From peter at javamonkey.com  Wed Mar  3 00:53:41 2004
From: peter at javamonkey.com (Peter Seibel)
Date: Tue, 02 Mar 2004 16:53:41 -0800
Subject: [cl-typesetting-devel] Small patch for hyphenation code
Message-ID: <m38yiikay2.fsf@javamonkey.com>

I noticed some bad hyphenations in my generated PDFs such as
hyphenating "store" as "s-tore". Checking with TeX it agrees with me
that there are no good hyphen breaks in "store". Anyway, I figured out
that the problem is cl-typesetting has no concept of \lefthyphenmin
and \righthyphenmin the way TeX does. Here's a patch that adds at
least an approximation of those concepts, set to the default TeX
values.

-Peter


Index: hyphenation-fp.lisp
===================================================================
--- hyphenation-fp.lisp	(revision 10)
+++ hyphenation-fp.lisp	(working copy)
@@ -39,6 +39,13 @@
                                   :directory '(:relative "hyphen-patterns"))
                    *cl-typesetting-base-directory*))
 
+
+(defvar *left-hyphen-minimum* 2
+  "Minimum number of characters that must precede a hyphen.")
+
+(defvar *right-hyphen-minimum* 3
+  "Minimum number of characters that must follow a hyphen.")
+
 ;; An hyphenation object is able to return the list
 ;; of hyphenation points for any word according to
 ;; the language it has been built for
@@ -65,10 +72,16 @@
 	 (result (hyphen-trie-find-exception word-seq (exception-trie hyphen-trie))))
     (unless result
       (setq result (hyphen-trie-find `(#\. , at word-seq #\.) (pattern-trie hyphen-trie))))
-    (mapcar #'first (remove-if
-	     #'(lambda (x) (or (eq (first x) :end) (eq (first x) 0)))
-	     result))))
+    (format t "result : ~a~%" result)
+    (mapcar #'first (remove-if 
+                     #'(lambda (x)
+                         (let ((idx (first x)))
+                           (or (eq idx :end)
+                               (< idx *left-hyphen-minimum*)
+                               (< (- (length word) idx) *right-hyphen-minimum*))))
+                     result))))
 
+
 ;; Format of the language file:
 ;; the first line has 'pattern'
 ;; one pattern per line

-- 
Peter Seibel                                      peter at javamonkey.com

         Lisp is the red pill. -- John Fraser, comp.lang.lisp



From divanov at aha.ru  Wed Mar  3 06:32:57 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Wed, 3 Mar 2004 09:32:57 +0300
Subject: [cl-typesetting-devel] Multi-page tables and layout
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2>
	<000001c400e4$86f2f770$a44d02c3@digo>
Message-ID: <000201c400e9$7eebf770$a44d02c3@digo>

Sorry about previously posting to the wrong list.

| Hello,
| 
| I am on my way to split a table automatically and want your opinion. My
| approach is to put separate rows into a vbox. Do you think this is the 
| right direction?
| 
| What was the intention of the splitable-p slot in the table-row class?
| 
| The weak point is that we don't have the position of the last item
| stroked into the page at our disposal. Is it possible for
| make-filled-vbox to return a dy-left or dy-allocated value indicating
| how much vertical space was actually used?
| 
| We do need a kind of top-level layout staff akin to DefDoc. Its unit
| conversion module could also be of value.
| --
| Sincerely,
| Dmitri Ivanov
| lisp.ystok.ru





From marc.battyani at fractalconcept.com  Wed Mar  3 23:21:50 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Thu, 4 Mar 2004 00:21:50 +0100
Subject: [cl-typesetting-devel] Small patch for hyphenation code
References: <m38yiikay2.fsf@javamonkey.com>
Message-ID: <034501c40176$500e7530$0a04a8c0@marc2>

 "Peter Seibel" <peter at javamonkey.com> writes:

> I noticed some bad hyphenations in my generated PDFs such as
> hyphenating "store" as "s-tore". Checking with TeX it agrees with me
> that there are no good hyphen breaks in "store". Anyway, I figured out
> that the problem is cl-typesetting has no concept of \lefthyphenmin
> and \righthyphenmin the way TeX does. Here's a patch that adds at
> least an approximation of those concepts, set to the default TeX
> values.

[patch...]

Committed to the repository!
(I just removed the format and moved the (length word) out of the loop)
It should work but I could not find the time to test it...

Marc




From marc.battyani at fractalconcept.com  Wed Mar  3 23:22:26 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Thu, 4 Mar 2004 00:22:26 +0100
Subject: [cl-typesetting-devel] Multi-page tables and layout
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo>
	<000201c400e9$7eebf770$a44d02c3@digo>
Message-ID: <034601c40176$646006c0$0a04a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> Sorry about previously posting to the wrong list.
>
> | Hello,
> |
> | I am on my way to split a table automatically and want your opinion. My
> | approach is to put separate rows into a vbox. Do you think this is the
> | right direction?
> |
> | What was the intention of the splitable-p slot in the table-row class?
> |
> | The weak point is that we don't have the position of the last item
> | stroked into the page at our disposal. Is it possible for
> | make-filled-vbox to return a dy-left or dy-allocated value indicating
> | how much vertical space was actually used?
> |
> | We do need a kind of top-level layout staff akin to DefDoc. Its unit
> | conversion module could also be of value.

Hi Dmitri,

I'm sorry but I can't find the time to reply in details before Friday or
Saturday. (Tomorrow I'm going to see customers in NL)

Marc




From peter at javamonkey.com  Thu Mar  4 20:22:35 2004
From: peter at javamonkey.com (Peter Seibel)
Date: Thu, 04 Mar 2004 12:22:35 -0800
Subject: [cl-typesetting-devel] Small patch for hyphenation code
In-Reply-To: <034501c40176$500e7530$0a04a8c0@marc2> (Marc Battyani's message
	of "Thu, 4 Mar 2004 00:21:50 +0100")
References: <m38yiikay2.fsf@javamonkey.com>
	<034501c40176$500e7530$0a04a8c0@marc2>
Message-ID: <m3ptbsgy5w.fsf@javamonkey.com>

"Marc Battyani" <marc.battyani at fractalconcept.com> writes:

>  "Peter Seibel" <peter at javamonkey.com> writes:
>
>> I noticed some bad hyphenations in my generated PDFs such as
>> hyphenating "store" as "s-tore". Checking with TeX it agrees with me
>> that there are no good hyphen breaks in "store". Anyway, I figured out
>> that the problem is cl-typesetting has no concept of \lefthyphenmin
>> and \righthyphenmin the way TeX does. Here's a patch that adds at
>> least an approximation of those concepts, set to the default TeX
>> values.
>
> [patch...]
>
> Committed to the repository!
> (I just removed the format and moved the (length word) out of the loop)
> It should work but I could not find the time to test it...

Whoops. Didn't mean to leave that FORMAT in there obviously. Thanks.

-Peter 

-- 
Peter Seibel                                      peter at javamonkey.com

         Lisp is the red pill. -- John Fraser, comp.lang.lisp



From divanov at aha.ru  Fri Mar  5 16:17:14 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Fri, 5 Mar 2004 19:17:14 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo>
	<000201c400e9$7eebf770$a44d02c3@digo>
	<034601c40176$646006c0$0a04a8c0@marc2>
Message-ID: <001101c402cd$560313d0$2a4302c3@digo>


Hello,

Main features
==============
- Basic top-level stuff
- Multi-page table working only if they are inline
- Multi-page table can have header and footer rows
- Implemented row-span for table cells
- New generic v-split as a replacement of split-lines.

Please test. The resulting pdf file is available in
    lisp.ystok.ru/code/multi-page.pdf

Other suggestions and notes
=========================
1) I suggest renaming the splitable-p slot into "keep" or "keep-together"
which accept values t (meaning keep own lines together) or :next (meaning
keep own lines together as well as with the next box).

2) The package name "typeset" is a bit long, and source looks too wording as
a result. It would be nice having nicknames like TSET, TPS, or even TT or
TY. I vote for TT.

It would be nice having an abbreviation for the paragraph macro - I vote for
universally recognized P.

IMHO put-source-code-string is also too long, let's have simply "put-source"
or the like.

3) AFAIR, code badly needs lots of restructuring and encapsulation. For
example, the pdf:page class should inherit from typeset:page, pdf:document -
from typeset:document, not vice versa.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru
-------------- next part --------------
A non-text attachment was scrubbed...
Name: multi-page.lisp
Type: application/octet-stream
Size: 7808 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/cl-typesetting-devel/attachments/20040305/6661821c/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: di-contrib.zip
Type: application/x-zip-compressed
Size: 9330 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/cl-typesetting-devel/attachments/20040305/6661821c/attachment.bin>

From marc.battyani at fractalconcept.com  Fri Mar  5 23:12:23 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Sat, 6 Mar 2004 00:12:23 +0100
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2>
	<001101c402cd$560313d0$2a4302c3@digo>
Message-ID: <011b01c40307$522928a0$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> Main features
> ==============
> - Basic top-level stuff
> - Multi-page table working only if they are inline
> - Multi-page table can have header and footer rows
> - Implemented row-span for table cells
> - New generic v-split as a replacement of split-lines.
>
> Please test. The resulting pdf file is available in
>     lisp.ystok.ru/code/multi-page.pdf

Very Great!

> Other suggestions and notes
> =========================
> 1) I suggest renaming the splitable-p slot into "keep" or "keep-together"
> which accept values t (meaning keep own lines together) or :next (meaning
> keep own lines together as well as with the next box).

Or maybe a number?

> 2) The package name "typeset" is a bit long, and source looks too wording
as
> a result. It would be nice having nicknames like TSET, TPS, or even TT or
> TY. I vote for TT.

OK, why not. Anybody has another suggestion ?

> It would be nice having an abbreviation for the paragraph macro - I vote
for
> universally recognized P.

OK.

> IMHO put-source-code-string is also too long, let's have simply
"put-source"
> or the like.

Or maybe "verbatim"
What does TeX uses ?

> 3) AFAIR, code badly needs lots of restructuring and encapsulation. For
> example, the pdf:page class should inherit from typeset:page,
pdf:document -
> from typeset:document, not vice versa.

No. cl-pdf can be used without cl-typesetting. At least for now ;-)

I will test and merge this into cl-typesetting this WE.
I any case thanks a lot Dmitri! This is a very great contribution!

All the best,

Marc




From divanov at aha.ru  Sat Mar  6 06:39:38 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Sat, 6 Mar 2004 09:39:38 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2>
	<001101c402cd$560313d0$2a4302c3@digo>
	<011b01c40307$522928a0$0303a8c0@marc2>
Message-ID: <000001c40348$5281bf60$da4d02c3@digo>

Hello Marc,

|> 1) I suggest renaming the splitable-p slot into "keep" or
|> "keep-together" which accept values t (meaning keep own lines
|> together) or :next (meaning keep own lines together as well as with
|> the next box).
|
| Or maybe a number?

Agree.

|> IMHO put-source-code-string is also too long, let's have simply
| "put-source"
|> or the like.
|
| Or maybe "verbatim"
| What does TeX uses ?

"verbatim" is even better. I would also prefer a global switch
*dont-hyphenate* or *hyphenatation-level* if possible.

Small code notes on typo.lisp

;;-- Instead of:
(defmethod (setf font) :around (font (style text-style))
  (when (stringp font)(setf font (pdf:get-font font)))
  (call-next-method font style))
;;-- More natural:
(defmethod (setf font) ((font string) (style text-style))
  (setf (font style) (pdf:get-font font)))

;; About insert-stuff
(defmethod insert-stuff ((obj symbol))
  `(put-string (princ-to-string ,obj)))

It could be of value to make difference between put-string and
put-source-code-string (or verbatim) just inside the insert-stuff. Peter
uses *significant-whitespace* for a similiar purpose:

(defmethod emit-pdf ((thing string))
  (if *significant-whitespace*
    (typeset::put-source-code-string thing)
    (typeset::put-string thing)))

A *significant-whitespace* equivalent can be incorporated at insert-stuff
level like this:

(defmethod insert-stuff ((obj symbol))
  `(funcall (if *significant-whitespace*
                (typeset::put-source-code-string thing)
                (typeset::put-string thing))
              (princ-to-string ,obj)))
Or,

(defmethod insert-stuff ((obj symbol))
  `(funcall *default-stringer* (princ-to-string ,obj)))
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From peter at javamonkey.com  Sat Mar  6 18:21:13 2004
From: peter at javamonkey.com (Peter Seibel)
Date: Sat, 06 Mar 2004 10:21:13 -0800
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables, and
	other contrib
In-Reply-To: <000001c40348$5281bf60$da4d02c3@digo> (Dmitri Ivanov's message
	of "Sat, 6 Mar 2004 09:39:38 +0300")
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net>
	<06b501c3fcab$6f89e0e0$0303a8c0@marc2>
	<16446.25586.351774.46397@mccarthy.genworks.com>
	<06ea01c3fcaf$8a86c3f0$0303a8c0@marc2>
	<000001c400e4$86f2f770$a44d02c3@digo>
	<000201c400e9$7eebf770$a44d02c3@digo>
	<034601c40176$646006c0$0a04a8c0@marc2>
	<001101c402cd$560313d0$2a4302c3@digo>
	<011b01c40307$522928a0$0303a8c0@marc2>
	<000001c40348$5281bf60$da4d02c3@digo>
Message-ID: <m3oer9g7l2.fsf@javamonkey.com>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> Hello Marc,
>
> |> 1) I suggest renaming the splitable-p slot into "keep" or
> |> "keep-together" which accept values t (meaning keep own lines
> |> together) or :next (meaning keep own lines together as well as with
> |> the next box).
> |
> | Or maybe a number?
>
> Agree.
>
> |> IMHO put-source-code-string is also too long, let's have simply
> | "put-source"
> |> or the like.
> |
> | Or maybe "verbatim"
> | What does TeX uses ?
>
> "verbatim" is even better. I would also prefer a global switch
> *dont-hyphenate* or *hyphenatation-level* if possible.

Yes, that would be nice. I've been tempted to hack that in but haven't
gotten around to it. I also would like a way to specify that a
particular string is *not* to be broken across a line even if it
contains whitespace, etc. Maybe that's just "verbatim"; I haven't
played around with what happens if I emit verbatim text in the middle
of a paragraph.

-Peter

-- 
Peter Seibel                                      peter at javamonkey.com

         Lisp is the red pill. -- John Fraser, comp.lang.lisp



From divanov at aha.ru  Sun Mar  7 07:10:14 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Sun, 7 Mar 2004 10:10:14 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2>
	<001101c402cd$560313d0$2a4302c3@digo>
	<011b01c40307$522928a0$0303a8c0@marc2>
Message-ID: <000101c40413$5d0ded20$d54302c3@digo>

Hello Marc,

|> 3) AFAIR, code badly needs lots of restructuring and encapsulation.
|> For example, the pdf:page class should inherit from typeset:page,
| pdf:document -
|> from typeset:document, not vice versa.
|
| No. cl-pdf can be used without cl-typesetting. At least for now ;-)

At least symbols pdf:*document*, pdf:*page* and some others should be
imported into typeset and then reexported.

I assumed that finalize-page can perform custom operations like adding
outline via finalize-fn. Perhaps finalize-page should become a generic and
do this in :before  or :after method for pdf::page argument (as well as
getting content).

BTW, I've run into the questionable method in stroke.lisp:

(defmethod stroke ((style text-style) x y)
  (when (font style)
    (setf *font* (font style)))
  (when (font-size style)
    (setf *font-size* (font-size style)))
  (when (text-x-scale style)
    (setf *text-x-scale* (text-x-scale style)))
  (when (color style)
    (setf *color* (color style))
    (pdf::set-color-fill *color*)))

Though *color* holds a foreground, or stroking, color value, this value is
assigned as a non-stroking PDF color. Must we have the following:

(defmethod stroke ((style text-style) x y)
  (when (font style)
    (setf *font* (font style)))
  (when (font-size style)
    (setf *font-size* (font-size style)))
  (when (text-x-scale style)
    (setf *text-x-scale* (text-x-scale style)))
  (when (color style)
    (setf *color* (color style))
    (pdf::set-color-stroke *color*))
  (when (background-color style)
    (setf *background-color* (background-color style))
    (pdf::set-color-fill *background-color*)))
?
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From marc.battyani at fractalconcept.com  Sun Mar  7 22:52:07 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Sun, 7 Mar 2004 23:52:07 +0100
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2>
	<000101c40413$5d0ded20$d54302c3@digo>
Message-ID: <050a01c40496$d2a746c0$0303a8c0@marc2>


I've merged the code from Dmitri Ivanov into the repository

Added splitable multi-pages tables
Added row-span in tables
Top level document structure
Several improvements

Thanks to Dmitri for all this. :)

Have fun with this and please report the bugs...
(Look at test.lisp for an example of the new functionalities) 

Marc



From marc.battyani at fractalconcept.com  Sun Mar  7 22:52:18 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Sun, 7 Mar 2004 23:52:18 +0100
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2>
	<000101c40413$5d0ded20$d54302c3@digo>
Message-ID: <050b01c40496$d95cce40$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> |> 3) AFAIR, code badly needs lots of restructuring and encapsulation.
> |> For example, the pdf:page class should inherit from typeset:page,
> | pdf:document -
> |> from typeset:document, not vice versa.
> |
> | No. cl-pdf can be used without cl-typesetting. At least for now ;-)
>
> At least symbols pdf:*document*, pdf:*page* and some others should be
> imported into typeset and then reexported.

Yes.

> I assumed that finalize-page can perform custom operations like adding
> outline via finalize-fn. Perhaps finalize-page should become a generic and
> do this in :before  or :after method for pdf::page argument (as well as
> getting content).

I have not looked enough at this code to have an opinion on this yet. :)

> BTW, I've run into the questionable method in stroke.lisp:
>
> (defmethod stroke ((style text-style) x y)
>   (when (font style)
>     (setf *font* (font style)))
>   (when (font-size style)
>     (setf *font-size* (font-size style)))
>   (when (text-x-scale style)
>     (setf *text-x-scale* (text-x-scale style)))
>   (when (color style)
>     (setf *color* (color style))
>     (pdf::set-color-fill *color*)))
>
> Though *color* holds a foreground, or stroking, color value, this value is
> assigned as a non-stroking PDF color. Must we have the following:
>
> (defmethod stroke ((style text-style) x y)
>   (when (font style)
>     (setf *font* (font style)))
>   (when (font-size style)
>     (setf *font-size* (font-size style)))
>   (when (text-x-scale style)
>     (setf *text-x-scale* (text-x-scale style)))
>   (when (color style)
>     (setf *color* (color style))
>     (pdf::set-color-stroke *color*))
>   (when (background-color style)
>     (setf *background-color* (background-color style))
>     (pdf::set-color-fill *background-color*)))

Not for text. When drawing text in pdf, the fill color is used. IIRC the
stroke color would apply when you want to have hollow characters for
instance.

Marc




From divanov at aha.ru  Tue Mar  9 08:20:13 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Tue, 9 Mar 2004 11:20:13 +0300
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000101c40413$5d0ded20$d54302c3@digo>
	<050a01c40496$d2a746c0$0303a8c0@marc2>
Message-ID: <000201c405af$c4418360$445702c3@digo>

Hello Marc,

| I've merged the code from Dmitri Ivanov into the repository
| ...snip...|


Two notes about the code merged.

1) You have inserted code starting with
    (defgeneric v-split (content dx dy &optional v-align)
in both specials.lisp and boxes.lisp.

2) stroke.lisp, (defmethod stroke ((line text-line) x y)
Thanks for  #+lispworks/#+lispworks - featured additions. I found out that
they are not actually needed though don't do any harm. After I have set
    (lw:set-default-character-element-type 'simple-char)
beforehand, your old version works fine. Sorry for bothering.

Two notes about test.lisp
1) To facilitate testing, it would be nice to parameterize banner and
fractal locations as follows:

(defun single-page-example (&optional (file #P"/tmp/ex.pdf")
     (banner #P"/tmp/banner.jpg") (fractal #P"/tmp/fractal.jpg"))

2) multi-page-example seems to artbitrary switch text
color to green for even pages 2 and 4. Not sure if this due to lack of any
initial color setting in the example or some other side effect.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From divanov at aha.ru  Tue Mar  9 09:52:57 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Tue, 9 Mar 2004 12:52:57 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2>
	<000101c40413$5d0ded20$d54302c3@digo>
	<050b01c40496$d95cce40$0303a8c0@marc2>
Message-ID: <000301c405bc$81993af0$445702c3@digo>

Hello Marc,

|> BTW, I've run into the questionable method in stroke.lisp:
|>
|> (defmethod stroke ((style text-style) x y)
|...snip...|

Thanks for the explanation.

BTW, is there any reason for having the two keywords: :centered and :center?
E.g.

(tt:paragraph (:h-align :centered) ...)
(tt:hbox (:align :center) ...)

IMHO if we had them unified to :center, that would give less mess ;-)
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From divanov at aha.ru  Tue Mar  9 10:27:18 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Tue, 9 Mar 2004 13:27:18 +0300
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000101c40413$5d0ded20$d54302c3@digo><050a01c40496$d2a746c0$0303a8c0@marc2>
	<000201c405af$c4418360$445702c3@digo>
Message-ID: <000401c405c1$35259920$445702c3@digo>

| 2) multi-page-example seems to artbitrary switch text
| color to green for even pages 2 and 4. Not sure if this due to lack of
| any initial color setting in the example or some other side effect.

I've discovered that the global value of *color* was retained (0.0 0.6 0.3)
from single-page-example after a non-local exit (aborting from a debugger
prompt). The text-style mechanism seems rather fragile :-)
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From marc.battyani at fractalconcept.com  Wed Mar 10 19:05:27 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 10 Mar 2004 20:05:27 +0100
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	and other contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2>
	<000101c40413$5d0ded20$d54302c3@digo>
	<050b01c40496$d95cce40$0303a8c0@marc2>
	<000301c405bc$81993af0$445702c3@digo>
Message-ID: <0fb701c406d2$a6df7e20$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> BTW, is there any reason for having the two keywords: :centered and
:center?
> E.g.
>
> (tt:paragraph (:h-align :centered) ...)
> (tt:hbox (:align :center) ...)
>
> IMHO if we had them unified to :center, that would give less mess ;-)

Hi Dmitri.

I will put :center everywhere.

Marc





From marc.battyani at fractalconcept.com  Wed Mar 10 19:06:00 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 10 Mar 2004 20:06:00 +0100
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000101c40413$5d0ded20$d54302c3@digo>
	<050a01c40496$d2a746c0$0303a8c0@marc2>
	<000201c405af$c4418360$445702c3@digo>
Message-ID: <0fb801c406d2$ba9089f0$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> Two notes about the code merged.
>
> 1) You have inserted code starting with
>     (defgeneric v-split (content dx dy &optional v-align)
> in both specials.lisp and boxes.lisp.

Corrected. (removed from specials.lisp)

> 2) stroke.lisp, (defmethod stroke ((line text-line) x y)
> Thanks for  #+lispworks/#+lispworks - featured additions. I found out that
> they are not actually needed though don't do any harm. After I have set
>     (lw:set-default-character-element-type 'simple-char)
> beforehand, your old version works fine. Sorry for bothering.

OK, I will put back the initial code.

> Two notes about test.lisp
> 1) To facilitate testing, it would be nice to parameterize banner and
> fractal locations as follows:
>
> (defun single-page-example (&optional (file #P"/tmp/ex.pdf")
>      (banner #P"/tmp/banner.jpg") (fractal #P"/tmp/fractal.jpg"))

Done.

Marc




From marc.battyani at fractalconcept.com  Wed Mar 10 19:06:57 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 10 Mar 2004 20:06:57 +0100
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000101c40413$5d0ded20$d54302c3@digo><050a01c40496$d2a746c0$0303a8c0@marc2><000201c405af$c4418360$445702c3@digo>
	<000401c405c1$35259920$445702c3@digo>
Message-ID: <0fbd01c406d2$dcdb60c0$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:


> | 2) multi-page-example seems to artbitrary switch text
> | color to green for even pages 2 and 4. Not sure if this due to lack of
> | any initial color setting in the example or some other side effect.
>
> I've discovered that the global value of *color* was retained (0.0 0.6
0.3)
> from single-page-example after a non-local exit (aborting from a debugger
> prompt). The text-style mechanism seems rather fragile :-)

Sure it is...
Peter Seibel has found some ones which are now corrected but there are
probably others.
A with-default-style macro which would bind the special variables, could be
used in some places to avoid this.
Or maybe I should just toss away all the *style-special-variables* and just
have a *current-style* holding an object.

Marc




From marc.battyani at fractalconcept.com  Wed Mar 10 19:11:34 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 10 Mar 2004 20:11:34 +0100
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	andother contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo>
	<m3oer9g7l2.fsf@javamonkey.com>
Message-ID: <0fc101c406d3$829363a0$0303a8c0@marc2>

 "Peter Seibel" <peter at javamonkey.com> writes:
> "Dmitri Ivanov" <divanov at aha.ru> writes:
> 
> > "verbatim" is even better. I would also prefer a global switch
> > *dont-hyphenate* or *hyphenatation-level* if possible.
> 
> Yes, that would be nice. I've been tempted to hack that in but haven't
> gotten around to it. I also would like a way to specify that a
> particular string is *not* to be broken across a line even if it
> contains whitespace, etc. Maybe that's just "verbatim"; I haven't
> played around with what happens if I emit verbatim text in the middle
> of a paragraph.

I think it should work.
I renamed #'put-source-code-string to #'verbatim

Marc



From divanov at aha.ru  Thu Mar 11 07:20:13 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Thu, 11 Mar 2004 10:20:13 +0300
Subject: [cl-typesetting-devel] New version
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000101c40413$5d0ded20$d54302c3@digo><050a01c40496$d2a746c0$0303a8c0@marc2><000201c405af$c4418360$445702c3@digo>
	<000401c405c1$35259920$445702c3@digo>
	<0fbd01c406d2$dcdb60c0$0303a8c0@marc2>
Message-ID: <000801c4073e$062b3c00$514d02c3@digo>

Hello Marc,

| A with-default-style macro which would bind the special variables,
| could be used in some places to avoid this.
| Or maybe I should just toss away all the *style-special-variables* and
| just have a *current-style* holding an object.

I think moving to a style object is the right direction. That would be more
modular and less error-prone than tons of special variables.  Furthermore,
that would provide leverage for introducing style taxonomy: paragraph
styles, character styles etc.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From divanov at aha.ru  Thu Mar 11 10:52:24 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Thu, 11 Mar 2004 13:52:24 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	andother contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo>
	<m3oer9g7l2.fsf@javamonkey.com>
	<0fc101c406d3$829363a0$0303a8c0@marc2>
Message-ID: <001201c40759$9d4f4c00$514d02c3@digo>

Hello Marc,

| I think it should work.
| I renamed #'put-source-code-string to #'verbatim

IMHO as matter of convenience, the put-string and verbatim functions should:
- if the argument were null, do nothing;
- if the argument were not a string, convert to string by means
princ-to-string (or some customizable?).

For completeness, NIL should not be output by default, so I suggest:
(defmethod insert-stuff ((obj null))
    (declare (ignore obj)))
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From divanov at aha.ru  Fri Mar 12 05:19:45 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Fri, 12 Mar 2004 08:19:45 +0300
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	andother contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo><m3oer9g7l2.fsf@javamonkey.com><0fc101c406d3$829363a0$0303a8c0@marc2>
	<001201c40759$9d4f4c00$514d02c3@digo>
Message-ID: <000901c407f5$97051090$bd4302c3@digo>

Hello,
 
| IMHO as matter of convenience, the put-string and verbatim functions
| should:
| - if the argument were null, do nothing;
| - if the argument were not a 
| string, convert to string by means princ-to-string (or some 
| customizable?).
| 
| For completeness, NIL should not be output by default, 
| so I suggest:
| (defmethod insert-stuff ((obj null))
|     (declare (ignore obj)))

In addition to
(defmethod insert-stuff ((obj symbol))
  `(put-string ,obj))
or
(defmethod insert-stuff ((obj symbol))
  `(funcall *default-stringer* ,obj))
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru





From marc.battyani at fractalconcept.com  Fri Mar 12 22:08:32 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Fri, 12 Mar 2004 23:08:32 +0100
Subject: [cl-typesetting-devel] Top-level layout, multi-page tables,
	andother contrib
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo>
	<m3oer9g7l2.fsf@javamonkey.com>
	<0fc101c406d3$829363a0$0303a8c0@marc2>
	<001201c40759$9d4f4c00$514d02c3@digo>
Message-ID: <199f01c4087e$8f504340$0303a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

Hi Dmitri,

#|
Sorry for my very late replies. I'm very interested by all this, but I still
have an incredible amount of workload (this week-end, I have 2 PCB to design
and 3 functionalities to add to a web application)]
|#

> IMHO as matter of convenience, the put-string and verbatim functions
should:
> - if the argument were null, do nothing;
> - if the argument were not a string, convert to string by means
> princ-to-string (or some customizable?).

The method to transform a lisp element into some cl-typesetting
representation is #'insert-stuff.
#'put-string is an utility function for #'insert-stuff of a string. But
insert stuff works at macro-expansion time. There is probably a need for a
run time #'typeset-object generic function.

> For completeness, NIL should not be output by default, so I suggest:
> (defmethod insert-stuff ((obj null))
>     (declare (ignore obj)))

This will probably not work because obj will not be nil at compile time so
the following method will be called. And this will output a nil at run time
(defmethod insert-stuff ((obj symbol))
  `(put-string (format nil "~a" ,obj)))

#'typeset-object should be used instead of put-string (with no format).

Marc




From divanov at aha.ru  Sat Mar 13 08:40:05 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Sat, 13 Mar 2004 11:40:05 +0300
Subject: [cl-typesetting-devel] Best way of underlining?
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo>
	<m3oer9g7l2.fsf@javamonkey.com>
	<0fc101c406d3$829363a0$0303a8c0@marc2>
	<001201c40759$9d4f4c00$514d02c3@digo>
	<199f01c4087e$8f504340$0303a8c0@marc2>
Message-ID: <000201c408d6$f17644e0$3f4d02c3@digo>

Hello,

Please advise how to output characters underlined into PDF.

What is the means for the form like follows?

Institution Phone: ___________
Institution Fax: ___________

Sometimes the fields have to be left blank, but sometimes - filled in with
text having the overal underline retained.

Can hrule be used accompanied by with-offset?
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From marc.battyani at fractalconcept.com  Sat Mar 13 13:02:04 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Sat, 13 Mar 2004 14:02:04 +0100
Subject: [cl-typesetting-devel] Best way of underlining?
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo><m3oer9g7l2.fsf@javamonkey.com><0fc101c406d3$829363a0$0303a8c0@marc2><001201c40759$9d4f4c00$514d02c3@digo><199f01c4087e$8f504340$0303a8c0@marc2>
	<000201c408d6$f17644e0$3f4d02c3@digo>
Message-ID: <00fc01c408fb$62898650$0a04a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:

> Please advise how to output characters underlined into PDF.
>
> What is the means for the form like follows?
>
> Institution Phone: ___________
> Institution Fax: ___________
>
> Sometimes the fields have to be left blank, but sometimes - filled in with
> text having the overal underline retained.
>
> Can hrule be used accompanied by with-offset?

No, you should use colored-box which is intended to be inlined.

Marc




From peter at javamonkey.com  Tue Mar 16 20:28:20 2004
From: peter at javamonkey.com (Peter Seibel)
Date: Tue, 16 Mar 2004 12:28:20 -0800
Subject: [cl-typesetting-devel] Portability oops
Message-ID: <m3r7vs8rkr.fsf@javamonkey.com>

Looks like the new top-level stuff has a dependence on some impl
dependent foo. Here's a patch to get rid of it (taking a reasonable
guess at what SYS::REMOVE-PROPERTIES does):

-Peter


Index: top-level.lisp
===================================================================
--- top-level.lisp	(revision 15)
+++ top-level.lisp	(working copy)
@@ -51,6 +51,11 @@
   (room-left :accessor room-left :initarg :room-left :initform 0)
 ))
 
+(defun remove-properties (plist keys)
+  (loop for (key value) on plist by #'cddr
+        unless (member key keys) nconc (list key value)))
+
+
 (defun draw-pages (content &rest args
                    &key (size *default-page-size*)
                         (orientation *default-page-orientation*)
@@ -78,7 +83,7 @@
                                      :footer-bottom footer-bottom
                                      ;; Move room-left into initialize-instance :after?
                                      :room-left (- height top-margin bottom-margin)
-                                     (sys::remove-properties args
+                                     (remove-properties args
                                        '(:size :orientation :bounds
                                          :header-top :footer-bottom :break))))))
       (when (and pdf:*page* (member break '(:before :always)))


-- 
Peter Seibel                                      peter at javamonkey.com

         Lisp is the red pill. -- John Fraser, comp.lang.lisp



From marc.battyani at fractalconcept.com  Tue Mar 16 21:56:37 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Tue, 16 Mar 2004 22:56:37 +0100
Subject: [cl-typesetting-devel] Portability oops
References: <m3r7vs8rkr.fsf@javamonkey.com>
Message-ID: <05fb01c40ba1$8eac6740$0a04a8c0@marc2>

"Peter Seibel" <peter at javamonkey.com> writes:

> Looks like the new top-level stuff has a dependence on some impl
> dependent foo. Here's a patch to get rid of it (taking a reasonable
> guess at what SYS::REMOVE-PROPERTIES does):

[patch...]

Thanks Peter,

I commited the changes to the repository.

(I'm trying to find some time to play with all this by the end of the week.)

Marc





From divanov at aha.ru  Wed Mar 17 09:24:17 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Wed, 17 Mar 2004 12:24:17 +0300
Subject: [cl-typesetting-devel] v-split failure and patch
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo><m3oer9g7l2.fsf@javamonkey.com><0fc101c406d3$829363a0$0303a8c0@marc2>
	<001201c40759$9d4f4c00$514d02c3@digo>
Message-ID: <001001c40c01$c0d5bc70$114302c3@digo>

Hello,

I have bumped into terrible looping bug. I am not sure that the
user-drawn-box was actually the culprit but the error has been detected
where user-drawn-box occurs. Please see the broken-user-box.lisp  attached
and the following trace.

CL-USER 13 > (tt::broken-user-box 34)
Table total height 511.
0 (METHOD TYPESET::STROKE (TYPESET:USER-DRAWN-BOX T T)) > ...
>> TYPESET::BOX : #<TYPESET:USER-DRAWN-BOX 21D74CCC>
>> TYPESET::X   : 50
>> TYPESET::Y   : 35.996597623924345
0 (METHOD TYPESET::STROKE (TYPESET:USER-DRAWN-BOX T T)) < ...
  << VALUE-0 : " B"

Error: Can't make a string of length 4194304.
  1 (abort) Return to level 0.
  2 Return to top loop level 0.


After patching the v-split method, it returns nil as the first value and the
outer draw-pages feeds a new page. The patched return-lines subfunction
simply detects pathological case where the text-lines list is a one-element
list consisting of the spurious text-line. Its boxes slot only includes
either text-style or h-spacing (including explicit or made by
make-hfill-glue) instances.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru
-------------- next part --------------
A non-text attachment was scrubbed...
Name: broken-user-box.lisp
Type: application/octet-stream
Size: 1687 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/cl-typesetting-devel/attachments/20040317/915bdce7/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: v-split-patch.diff
Type: application/octet-stream
Size: 4826 bytes
Desc: not available
URL: <https://mailman.common-lisp.net/pipermail/cl-typesetting-devel/attachments/20040317/915bdce7/attachment-0001.obj>

From divanov at aha.ru  Wed Mar 17 10:10:06 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Wed, 17 Mar 2004 13:10:06 +0300
Subject: [cl-typesetting-devel] Portability oops
References: <m3r7vs8rkr.fsf@javamonkey.com>
Message-ID: <001101c40c27$733caa70$114302c3@digo>

Hello Peter,

| Looks like the new top-level stuff has a dependence on some impl
| dependent foo. Here's a patch to get rid of it (taking a reasonable
| guess at what SYS::REMOVE-PROPERTIES does):
| ...snip...|
|  
| +(defun remove-properties (plist keys)
| +  (loop for (key value) on plist by #'cddr
| +        unless (member key keys) nconc (list key value)))

Sorry, this excerpt is from my lw-compat library:

#+lispworks 
(eval-when (:compile-toplevel :load-toplevel :execute)
    (import 'system::remove-properties))
#-lispworks
(defun remove-properties (plist keys)
  (loop for (key value) on plist by #'cddr
          unless (member key keys :test #'eq) nconc (list key value)))

--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From divanov at aha.ru  Mon Mar 22 11:24:35 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Mon, 22 Mar 2004 14:24:35 +0300
Subject: [cl-typesetting-devel] v-split failure and patch
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo><m3oer9g7l2.fsf@javamonkey.com><0fc101c406d3$829363a0$0303a8c0@marc2><001201c40759$9d4f4c00$514d02c3@digo>
	<001001c40c01$c0d5bc70$114302c3@digo>
	<133d01c40fee$019ebd70$0a04a8c0@marc2>
Message-ID: <002a01c41000$6f408400$7a4302c3@digo>

Hello Marc,

| OK thanks Dmitri. It's committed to the repository. (I just moved a
| badly placed nreverse which was broking the single-page-example.)

Agree that the nreverse is missing as a default cond clause. Sorry.
--
Sincerely,
Dmitri Ivanov
list.ystok.ru




From marc.battyani at fractalconcept.com  Mon Mar 22 09:13:56 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Mon, 22 Mar 2004 10:13:56 +0100
Subject: [cl-typesetting-devel] v-split failure and patch
References: <A583F36C-689D-11D8-9105-000A95CEC334@nittin.net><06b501c3fcab$6f89e0e0$0303a8c0@marc2><16446.25586.351774.46397@mccarthy.genworks.com><06ea01c3fcaf$8a86c3f0$0303a8c0@marc2><000001c400e4$86f2f770$a44d02c3@digo><000201c400e9$7eebf770$a44d02c3@digo><034601c40176$646006c0$0a04a8c0@marc2><001101c402cd$560313d0$2a4302c3@digo><011b01c40307$522928a0$0303a8c0@marc2><000001c40348$5281bf60$da4d02c3@digo><m3oer9g7l2.fsf@javamonkey.com><0fc101c406d3$829363a0$0303a8c0@marc2><001201c40759$9d4f4c00$514d02c3@digo>
	<001001c40c01$c0d5bc70$114302c3@digo>
Message-ID: <133d01c40fee$019ebd70$0a04a8c0@marc2>

"Dmitri Ivanov" <divanov at aha.ru> writes:
> I have bumped into terrible looping bug. I am not sure that the
> user-drawn-box was actually the culprit but the error has been detected
> where user-drawn-box occurs. Please see the broken-user-box.lisp  attached
> and the following trace.
...
 > After patching the v-split method, it returns nil as the first value and
the
> outer draw-pages feeds a new page. The patched return-lines subfunction
> simply detects pathological case where the text-lines list is a
one-element
> list consisting of the spurious text-line. Its boxes slot only includes
> either text-style or h-spacing (including explicit or made by
> make-hfill-glue) instances.

OK thanks Dmitri. It's committed to the repository. (I just moved a badly
placed nreverse which was broking the single-page-example.)

This week I have to write some reports with cl-typesetting so I will have
the opportunity to test the new functionalities :)

Marc




From erik at nittin.net  Wed Mar 24 15:35:16 2004
From: erik at nittin.net (Erik Enge)
Date: Wed, 24 Mar 2004 10:35:16 -0500
Subject: [cl-typesetting-devel] Image in upper-right corner.
Message-ID: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>

Is there any way I can make an image stick in the upper-right corner of 
a
page regardless of how big it is?  Currently, my code looks like this:

  (typeset:cell ()
         (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20))

That cell is in a row and the table has two rows which are half the 
size of
the width of the page.  Sometimes the image is 43x50 pixels and sometime
its 189x26 and yet other times various sizes.  Is there any way I can 
align it
with the upper-right corner regardless of its size?  The various images 
move
around today depending on their sizes.

Thanks,

Erik.




From divanov at aha.ru  Wed Mar 24 15:54:30 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Wed, 24 Mar 2004 18:54:30 +0300
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
Message-ID: <005601c411bb$3bb39d40$845702c3@digo>

Hello Erik,

| Is there any way I can make an image stick in the upper-right corner of
| a page regardless of how big it is?  Currently, my code looks like this:
|
|   (typeset:cell ()
|          (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20))
|
| That cell is in a row and the table has two rows which are half the
| size of
| the width of the page.  Sometimes the image is 43x50 pixels and
| sometime its 189x26 and yet other times various sizes.  Is there any
| way I can align it
| with the upper-right corner regardless of its size?  The various images
| move around today depending on their sizes.

I would recommend tt:draw-pages with a page header containing the image. The
header should be specified as function instead of being static content.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From marc.battyani at fractalconcept.com  Wed Mar 24 16:32:06 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 24 Mar 2004 17:32:06 +0100
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
Message-ID: <060101c411bd$8d21f300$0a04a8c0@marc2>

"Erik Enge" <erik at nittin.net> writes:

> Is there any way I can make an image stick in the upper-right corner of 
> a
> page regardless of how big it is?  Currently, my code looks like this:
> 
>   (typeset:cell ()
>          (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20))
> 
> That cell is in a row and the table has two rows which are half the 
> size of
> the width of the page.  Sometimes the image is 43x50 pixels and sometime
> its 189x26 and yet other times various sizes.  Is there any way I can 
> align it
> with the upper-right corner regardless of its size?  The various images 
> move
> around today depending on their sizes.


Upper-right corner of the page or of the cell?
Have you tried something like this in the cell?

:hfill (image :file fractal-jpg :dx 15 :dy 15 :inline t)
:vfill

Marc




From marc.battyani at fractalconcept.com  Wed Mar 24 16:34:57 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 24 Mar 2004 17:34:57 +0100
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<060101c411bd$8d21f300$0a04a8c0@marc2>
Message-ID: <060f01c411bd$f24ed400$0a04a8c0@marc2>

"Marc Battyani" <marc.battyani at fractalconcept.com> writes:
> "Erik Enge" <erik at nittin.net> writes:
>
> > Is there any way I can make an image stick in the upper-right corner of
> > a
> > page regardless of how big it is?  Currently, my code looks like this:
> >
> >   (typeset:cell ()
> >          (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20))
> >
> > That cell is in a row and the table has two rows which are half the
> > size of
> > the width of the page.  Sometimes the image is 43x50 pixels and sometime
> > its 189x26 and yet other times various sizes.  Is there any way I can
> > align it
> > with the upper-right corner regardless of its size?  The various images
> > move
> > around today depending on their sizes.
>
>
> Upper-right corner of the page or of the cell?
> Have you tried something like this in the cell?
>
> :hfill (image :file fractal-jpg :dx 15 :dy 15 :inline t)
> :vfill

Or wrapping the image in a paragraph with a :align :right style ?

Marc




From erik at nittin.net  Wed Mar 24 17:33:13 2004
From: erik at nittin.net (Erik Enge)
Date: Wed, 24 Mar 2004 12:33:13 -0500
Subject: [cl-typesetting-devel] Image in upper-right corner.
In-Reply-To: <060f01c411bd$f24ed400$0a04a8c0@marc2>
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<060101c411bd$8d21f300$0a04a8c0@marc2>
	<060f01c411bd$f24ed400$0a04a8c0@marc2>
Message-ID: <53F60E34-7DB9-11D8-A1C2-000A95CEC334@nittin.net>

On Mar 24, 2004, at 11:34 AM, Marc Battyani wrote:

>> Upper-right corner of the page or of the cell?

Of the cell, which in this case happens to be the upper-right corner of 
the page too.

>> :hfill (image :file fractal-jpg :dx 15 :dy 15 :inline t)
>> :vfill

I tried

     (typeset:cell ()
       :hfill (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20 
:inline t))

Though that did not help much.  Did I misunderstand?

> Or wrapping the image in a paragraph with a :align :right style ?

Tried that but that did not work, the image stayed in the same spot.

Erik.




From erik at nittin.net  Wed Mar 24 17:33:45 2004
From: erik at nittin.net (Erik Enge)
Date: Wed, 24 Mar 2004 12:33:45 -0500
Subject: [cl-typesetting-devel] Image in upper-right corner.
In-Reply-To: <005601c411bb$3bb39d40$845702c3@digo>
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<005601c411bb$3bb39d40$845702c3@digo>
Message-ID: <66AD0979-7DB9-11D8-A1C2-000A95CEC334@nittin.net>


On Mar 24, 2004, at 10:54 AM, Dmitri Ivanov wrote:

> I would recommend tt:draw-pages with a page header containing the 
> image. The
> header should be specified as function instead of being static content.

Would this header should up on each page?  I believe I only want it on 
the first page.

Erik.




From marc.battyani at fractalconcept.com  Wed Mar 24 17:46:50 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Wed, 24 Mar 2004 18:46:50 +0100
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<060101c411bd$8d21f300$0a04a8c0@marc2>
	<060f01c411bd$f24ed400$0a04a8c0@marc2>
	<53F60E34-7DB9-11D8-A1C2-000A95CEC334@nittin.net>
Message-ID: <06c301c411c7$fdbcf600$0a04a8c0@marc2>

 "Erik Enge" <erik at nittin.net> writes:

> On Mar 24, 2004, at 11:34 AM, Marc Battyani wrote:
> 
> >> Upper-right corner of the page or of the cell?
> 
> Of the cell, which in this case happens to be the upper-right corner of 
> the page too.
> 
> >> :hfill (image :file fractal-jpg :dx 15 :dy 15 :inline t)
> >> :vfill
> 
> I tried
> 
>      (typeset:cell ()
>        :hfill (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20 
> :inline t))
> 
> Though that did not help much.  Did I misunderstand?
> 
> > Or wrapping the image in a paragraph with a :align :right style ?
> 
> Tried that but that did not work, the image stayed in the same spot.

I will have a closer look at this.

Marc



From erik at nittin.net  Wed Mar 24 18:02:29 2004
From: erik at nittin.net (Erik Enge)
Date: Wed, 24 Mar 2004 13:02:29 -0500
Subject: [cl-typesetting-devel] Image in upper-right corner.
In-Reply-To: <06c301c411c7$fdbcf600$0a04a8c0@marc2>
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<060101c411bd$8d21f300$0a04a8c0@marc2>
	<060f01c411bd$f24ed400$0a04a8c0@marc2>
	<53F60E34-7DB9-11D8-A1C2-000A95CEC334@nittin.net>
	<06c301c411c7$fdbcf600$0a04a8c0@marc2>
Message-ID: <6A5189A4-7DBD-11D8-A1C2-000A95CEC334@nittin.net>


On Mar 24, 2004, at 12:46 PM, Marc Battyani wrote:

> I will have a closer look at this.

Thanks.  In the mean time I have decided I will only use same-sized 
images
and I will have them centered (using :h-align :center in 
typeset:paragraph).

Erik.




From divanov at aha.ru  Thu Mar 25 06:33:42 2004
From: divanov at aha.ru (Dmitri Ivanov)
Date: Thu, 25 Mar 2004 09:33:42 +0300
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<005601c411bb$3bb39d40$845702c3@digo>
	<66AD0979-7DB9-11D8-A1C2-000A95CEC334@nittin.net>
Message-ID: <000301c4123f$0c7880c0$fd5702c3@digo>

Hello Erik,

|> I would recommend tt:draw-pages with a page header containing the 
|> image. The
|> header should be specified as function instead of being static
|> content.
| 
| Would this header should up on each page?  I believe I only want it on 
| the first page.

It will. In your case, this is inappropriate. Sorry for misunderstanding.
--
Sincerely,
Dmitri Ivanov
lisp.ystok.ru




From marc.battyani at fractalconcept.com  Fri Mar 26 15:02:59 2004
From: marc.battyani at fractalconcept.com (Marc Battyani)
Date: Fri, 26 Mar 2004 16:02:59 +0100
Subject: [cl-typesetting-devel] Image in upper-right corner.
References: <D9628226-7DA8-11D8-A1C2-000A95CEC334@nittin.net>
	<060101c411bd$8d21f300$0a04a8c0@marc2>
	<060f01c411bd$f24ed400$0a04a8c0@marc2>
	<53F60E34-7DB9-11D8-A1C2-000A95CEC334@nittin.net>
Message-ID: <042001c41343$6e979160$0a04a8c0@marc2>

"Erik Enge" <erik at nittin.net> wrote:

OK i've looked at this:

> I tried
>
>      (typeset:cell ()
>        :hfill (typeset:image :file #p"/somwhere.jpg" :dx 200 :dy 20
> :inline t))
>
> Though that did not help much.  Did I misunderstand?

When you do this, you are left aligned by default so you will get one hfill
on each side of the image so it will end up centered.

> > Or wrapping the image in a paragraph with a :align :right style ?
>
> Tried that but that did not work, the image stayed in the same spot.

Sorry it's not :align but :h-align
I tried this in the single-page-example in test.lisp and for me it works:

(cell (:background-color '(1 1 1))
       (paragraph (:h-align :right)
           (image :file fractal-jpg :dx 15 :dy 15 :inline t)))

Marc