaks multi-line
> scope of It
Index: mdoc_html.c
===
RCS file: /cvs/src/usr.bin/mandoc/mdoc_html.c,v
retrieving revision 1.8
diff -u -p -r1.8 mdoc_html.c
--- mdoc_html.c 2 Mar 2010 00:38:59 - 1.8
+++ mdoc_html.c 3 Apr 2010 16:55:13 -
@@ -753,6 +753,9 @@ mdoc_xr_pre(MDOC_ARGS)
struct htmlpair tag[2];
const struct mdoc_node *nn;
+ if (NULL == n->child)
+ return(0);
+
PAIR_CLASS_INIT(&tag[0], "link-man");
if (h->base_man) {
Index: mdoc_term.c
===
RCS file: /cvs/src/usr.bin/mandoc/mdoc_term.c,v
retrieving revision 1.71
diff -u -p -r1.71 mdoc_term.c
--- mdoc_term.c 26 Mar 2010 01:22:05 - 1.71
+++ mdoc_term.c 3 Apr 2010 16:55:18 -
@@ -1275,7 +1275,10 @@ termp_xr_pre(DECL_ARGS)
{
const struct mdoc_node *nn;
- assert(n->child && MDOC_TEXT == n->child->type);
+ if (NULL == n->child)
+ return(0);
+
+ assert(MDOC_TEXT == n->child->type);
nn = n->child;
term_word(p, nn->string);
Index: mdoc_validate.c
===
RCS file: /cvs/src/usr.bin/mandoc/mdoc_validate.c,v
retrieving revision 1.46
diff -u -p -r1.46 mdoc_validate.c
--- mdoc_validate.c 3 Apr 2010 16:30:42 - 1.46
+++ mdoc_validate.c 3 Apr 2010 16:55:20 -
@@ -129,7 +129,7 @@ static v_post posts_text1[] = { eerr_eq
static v_post posts_vt[] = { post_vt, NULL };
static v_post posts_wline[] = { bwarn_ge1, herr_eq0, NULL };
static v_post posts_wtext[] = { ewarn_ge1, NULL };
-static v_post posts_xr[] = { eerr_ge1, NULL };
+static v_post posts_xr[] = { ewarn_ge1, NULL };
static v_prepres_an[] = { pre_an, NULL };
static v_prepres_bd[] = { pre_display, pre_bd, NULL };
static v_prepres_bl[] = { pre_bl, NULL };
Re: scrotwm patch to change wm.
On Thu, Apr 01, 2010 at 03:31:15AM +0100, Edd Barrett wrote:
> Works well here, but the dmenu does not follow colour scheme.
This diff fixes the colour scheme of dmenu by treating the alt_wm feature as a
custom_spawn, meaning it goes through the $some_color expansions just like
normal dmenu does.
So this adds a config variable 'program[alt_wm_menu]' which defaults to dmenu
with colours matching the current scheme :) I updated the manual to reflect
this.
What do you think?
Tested i386.
opencvs server: Diffing inside .
Index: scrotwm.1
===
RCS file: /scrotwm/scrotwm/scrotwm.1,v
retrieving revision 1.28
diff -N -u -p -u scrotwm.1
--- scrotwm.1 7 Oct 2009 03:19:11 - 1.28
+++ scrotwm.1 3 Apr 2010 16:20:47 -
@@ -79,6 +79,12 @@ Enabling or disabling an option is done by using 1 or
The file supports the following keywords:
.Pp
.Bl -tag -width "title_class_enabledXXX" -offset indent -compact
+.It Cm alt_wms
+A comma separated list of alternative window managers for use with
+exec_alt_wm. The menu program specified by the 'alt_wm_menu' program is
+used to show the specified executables as choices of alternative window
+managers. When one is selected, scrotwm will replace itself with the selected
+executable.
.It Cm color_focus
Border color of the currently focussed window.
.It Cm color_unfocus
@@ -192,6 +198,8 @@ xlock
initscreen.sh
.It Cm menu
dmenu_run \-fn $bar_font \-nb $bar_color \-nf $bar_font_color \-sb $bar_border
\-sf $bar_color
+.It Cm alt_wm_menu
+dmenu \-fn $bar_font \-nb $bar_color \-nf $bar_font_color \-sb $bar_border
\-sf $bar_color
.El
.Pp
Custom programs in the configuration file are specified as follows:
@@ -259,6 +267,8 @@ The default key bindings are described below:
term
.It Cm M-p
menu
+.It Cm M-r
+exec_alt_wm
.It Cm M-S-q
quit
.It Cm M-q
@@ -343,6 +353,8 @@ Menu
(see
.Sx PROGRAMS
above)
+.It Cm exec_alt_wm
+Execute an alternative window manager
.It Cm quit
Quit
.Nm
Index: scrotwm.c
===
RCS file: /scrotwm/scrotwm/scrotwm.c,v
retrieving revision 1.281
diff -N -u -p -u scrotwm.c
--- scrotwm.c 13 Jan 2010 23:22:31 - 1.281
+++ scrotwm.c 3 Apr 2010 16:20:50 -
@@ -179,6 +179,7 @@ int cycle_visible = 0;
intterm_width = 0;
intfont_adjusted = 0;
unsigned int mod_key = MODKEY;
+intret_status = -1; /* store return status of fork/exec */
/* dialog windows */
double dialog_ratio = .6;
@@ -385,6 +386,15 @@ struct quirk {
intquirks_size = 0, quirks_length = 0;
struct quirk *quirks = NULL;
+/* alternative window managers */
+struct alt_wm {
+ SLIST_ENTRY(alt_wm) entries;
+ char*wm;
+};
+SLIST_HEAD(head, alt_wm) alt_wms;
+void exec_alt_wm(struct swm_region *r, char **args);
+void free_alt_wm_list();
+
/* events */
#ifdef SWM_DEBUG
void
@@ -587,7 +597,7 @@ sighdlr(int sig)
switch (sig) {
case SIGCHLD:
- while ((pid = waitpid(WAIT_ANY, NULL, WNOHANG)) != -1) {
+ while ((pid = waitpid(WAIT_ANY, &ret_status, WNOHANG)) != -1) {
DNPRINTF(SWM_D_MISC, "reaping: %d\n", pid);
if (pid <= 0)
break;
@@ -1135,6 +1145,150 @@ restart(struct swm_region *r, union arg *args)
quit(NULL, NULL);
}
+/* execute a new window manager */
+void
+exec_alt_wm(struct swm_region *r, char **args)
+{
+ int fd[2], fd1[2], pipe_written = 0;
+ int pipe_read = 0, found_choice = 0;
+ int max_wm_len = -1, cur_wm_len;
+ int pipe_in_sz = 0, i, pid;
+ char*new_wm = NULL, *buf = NULL, *pipe_in;
+ struct alt_wm *wm_node;
+
+ if (SLIST_EMPTY(&alt_wms)) {
+ fprintf(stderr, "exec_alt_wm: no alt_wms specified\n");
+ return;
+ }
+
+ if ((pipe(fd) == -1) || (pipe(fd1) == -1))
+ err(1, "exec_alt_wm: cannot pipe");
+
+ if (signal(SIGPIPE, SIG_IGN) == SIG_ERR)
+ err(1, "exec_alt_wm: cannot disable SIGPIPE");
+
+ /* work out how many wms and the longest name */
+ SLIST_FOREACH(wm_node, &alt_wms, entries) {
+ cur_wm_len = strlen(wm_node->wm);
+ pipe_in_sz = pipe_in_sz + cur_wm_len + 1; /* +1 \n */
+ if (max_wm_len < cur_wm_len)
+ max_wm_len = cur_wm_len;
+ }
+ pipe_in_sz ++; /* \0 */
+
+ pid = fork();
+ switch (pid) {
+ case -1:
+ err(1, "exec_alt_wm: cannot fork");
+ break;
+ case 0: /* we are the child */
+
Re: update: games/zoom 1.1.4
On Sat, 3 Apr 2010 12:34:30 +0200 Landry Breuil
wrote:
> On Sat, Apr 03, 2010 at 12:12:54AM -0700, J.C. Roberts wrote:
> > This is a simple minor update for games/zoom. It fixes a few bugs
> > in the handling of various z-code versions.
> >
> > tested i386.
>
> segfaults when ran over X forwarding, but previous version segfaults
> too.. i suppose it's expected with all those crusty old software.
>
I haven't been able to recreate the segfault you mentioned.
What kind of X forwarding are you using?
I just tried a simple:
$ ssh -Y j...@10.10.10.48 'zoom ZorkI.z3'
j...@10.10.10.48's password:
And it works fine?
If you do not give zoom an input file, the interpreter within the
program "panics" but the program itself shows its "usage" and properly
exits on error.
As for ancient and crusty, yep. It doesn't get much worse. The games
played through zoom and similar Z-Code interpreters are older than some
of the OpenBSD developers.
A friend asked me a historical question about games, and the answer
seemed to be Zork (1977-1979) or the vax port of it, called Dungeon. I
really don't know enough about Z-Code or its interpreters (e.g. zoom).
> > +PKGNAME= ${DISTNAME}
>
> Unneeded... as stated hundreds of times here :)
>
Crap! I remembered to delete the patch level, but I forgot to delete the
whole line.
jcr
Re: mandoc errors in ports
Ingo Schwarze wrote: Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200: textproc/sgmlformat instant.1:0:1: error: document has no title/section x11/sclock sclock.1:0:1: error: document has no title/section Are these really man(7) pages without a .TH header macro? Must have a look later... In any case, this issue need not be fatal. When we have no title, why not just use "unknown" and prod on? For mdoc(7), mandoc(1) already does something similar in one place. Kristaps, OK? Since someone is talking about man pages, I have gotten the following for a long time: Rebuilding whatis databases: Couldn't find subject in old manpage /usr/local/man/man3p/SVG::Extension.3p Unknown manpage type /usr/local/man/man3p/SVG::Extension.3p Couldn't find subject in old manpage /usr/local/man/man3p/Text::Autoformat.3p Unknown manpage type /usr/local/man/man3p/Text::Autoformat.3p -- A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects. -- Robert Heinlein
Re: mandoc errors in ports
Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200:
> textproc/sgmlformat
> instant.1:0:1: error: document has no title/section
> x11/sclock
> sclock.1:0:1: error: document has no title/section
Are these really man(7) pages without a .TH header macro?
Must have a look later...
In any case, this issue need not be fatal.
When we have no title, why not just use "unknown" and prod on?
For mdoc(7), mandoc(1) already does something similar in one place.
Kristaps, OK?
Remaining:
> misc/omconsole
> omconsole.1:36:5: error: requires line arguments > 0 (has 0)
> net/bluetooth-tools
> btkey.1:123:2: error: bad child for parent context
> net/openvpn_bsdauth
> openvpn_bsdauth.8:22:16: error: multi-line scope breaks multi-line
> scope of Xo
> net/ipcalc
> ipcalc.1:70:2: error: multi-line scope breaks multi-line scope of It
> sysutils/pftop
> pftop.8:146:2: error: multi-line scope breaks multi-line scope of It
> sysutils/login_ldap
> login_ldap.8:283:2: error: multi-line scope breaks multi-line scope of
> It
> sysutils/login_fingerprint
> login_fingerprint.8:72:2: error: multi-line scope breaks multi-line
> scope of It
--- man_validate.c.00 Fri Apr 2 13:37:07 2010
+++ man_validate.c Sat Apr 3 16:52:37 2010
@@ -163,8 +163,11 @@ check_root(CHKARGS)
if (NULL == m->first->child)
return(man_nerr(m, n, WNODATA));
- if (NULL == m->meta.title)
- return(man_nerr(m, n, WNOTITLE));
+ if (NULL == m->meta.title) {
+ if ( ! man_nwarn(m, n, WNOTITLE))
+ return(0);
+ m->meta.title = mandoc_strdup("unknown");
+ }
return(1);
}
Re: mandoc errors in ports
Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200:
> games/sudoku-solver
> sudoku-solver.6:33:2: error: too many list types
> sysutils/tabled
> tabled.8:53:2: error: too many list types
This one looks recoverable, too.
When two list types are specified for the same list,
just use the first type, discard the second one, and warn.
Again, this ought to be reported upstream.
Kristaps, OK?
Remaining:
> textproc/sgmlformat
> instant.1:0:1: error: document has no title/section
> x11/sclock
> sclock.1:0:1: error: document has no title/section
> misc/omconsole
> omconsole.1:36:5: error: requires line arguments > 0 (has 0)
> net/bluetooth-tools
> btkey.1:123:2: error: bad child for parent context
> net/openvpn_bsdauth
> openvpn_bsdauth.8:22:16: error: multi-line scope breaks multi-line
> scope of Xo
> net/ipcalc
> ipcalc.1:70:2: error: multi-line scope breaks multi-line scope of It
> sysutils/pftop
> pftop.8:146:2: error: multi-line scope breaks multi-line scope of It
> sysutils/login_ldap
> login_ldap.8:283:2: error: multi-line scope breaks multi-line scope of
> It
> sysutils/login_fingerprint
> login_fingerprint.8:72:2: error: multi-line scope breaks multi-line
> scope of It
--- libmdoc.h Sat Apr 3 16:11:50 2010
+++ libmdoc.h.02Sat Apr 3 15:06:06 2010
@@ -167,6 +167,7 @@
#defineARGV_ARG(1)
#defineARGV_WORD (2)
void mdoc_argv_free(struct mdoc_arg *);
+void mdoc_argn_free(struct mdoc_arg *, int);
int mdoc_args(struct mdoc *, int,
int *, char *, enum mdoct, char **);
int mdoc_zargs(struct mdoc *, int,
--- mdoc_argv.c Sat Apr 3 16:11:51 2010
+++ mdoc_argv.c.02 Sat Apr 3 16:19:40 2010
@@ -285,7 +285,7 @@
void
mdoc_argv_free(struct mdoc_arg *p)
{
- int i, j;
+ int i;
if (NULL == p)
return;
@@ -297,23 +297,28 @@
}
assert(p->argc);
- /* LINTED */
- for (i = 0; i < (int)p->argc; i++) {
- if (0 == p->argv[i].sz)
- continue;
- if (NULL == p->argv[i].value)
- continue;
+ for (i = p->argc - 1; i >= 0; i--)
+ mdoc_argn_free(p, i);
- /* LINTED */
- for (j = 0; j < (int)p->argv[i].sz; j++)
- if (p->argv[i].value[j])
- free(p->argv[i].value[j]);
+ free(p->argv);
+ free(p);
+}
- free(p->argv[i].value);
+
+void
+mdoc_argn_free(struct mdoc_arg *p, int iarg)
+{
+ struct mdoc_argv *arg = &p->argv[iarg];
+ int j;
+
+ if (arg->sz && arg->value) {
+ for (j = arg->sz - 1; j >= 0; j--)
+ free(arg->value[j]);
+ free(arg->value);
}
- free(p->argv);
- free(p);
+ for (--p->argc; iarg < (int)p->argc; iarg++)
+ p->argv[iarg] = p->argv[iarg+1];
}
--- mdoc_validate.c Sat Apr 3 16:13:21 2010
+++ mdoc_validate.c.02 Sat Apr 3 15:31:47 2010
@@ -618,8 +618,12 @@
case (MDOC_Inset):
/* FALLTHROUGH */
case (MDOC_Column):
- if (type >= 0)
- return(mdoc_nerr(mdoc, n, EMULTILIST));
+ if (type >= 0) {
+ if ( ! mdoc_nwarn(mdoc, n, EMULTILIST))
+ return(0);
+ mdoc_argn_free(n->args, pos);
+ break;
+ }
type = n->args->argv[pos].arg;
break;
case (MDOC_Compact):
Re: mandoc errors in ports
Nicholas Marriott wrote on Sat, Apr 03, 2010 at 08:10:47AM +0100: > mandoc doesn't understand Spanish, it expects NAME not NOMBRE. In the long run, we might or might not consider adding proper support for section names in national languages. For some sections, it makes a difference because some macros behave differently in some sections. For now, just treating section names in national languages as custom sections is probably the way to go (see the patch it sent out). > I'd guess this should probably be a warning not an error? Right, it seems it is possible to continue parsing, and even the header line is rendered just fine.
Re: mandoc errors in ports
Hi naddy,
Christian Weisgerber wrote on Sat, Apr 03, 2010 at 03:11:26AM +0200:
> Here's the list of ports that pull in to format their
> man pages, but mandoc(1) errors out:
>
> benchmarks/smtp-benchmark
> smtpsend.8:38:1: error: blank line disallowed
> devel/ctm
> ctm.1:42:1: error: blank line disallowed
> mail/smtp-vilter
> smtp-vilter.8:186:1: error: blank line disallowed
> mail/movemail
> movemail.1:61:1: error: blank line disallowed
> net/cnupm
> cnupmstat.8:52:1: error: blank line disallowed
> sysutils/openpoppassd
> openpoppassd.8:35:1: error: blank line disallowed
> x11/scrotwm
> scrotwm_es.1:21:2: error: NAME section must be first
These should be converted to warnings, it is still possible to
continue parsing, and the result will probably be more or less
reasonable.
See below for a diff; Kristaps: OK?
All the same, the first block of pages should be reported upstream.
Blank lines are not valid syntax in mdoc(7) except in literal context.
The authors might consider using .Pp instead. That might not be what
they want in some cases, but it will in many.
Yours,
Ingo
Remaining:
> games/sudoku-solver
> sudoku-solver.6:33:2: error: too many list types
> misc/omconsole
> omconsole.1:36:5: error: requires line arguments > 0 (has 0)
> net/openvpn_bsdauth
> openvpn_bsdauth.8:22:16: error: multi-line scope breaks multi-line
> scope of Xo
> net/ipcalc
> ipcalc.1:70:2: error: multi-line scope breaks multi-line scope of It
> net/bluetooth-tools
> btkey.1:123:2: error: bad child for parent context
> sysutils/tabled
> tabled.8:53:2: error: too many list types
> sysutils/pftop
> pftop.8:146:2: error: multi-line scope breaks multi-line scope of It
> sysutils/login_ldap
> login_ldap.8:283:2: error: multi-line scope breaks multi-line scope of
> It
> sysutils/login_fingerprint
> login_fingerprint.8:72:2: error: multi-line scope breaks multi-line
> scope of It
> textproc/sgmlformat
> instant.1:0:1: error: document has no title/section
> x11/sclock
> sclock.1:0:1: error: document has no title/section
Index: mdoc.c
===
RCS file: /cvs/src/usr.bin/mandoc/mdoc.c,v
retrieving revision 1.37
diff -u -p -r1.37 mdoc.c
--- mdoc.c 2 Apr 2010 12:39:47 - 1.37
+++ mdoc.c 3 Apr 2010 12:16:22 -
@@ -596,8 +596,12 @@ parsetext(struct mdoc *m, int line, char
for (i = 0; ' ' == buf[i]; i++)
/* Skip leading whitespace. */ ;
- if ('\0' == buf[i])
- return(mdoc_perr(m, line, 0, ENOBLANK));
+ if ('\0' == buf[i]) {
+ if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
+ return(0);
+ if ( ! mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL))
+ return(0);
+ }
/*
* Break apart a free-form line into tokens. Spaces are
Index: mdoc_validate.c
===
RCS file: /cvs/src/usr.bin/mandoc/mdoc_validate.c,v
retrieving revision 1.44
diff -u -p -r1.44 mdoc_validate.c
--- mdoc_validate.c 2 Apr 2010 12:39:47 - 1.44
+++ mdoc_validate.c 3 Apr 2010 12:16:22 -
@@ -1306,8 +1306,9 @@ post_sh_head(POST_ARGS)
* non-CUSTOM has a conventional order to be followed.
*/
- if (SEC_NAME != sec && SEC_NONE == mdoc->lastnamed)
- return(mdoc_nerr(mdoc, mdoc->last, ESECNAME));
+ if (SEC_NAME != sec && SEC_NONE == mdoc->lastnamed &&
+ ! mdoc_nwarn(mdoc, mdoc->last, ESECNAME))
+ return(0);
if (SEC_CUSTOM == sec)
return(1);
if (sec == mdoc->lastnamed)
Re: graphics/cairo with XCB backend
On Tue Mar 30, 2010 at 10:45:16AM +0200, David Coppa wrote: > > I need it for a port I'm working on: the new stable branch (v3.x) of > the awesome window manager. > So it's just nice to hear that anybody works on awesome WM. -- http://www.sizeofvoid.org - raf...@sizeofvoid.org Jabber: z...@jabber.ccc.de GPG/PGP: 0x8790CC0B
Re: update: games/zoom 1.1.4
On Sat, Apr 03, 2010 at 12:12:54AM -0700, J.C. Roberts wrote:
> This is a simple minor update for games/zoom. It fixes a few bugs in
> the handling of various z-code versions.
>
> tested i386.
segfaults when ran over X forwarding, but previous version segfaults
too.. i suppose it's expected with all those crusty old software.
> +PKGNAME= ${DISTNAME}
Unneeded... as stated hundreds of times here :)
Thanks for the diff.
Landry
update: games/zoom 1.1.4
This is a simple minor update for games/zoom. It fixes a few bugs in
the handling of various z-code versions.
tested i386.
jcr
Index: Makefile
===
RCS file: /cvs/ports/games/zoom/Makefile,v
retrieving revision 1.11
diff -N -u -p Makefile
--- Makefile10 Aug 2009 06:31:43 - 1.11
+++ Makefile3 Apr 2010 02:38:47 -
@@ -2,8 +2,8 @@
COMMENT= Z-code interpreter for X11
-DISTNAME= zoom-1.1.2
-PKGNAME= ${DISTNAME}p0
+DISTNAME= zoom-1.1.4
+PKGNAME= ${DISTNAME}
CATEGORIES=games
HOMEPAGE= http://www.logicalshift.co.uk/unix/zoom/
Index: distinfo
===
RCS file: /cvs/ports/games/zoom/distinfo,v
retrieving revision 1.5
diff -N -u -p distinfo
--- distinfo6 Dec 2007 17:03:56 - 1.5
+++ distinfo3 Apr 2010 02:38:47 -
@@ -1,5 +1,5 @@
-MD5 (zoom-1.1.2.tar.gz) = d5FYhceV44knzVGD5xBljQ==
-RMD160 (zoom-1.1.2.tar.gz) = pSXMbFFxekT/H8sRzbz5Df7ZlxU=
-SHA1 (zoom-1.1.2.tar.gz) = C2R15TZ993pPvV+z3Ad0oz4jLGo=
-SHA256 (zoom-1.1.2.tar.gz) = M3du2XDEReQTfJyA9pbmNKVRbZd/H11a18Wq8svLHfQ=
-SIZE (zoom-1.1.2.tar.gz) = 2152557
+MD5 (zoom-1.1.4.tar.gz) = V/vAalDK9gEwf59gSZBB3A==
+RMD160 (zoom-1.1.4.tar.gz) = IxyrlsL5fYlJtw4nAZ/I1VQCCdw=
+SHA1 (zoom-1.1.4.tar.gz) = CiVot1EAVpS+cIXkCGB63F33UCs=
+SHA256 (zoom-1.1.4.tar.gz) = iGtdsBExTS5LS9Pj69KqOIT04x989o1u6kVAe4KWzGw=
+SIZE (zoom-1.1.4.tar.gz) = 2177409
Index: pkg/PLIST
===
RCS file: /cvs/ports/games/zoom/pkg/PLIST,v
retrieving revision 1.2
diff -N -u -p pkg/PLIST
--- pkg/PLIST 15 Sep 2004 00:46:13 - 1.2
+++ pkg/PLIST 3 Apr 2010 02:38:47 -
@@ -1,5 +1,5 @@
@comment $OpenBSD: PLIST,v 1.2 2004/09/15 00:46:13 espie Exp $
-bin/zoom
+...@bin bin/zoom
share/doc/zoom/
share/doc/zoom/README
share/examples/zoom/
Re: mandoc errors in ports
mandoc doesn't understand Spanish, it expects NAME not NOMBRE. I'd guess this should probably be a warning not an error? On Fri, Apr 02, 2010 at 09:56:44PM -0500, Marco Peereboom wrote: > > x11/scrotwm > > scrotwm_es.1:21:2: error: NAME section must be first > > I don't understand the issue. What does this mean? >
