Bug#410011: html2ps: Honour PAPERSIZE environment variable

2007-02-06 Thread Trent Buck
Package: html2ps
Version: 1.0b5-1
Severity: wishlist

AFAICT, html2ps always generates pages of size letter, as this is what
is hard-coded in /etc/html2psrc.

The libpaper-utils package provides a mechanism for setting a
host-wide paper size (/etc/papersize), which can be overridden by
users using the PAPERSIZE environment variable.  The desired paper
size can be queried with `paperconf -n`.

It would be nice if these values were used rather where available, so
that my scripts do not need to add the arguments

-S "@html2ps { paper { type: `paperconf -n` } }"

when invoking html2ps.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]



Bug#410012: html2ps: Unnecessary repetition of author metadata

2007-02-06 Thread Trent Buck
Package: html2ps
Version: 1.0b5-1
Severity: wishlist

Consider the following transcript:

$ cat [123].xhtml
http://www.w3.org/1999/xhtml";>1
http://www.w3.org/1999/xhtml";>2
http://www.w3.org/1999/xhtml";>3
$ html2ps [123].xhtml | ps2pdf - tmp.pdf
$ pdfinfo tmp.pdf | grep ^Author:
Author: Foo + Foo + Bar

html2ps generates author metadata by simply concatenating the author
metadata of each source document (interspersed with " + ").  When
multiple documents are by the same author, this is unnecessarily
repetetive.

Consider the case when a large document is split into 80 separate HTML
documents.  Instead of "Author: Trent Buck", the resulting PDF will
contain "Author: Trent Buck + Trent Buck + ... + Trent Buck".

html2ps should accumulate a list of author metadata values as it reads
in each source document, and only add an element to the list if there
is not already an identical element in the list.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]



Bug#410011: html2ps: Honour PAPERSIZE environment variable

2007-02-07 Thread Trent Buck
On Wed, Feb 07, 2007 at 01:10:32PM +1100, Trent Buck wrote:
> Package: html2ps
> Version: 1.0b5-1
> Severity: wishlist
> 
> AFAICT, html2ps always generates pages of size letter, as this is what
> is hard-coded in /etc/html2psrc.
> 
> The libpaper-utils package provides a mechanism for setting a
> host-wide paper size (/etc/papersize), which can be overridden by
> users using the PAPERSIZE environment variable.  The desired paper
> size can be queried with `paperconf -n`.
> 
> It would be nice if these values were used rather where available, so
> that my scripts do not need to add the arguments
> 
> -S "@html2ps { paper { type: `paperconf -n` } }"
> 
> when invoking html2ps.

Looking at the html2ps source code, line 29 does call paperconf.
Therefore, html2ps already behaves correctly, and the default
/etc/html2psrc is breaking it.

The "paper { type: letter; }" setting should be removed from
/etc/html2psrc.
-- 
Trent Buck, Student Errant


signature.asc
Description: Digital signature


Bug#410011: html2ps: Honour PAPERSIZE environment variable

2007-02-07 Thread Trent Buck
tag 410011 + patch
thank you

>> AFAICT, html2ps always generates pages of size letter, as this is
>> what is hard-coded in /etc/html2psrc.
> 
> Looking at the html2ps source code, line 29 does call paperconf.
> Therefore, html2ps already behaves correctly, and the default
> /etc/html2psrc is breaking it.
> 
> The "paper { type: letter; }" setting should be removed from
> /etc/html2psrc.

That doesn't work either, because of a quoting bug in the html2ps
source.  The attached patch fixes it.
-- 
Trent Buck, Student Errant
--- /tmp/html2ps-1.0b5/html2ps	2007-02-08 13:25:25.0 +1100
+++ /usr/bin/html2ps	2007-02-08 15:36:53.0 +1100
@@ -28,8 +28,8 @@
 $ug='/usr/share/doc/html2ps/html2ps.html';
 $default_paper=`paperconf` || 'a4';
 
-$conf=<<'EOR';
[EMAIL PROTECTED] {
+$conf=<<"EOR";
[EMAIL PROTECTED] {
   package {
 PerlMagick: 0;
 ImageMagick: 0;
@@ -344,7 +344,7 @@
 
 A:link, HR { color: black }
 
[EMAIL PROTECTED] {
[EMAIL PROTECTED] {
   margin-left: 2.5cm;
   margin-right: 2.5cm;
   margin-top: 3cm;


signature.asc
Description: Digital signature


Bug#401855: scheme48: FTBFS on kfreebsd-i386: kfreebsd-i386 not in the architecture list

2007-03-15 Thread Trent Buck
On Wed, Dec 06, 2006 at 01:45:21PM +0100, Petr Salinger wrote:
> The current version fails to build on GNU/kFreeBSD, because
> kfreebsd-i386 is not listed in the Architecture: field.
> 
> Could you please add kfreebsd-i386 to the list of the supported
> architectures?

There are many variations of the original "i386" architecture (and
other archiectures) that should be in the Architecture field, but are
not currently.

scsh has a similar problem (only supports 32-bit architectures); it
uses the `type-handling' package as a workaround.  Perhaps a similar
approach could be adopted here.
-- 
Trent Buck, Student Errant


signature.asc
Description: Digital signature


Bug#410401: scheme48: Please package new upstream version

2007-03-15 Thread Trent Buck
On Sat, Feb 10, 2007 at 01:55:16PM +0100, Lionel Elie Mamane wrote:
> I tried to upgrade the package to version 1.5 (in the same batch as
> addressing bug #401855), but I see that you switched the package to
> quilt, which means I don't know how to do the merge between
> Debian-specific changes and upstream changes.

I have packaged Scheme48 1.5 for Debian, based on the existing 1.3
package.  You can find it here:

http://twb.ath.cx/~twb/debian/pool/scheme48/

It includes experimental support for a 32-bit amd64 package.

It uses CDBS.

It does NOT include the work described in assume-64-bit.diff, because
I lost interest.  The prelude to assume-64-bit.diff describes the
additional work needed to make Scheme48 work on pure 64-bit
architectures.

It does not resolve bug #401855.

> Jorgen, I'll let you the honour of upgrading the package to 1.5.

Jorgen is no longer an active Debian developer.

PS: Please note that SLIME48[1], a slime/swank backend for Scheme48,
only works with Scheme48 1.3, and the author (Taylor Campbell) has no
plans for adding support for Scheme48 1.5.  Therefore, upgrading
Debian's Slime48 package to 1.5 will break SLIME48.  SLIME48 is not
currently part of Debian.

[1] http://mumble.net/~campbell/slime48.html
-- 
Trent Buck, Student Errant


signature.asc
Description: Digital signature


Bug#410401: scheme48: Please package new upstream version

2007-03-15 Thread Trent Buck
On Thu, Mar 15, 2007 at 05:15:01PM +0100, Lionel Elie Mamane wrote:
> > Jorgen is no longer an active Debian developer.
> 
> I know; he resigned _after_ the email you are replying to and possibly
> in reaction to it.

Understood.  My remark was primarily for the record.

> As such, scheme48 is mostly orphaned in Debian. Do you want to adopt
> the Scheme48 package?

I'd be happy to, assuming the demands are as minimal as they appear to
be (one hour per week or less).  The only problem is...

...I'm not a DD, and Jorgen was my sponsor.  So I haven't bothered to
find a new one yet because I haven't needed to (for the packages I
currently maintain).  Presumably I need to find a new sponsor before I
can adopt s48.
-- 
Trent Buck, Student Errant


signature.asc
Description: Digital signature


Bug#421385: $HOME/.html2psrc not read

2007-04-28 Thread Trent Buck
Package: html2ps
Version: 1.0b5-4
Severity: important
Tags: patch

The #410011 fix applied as 1.0b5-4 1) did not fix #410011; and 2)
introduced further errors.  The most important of these is that users'
config files are not read.  The attached patch (in ADDITION to the
1.0b5-4 patch, NOT instead of the 1.0b5-4 patch) fixes both (1) and
(2).
diff -ur html2ps-1.0b5/debian/config/html2psrc 
html2ps-1.0b5/debian/config/html2psrc
--- html2ps-1.0b5/debian/config/html2psrc   2007-04-28 23:54:57.0 
+1000
+++ html2ps-1.0b5/debian/config/html2psrc   2007-04-29 00:19:23.784198087 
+1000
@@ -10,9 +10,6 @@
 libwww-perl: 1;
 path: "/usr/X11R6/bin:/usr/bin";
   }
-  paper {
-type: letter;
-  }
   hyphenation {
 en {
   file: "/usr/share/texmf/tex/generic/hyphen/ushyph1.tex";
diff -ur html2ps-1.0b5/html2ps html2ps-1.0b5/html2ps
--- html2ps-1.0b5/html2ps   2007-04-28 23:54:57.0 +1000
+++ html2ps-1.0b5/html2ps   2007-04-29 00:18:17.017582576 +1000
@@ -88,7 +88,7 @@
 middle: 2cm;
   }
   xref {
-text: "[p $N]";
+text: "[p \$N]";
 passes: 1;
   }
   quote {
@@ -135,8 +135,8 @@
   }
   titlepage {
 content: "
-  $T
-  $[author]";
+  \$T
+  \$[author]";
 margin-top: 4cm;
   }
   font {
@@ -268,7 +268,7 @@
 fuchsia: FF00FF;
 aqua: 00;
   }
-  html2psrc: "$HOME/.html2psrc";
+  html2psrc: "\$HOME/.html2psrc";
   imgalt: "[IMAGE]";
   datefmt: "%e %b %Y  %R";
   locale: "";


Bug#410401: scheme48: Please package new upstream version

2007-06-21 Thread Trent Buck
On Thu, Jun 21, 2007 at 02:37:39PM +0200, Lionel Elie Mamane wrote:
> On Fri, Mar 16, 2007 at 03:43:41AM +1100, Trent Buck wrote:
> > On Thu, Mar 15, 2007 at 05:15:01PM +0100, Lionel Elie Mamane wrote:
> 
> >> As such, scheme48 is mostly orphaned in Debian. Do you want to
> >> adopt the Scheme48 package?
> 
> > I'd be happy to, assuming the demands are as minimal as they appear to
> > be (one hour per week or less).  The only problem is...
> 
> > ...I'm not a DD, and Jorgen was my sponsor.  So I haven't bothered to
> > find a new one yet (...)
> 
> What's the situation? I intended to sponsor your, but as my failure to
> do any step in that direction in several months show, it will not
> happen in a reasonable timeframe. Have you looked for another sponsor?
> 
> (You are welcome to take scsh, too, if you are interested.)

Well, let me summarize the salient data:

- the s48 package is easy to maintain;
- the scsh package is a horrendous spaghetti of unannotated changes to
  the source code; and
- I haven't actually *used* s48 at all lately, I still use scsh a bit.

I'm happy to take over s48 maintenance.  I can probably maintain scsh
as well, but I don't have the time to "clean up" the diff.gz (e.g. to
use quilt).

Features that I have ready for (almost) immediate upload to Debian are:

- scsh 0.6.7
- s48 1.6
- amd64 support (via i386 libs)

One feature that I have started on, but not finished, is support for
pure 64-bit systems.

I'm also reluctant to upload s48 1.6 because it *will* break slime48
in Debian.  On the other hand, slime48 isn't available for Debian, and
I suppose anyone using slime48 can always install the Etch package or
compile s48 1.3 by hand, outside of dpkg.

How about I upload the work I've already done to mentors.debian.net,
but mark it as no-sponsor-needed?  I'll try to do that this week, but
I'll hold off any other action until I hear back from you.
-- 
Trent Buck


signature.asc
Description: Digital signature


Bug#410401: scheme48: Please package new upstream version

2007-06-21 Thread Trent Buck
On Thu, Jun 21, 2007 at 06:25:19PM +0200, Lionel Elie Mamane wrote:
> > - scsh 0.6.7
> 
> That is already in Debian, so I don't see what you mean there. You
> mean an upload just changing Maintainer and/or Uploader lines?

Oh; it must have happened without me noticing.  My main machines are
amd64 at the moment, so it's probably I'd not have seen it.

> > I'm also reluctant to upload s48 1.6 because it *will* break slime48
> > in Debian.  On the other hand, slime48 isn't available for Debian, and
> > I suppose anyone using slime48 can always install the Etch package or
> > compile s48 1.3 by hand, outside of dpkg.
> 
> IMHO, we also cannot stay frozen on s48 1.3 forever because slime48 is
> not following and never will. If you think slime48 is that important,
> then at most do two parallel packages of 1.3 and "latest", but I
> wouldn't bother in your shoes - just latest.

Yes, this is the other side of the fence.  What I'd like to do is find
out who the hell is actually using s48 on Debian, and what proportion
of those users care about slime48.  Any thoughts on how to find those
people?

> I don't think you understood me correctly: I will not be a reliable
> sponsor for the months to come. You need to find another sponsor. So
> tagging them no-sponsor-need doesn't seem a good idea to me.

OK, understood.

Should I file ITA bugs to wnpp for s48 and scsh?
-- 
Trent Buck


signature.asc
Description: Digital signature