pombreda / txt2tags Goto Github PK

Version: txt2tags-2.5
File: usr/share/man/zh/man1/txt2tags.1.gz

man generates the following warning:

usr/share/man/zh/man1/txt2tags.1.gz 3: warning: can't find character with input 
code 144

What steps will reproduce the problem?
The internal processing of a t2t file is the following (see convert method)
the body is split into lines then for each line
1. apply preproc
2. handle t2t tags
3. apply post proc

The problem appears when you have preprocs that introduce '\n'. Then the
lines are really multi-lines and any regex using ^ or $ will not act as
expected. Because one expects ^/$ to be begin/end of line, and strictly
speaking they stand for the beginning/end of the entire string against  
which the regex is applied. See the following example:


%!preproc: "^AAA$" "BBB\nDDD"
%!preproc: "^BBB$" "CCC"

BBB
AAA


What is the expected output? What do you see instead?
The expected output is 

CCC
CCC
DDD

But what you get is

CCC
BBB
DDD

This problem is really annoying when you are relying on internal txt2tags
regex (like the ones processing ``` or """ tags) that you cannot modify...

What version of the product are you using? On what operating system?
2.5 and svn version (Ubuntu Karmic)

Please provide any additional information below.

What steps will reproduce the problem?

1. Create an arbitrary (valid) txt2tags file, say test.t2t.

2. Generate tex source from this file, using the encoding 'utf-8':

> txt2tags --encoding utf-8 -ttex test.t2t

3. Compile the generated tex file using latex or pdflatex:

> pdflatex test.tex


What is the expected output? What do you see instead?

I would have expected to the file to compile without problems.
Instead, the following error is generated:

! LaTeX Error: File `utf-8.def' not found.


What version of the product are you using? On what operating system?

I have encountered this issue with the current stable version and an older
version:
- txt2tags 2.5 on openSUSE 11.1
- txt2tags 2.3 on XUbuntu 8.10

Please provide any additional information below.

The issue can also be triggered by using the %!encoding directive instead
of the command-line option --encoding.

The cause of this issue seems to be that the 'utf-8' encoding is not
translated to its LaTeX equivalent 'utf8', which is supplied as an option
to the 'inputenc' package. So in the case of 'utf-8', txt2tags generates

  \usepackage[utf-8]{inputenc}

while it should generate

  \usepackage[utf8]{inputenc}

What steps will reproduce the problem?
1. use %!include: dir/fileinc.t2t
2. within fileinc.t2t use %!include: otherfile.t2t
3. txt2tags does not find this file if it is in the dir directory

What is the expected output? What do you see instead?
One would expect the inclusion to be relative to the current file

What version of the product are you using? On what operating system?
all versions

Please provide any additional information below.
Here is for review a patch that
- permits include@include and image@include
- add a new macro %%CURRENTFILE

Hi Aurelio and others,

as an enhancement for txt2tags tex output I suppose putting " around image
filenames. This way latex could also recognize filenames with e.g.
underscores and whitespaces.

[image.jpg] should become \includegraphics{"image".jpg}

[sub folder/image.jpg] should become \includegraphics{"sub folder/image".jpg}

[im_age.jpg] should become \includegraphics{"im_age".jpg}

In XHTML, the BLOCKQUOTE tag must have block elements inside. So any paragraph 
must be 
declared inside it:

<blockquote><p>Hello World.</p></blockquote>

Currently txt2tags doesn't add the P tag.

This rule may not be true in HTML, check it.

Jonatan Liljedahl wrote me:

I think you have gotten the IMG align tag the wrong way, here's a quote
from the HTML spec:

13.7.4 Alignment

Attribute definitions

align = bottom|middle|top|left|right
   Deprecated. This attribute specifies the position of an IMG, OBJECT,
or APPLET with respect to its context.

The following values for align concern the object's position with
respect to surrounding text:

   * bottom: means that the bottom of the object should be vertically
aligned with the current baseline. This is the default value.
   * middle: means that the center of the object should be vertically
aligned with the current baseline.
   * top: means that the top of the object should be vertically aligned
with the top of the current text line.

Two other values, left and right, cause the image to float to the
current left or right margin. They are discussed in the section on
floating objects.

But in the code it seems your thought is that it would affect horisontal
alignment. Note that 'middle' refers to vertical alignment, and that
'left' and 'right' is not about alignment but of floating images, (which
is often nice if the image is inside text). But to center an image you
should not use align=middle.

[deleted issue]

Txt2tags should never touch anything inside the ''tagged mark'', even the 
special escapes. Today, 
doFinalEscape() is performed inside tagged texts. Example in tex:�

$ echo -e "\nTest ''pass/pass''" | txt2tags -i- -o- -H -t lout

@LP
Test pass"/"pass


-- thanks Barrie Stott for reporting

Forwarded bug from Debian (still outstanding in 2.5)

From: Luis Mochan <[email protected]>
To: Debian Bug Tracking System <[email protected]>
Subject: txt2tags: bad html links with accented characters
Date: Fri, 01 May 2009 20:29:13 -0500
Package: txt2tags
Version: 2.3-1
Severity: normal
Tags: l10n

I have found documents in spanish that have accented characters in
their addresses, such as 

   http://my.computer/a_ú

or addresses such as 

   http://my.computer/some.page?somid:0,otherid:a

(they might violate the naming convention, but I have found them
frequently enough). Txt2tags assumes that the extraneous characters
(such as ú or :) signal the end of the address and produces incorrect
links such as

    <a href="http://my.computer/a_">http://my.computer/a_</a>ú

%!target and %!includeconf are the only commands that do not support the 
optional target specification, such as:

%!csv(html): foo.csv

Where foo.csv will only be included when the target is HTML.

OK
   %!target: html
   %!includeconf: foo.conf

NOT OK
   %!target(txt): html
   %!includeconf(html): foo.conf

Today, the last two commands will be executed as if the user has not specified 
the target. So it will always be true.

This can confuse the user and bring unexpected problems.

Txt2tags should raise an error when user tries to specify a target in commands 
that do not support it.

I've made a patch for adding the creole target, you'll find it as attachment to 
this "feature request".

I hope it's correct and accurate, so it could enter the svn code.

Keep in mind the creole markup is quite limited compared to txt2tags (no 
strikeout, no underlined text), but some wiki engine are compatible with it so 
it may be a useful inclusion.

for blogging,i hacked txt2tags.py v2.4, so share at first, i will try
to merge into v2.5:
   + for make TOC can click,and make TOC anchor unique:
       - e.g http://blog.zoomquiet.org/pyblosxom/utility/Subversion/
       - in this page include several t2t export xhtml body
       - include several TOC guider list
       - with my hacked,every TOC list can click jump to that chapter
       - and every chapter can click back the right entry top
       hacking: appended one global ver as
       XHTML_TOPTOP = ""
       ...
       # hacked for generate unique top anchor seed
       def doHeader(headers, config):
           #100419:zoomq for toc fixed in multipages
           global XHTML_TOPTOP
           ...
           ret =  string.split(template, '\n')
           #   zoomq:100419 for toc in multi pages
           for i in range(len(ret)):
               if "<h1>" in ret[i]:
                   #print ret[i][4:-5]
                   XHTML_TOPTOP =
base64.b64encode(base64.b32encode(ret[i][4:-6]))[:9]
                   ret[i] = "<a name='toptop%s'
id='toptop%s'></a>%s"%(XHTML_TOPTOP
                       ,XHTML_TOPTOP
                       ,ret[i]
                       )
                   break
           #return string.split(template, '\n')
           return ret

       ...
       class TitleMaster:
           import base64
           ...

           # hacked for generate unique anchor seed for every TOC title
           def _get_tagged_anchor(self):
           ...
               if CONF['toc'] and self.level <= CONF['toc-level']:
               self.anchor_count = self.anchor_count + 1
               # zoomq:100414 fixed link as unique str
               unique =
base64.b64encode(base64.b32encode(self._get_full_title_text()))
               label = label or '%s%s%s'%(
                       self.anchor_prefix
                       , self.anchor_count
                       , unique[:9]
                       )
           ...
           # hacked for plugin jump back the top link usage
XHTML_TOPTOP as unique anchor seed
           def get(self):
               ...
               global XHTML_TOPTOP
               full_title = "<A href='#toptop%s'> %s </A>"
%(XHTML_TOPTOP, self._get_full_title_text())
           ...
           # hacked for generate every TOC title jump into unique anchor
           def dump_marked_toc(self, max_level=99):
               ...
                   for level, count_id, txt, label in self.toc:
                       # zoomq:100414 fixed link as unique str
                       tocStr = "%s %s%s"%(count_id,txt, label)
                       uniqueStr = base64.b64encode(base64.b32encode(tocStr))
                       if level > max_level: continue   # ignore
                       indent = '  '*level
                       id_txt = string.lstrip('%s %s'%(count_id, txt))
                       label = label or self.anchor_prefix+`toc_count`
                       label += uniqueStr[:9]

What steps will reproduce the problem?
1. Create txt2tags document with underscores in title (my_title)
2. Convert it to tex

What is the expected output? 
my\_title

What do you see instead?
my_title

What version of the product are you using? 
trunk

What steps will reproduce the problem?
1. edit your manpage http://txt2tags.sourceforge.net/manpage.t2t
2. add something (at least 2 words) you want to underline __plop plip__
3. convert and see

What is the expected output? What do you see instead?
I expect to see all words and the spaces between them underlined but
nothing is underlined.

What version of the product are you using? On what operating system?
txt2tags-2.5, python-2.6.2, slackware package : http://txt2tags.javila.net/
on  slackware-13.0 and salix OS.

++ fredg

The %!csv command includes a table, but to specify the table header?

Maybe we can use the first line of the CSV as the table header. But this should 
be an option to the user. But how?

This is the current format:

%!csv(html): foo.csv

Where can we include an option for headers? Or maybe always using the first 
line as headers should be the default behavior?

What steps will reproduce the problem?
1. echo '<docbook & test>' | txt2tags -i- -o- -t dbk

What is the expected output? What do you see instead?
Expected output: …<title>&lt;docbook &amp; test&gt;</title>…
Actual output: …<title><docbook & test></title>…


What version of the product are you using? On what operating system?
2.6b (r92)  OSX 10.5.8


Please provide any additional information below.

Also the following change should probably be made to trunk/txt2tags:
@@ 836
-       'TOC'                  : '</index>'                          ,
+       'TOC'                  : '<index/>'                          ,

What steps will reproduce the problem?
1. Enter [""C:\Documents and Settings\image"".png] in a t2t file
2. Export file to latex

What is the expected output? 
\includegraphics{"C:/Documents and Settings/image".png}

What do you see instead?
\includegraphics{"C:$\backslash$Documents and Settings$\backslash$image".png}

What version of the product are you using? On what operating system?
t2t 2.5, Ubuntu Linux Karmic Koala

- when i put in my .t2t file :
[hydrogenAudio foobar2000 Wiki
http://wiki.hydrogenaudio.org/index.php?title=Foobar2000:Foobar2000#Specific_Gui
des]

- I see this as a href link :
http://wiki.hydrogenaudio.org/index.php?title=Foobar2000

- txt2tags 2.5 is used

The "marked.t2t" CSV test fail for the txt target : %%date is not escaped.

It would allow to add user modifications as server-side clones, a feature
of googlecode for mercurial:
http://code.google.com/p/support/issues/detail?id=2946

For example, I have a collection of config.t2t to:
-convert --help output to t2t
-postproc tex target to allow beamer presentations
-add equation support to wiki/tex
and would really like to put them in a central place under version control.

Anyway the conversion is quite simple:
svn checkout http://txt2tags.googlecode.com/svn/trunk/ txt2tags.svn
echo "aureliojargas=Aurélio Marinho Jargas" > authors
hg convert -A authors --datesort txt2tags.svn txt2tags.hg

So give it a try.

From Thomas Hafner:

Unfortunately //foo// does not work inside a hyperlink's description text. I 
circumvented the defect by these postproc definitions:
%!postproc(html): \(/ <I>
%!postproc(html): /\) </I>

Here goes an example how it is used:

[Multiple Dispatch, direkt in (/Common Lisp/) 
http://en.wikipedia.org/wiki/Multiple_dispatch#Common_Lisp]

Don't take me wrong: I don't suggest markup like (/foo/) at all, it's just my 
private workaround. I'd rather like to have //foo// working inside the 
hyperlink, at least for targets where it works. HTML is such a target. HTML 
for the example above looks like this:

<A 
HREF="http://en.wikipedia.org/wiki/Multiple_dispatch#Common_Lisp">Mu
ltiple Dispatch, direkt in <I>Common Lisp</I></A>

I have developed a new experimental feature that allows the outputting of 
multiple [X]HTML files.
The feature effectively allows you to insert page breaks & specify the 
file-name & title of the new page.

The user simply adds lines of the form "%!output: <file-name> <page-title>" to 
their source file.
Whenever txt2tags encounters this it creates a new output file & handles 
writing headers, footers etc.
Subsequent output is directed to the new file.
It also correctly creates any table of contents so links point to the 
appropriate files.

I've also added support for "%!meta: <name> <content>" which adds 'meta' tags 
to the header of the current output file.

Combined these features allow the easy generation of help books for Mac OS X.  
[My primary focus.]

If there's any interest I can supply a patch.  The code is based on svn HEAD.
It also includes a few other small changes including the ability to specify a 
CSS class for each table instance.

txt2tags 2.5 require Python 2.0

I now use in the trunk :
enumerate() from python 2.3 and conditional expressions from python 2.5

Which Python version do we want or need to support ?

 Python | Year
--------+------
 2.0      2000
 2.1      2001
 2.2      2001
 2.3      2003
 2.4      2004
 2.5      2006
 2.6      2008

I was told by Alexandre Felipe Muller de Souza that the Trac wiki syntax is 
compatible with MoinMoin.

We need to confirm if all the marks txt2tags handles are really 100% compatible 
between the two wiki systems.

Confirming that, we can add a new Trac target (-t trac) that will use the 
current MoinMoin settings.

error "formal argument cannot be an instance variable: {|@ext|}"

Tk is old and ugly.
Is there a lot of people using txt2tags with GUI and not in command-line ?

Two little typos related with LaTeX's subparagraph

---------

svn diff

Index: txt2tags
===================================================================
--- txt2tags    (revisión: 86)
+++ txt2tags    (copia de trabajo)
@@ -554,6 +554,7 @@
        blocktitle3Open     blocktitle3Close

        paragraphOpen       paragraphClose
+       subparagraphOpen    subparagraphClose
        blockVerbOpen       blockVerbClose
        blockQuoteOpen      blockQuoteClose blockQuoteLine
        blockCommentOpen    blockCommentClose
@@ -844,7 +845,7 @@
                'title3'               : '~A~\\subsubsection*{\a}',
                # title 4/5: DIRTY: para+BF+\\+\n
                'title4'               : '~A~\\paragraph{}\\textbf{\a}\\\\\n',
-               'title5'               : '~A~\\paragraph{}\\textbf{\a}\\\\\n',
+               'title5'               :
'~A~\\subparagraph{}\\textbf{\a}\\\\\n',
                'numtitle1'            : '\n~A~\section{\a}'      ,
                'numtitle2'            : '~A~\\subsection{\a}'    ,
                'numtitle3'            : '~A~\\subsubsection{\a}' ,

What steps will reproduce the problem?
1.$ cp sample.t2t sample2.t2t
2.$ txt2tags --slides -t art sample.t2t
3.$ txt2tags --slides -t art sample2.t2t
4. sample.art and sample2.art are OK
5.$ txt2tags --slides -t art sample.t2t sample2.t2t
6. sample2.art is NOT OK : "buggy slide" in more between slide header and first 
slide.
7. if you pass more files, they all have the same bug except the first.

XHTML 1.1 became a W3C Recommendation on May 31, 2001.
Txt2tags only support HTML 1.0 Transitional.
Two friends of mine have prefered AsciiDoc instead of txt2tags because of that.

Add the LaTeX-beamer (presentation) target, suggested by Pascal Steichen. He 
have made a patch 
for txt2tags v2.3.2, available at http://pst.libre.lu/toolz/t2t/.

fix txt2tags with follow:

'adoc': """\
= %(HEADER1)s
%(HEADER2)s
%(HEADER3)s
""",

Let the user inform the desired target in any case combination:

-t html
-t HTML
-t Html

%!target: HTML
%!target: Html

The case shouldn't be a problem for the user, as it is today:

$ echo | txt2tags -i- -o- -t HTML
txt2tags: Error: Invalid target 'HTML' (try --help)

What steps will reproduce the problem?
1. If I put some configs in a file and call them by
%!includeconf: myfile, they produced the expected results
2. If I use the command line instead, with
--config-file=myfile some stuff is broken. 

What is the expected output? What do you see instead?
It's seems that target specific rules (like
%!preproc(html): TARGET html
%!preproc(tex): TARGET tex
)
are now not so specific. The last one overrides the previous for any target.


What version of the product are you using? On what operating system?
appears on version 2.5 Ubuntu Karmic and latest svn version

Please provide any additional information below.

Looking inside the script, there is a "suspicious" comment line
2233, in the method get_raw_config:
# Remove the 'target' item of all configs 

So removing this part seems to correct the bug (see patch, note that this
is a simpler and better correction than the one I posted earlier on the
mailing list)

What steps will reproduce the problem?
I put a
%!includeconf(tex): texconfig.t2t
in a t2t document and my texconfig.t2t contains
%!style: aeguill
The aeguill style appears not only in tex document but also in a generated 
xhtml document where it shouldn't:
<link rel="stylesheet" type="text/css" href="aeguill" />

What version of the product are you using? On what operating system?
txt2tags 2.5 (installed through macports)
Debian GNU/Linux 2.6.26-1-686

Florent Gallaire wrote:

"," is the default CSV delimiter (for spreadsheet software and for
python csv library), and the only one supported today, as it's easier
to edit by hand tables with tab delimiter, an option to set delimiter
for each table will be fine too!

What steps will reproduce the problem?
1. create a txt2tags document with titles using = (non numbered titles)
2. use a Table of Content
3. export txt2tags source to tex

What is the expected output? What do you see instead?

The table of content should list the titles header, but it's only working with 
numbered titles 
(+)
It seems because of the limitations of LaTeX, you can't use a TOC with non 
numbered 
titles:
http://web.image.ufl.edu/help/latex/toc.shtml

Maybe it could be possible to find a workaround with txt2tags to allow both 
styles? 

What version of the product are you using? On what operating system?
txt2tags svn 2.6b on linux

Please provide any additional information below.

$ echo | txt2tags -i- -o-
txt2tags: Error: No target specified (try --help)

Maybe trying to convert an old v1.x file?
$

Improve this error message to be more informative and friendly.
The old 1.x series are dead. Try something like:

"You need to inform one target which the -t option or the %!target command.

Example: txt2tags -t html file.t2t"

(Thanks Leonardo Rosa)

What steps will reproduce the problem?
1. use svn

Jonatan Liljedahl wrote me:

• lout can do links:

 "http://snark.ptc.spbu.ru/~uwe/lout/lout.html" @ExternalLink { Lout
Home Page }

• lout can do crosslinks:

 @PageMark foo

or in the section titles:

 @Section @Title { Foo what? }
 @Tag { foo }
 @Begin

and then

 see foo @CrossLink { this page } for more info

• citation marks: "foo" should be converted to ``foo''

• ability to specify custom header, needed to tweak the output. perhaps
when given a %!style, replace the header with:

 @Include { style }
 @Document
 //

• auto-convert images to EPS with imagemagicks 'convert' ?

• lout can do tables

• special chars in titles are not escaped, so a title like 'GNU/Linux'
gives errors.

Can you provide a example latex file which can be converted into MoinMoin
markup

for support SyntaxHighlighter hope::

   self['entryext'] = ext
   self['filebasename'] = file_basename

       - so hacked class BlockMaster can auto generate include right
Syntax support code area!
       hacking:
       BlockMaster->def verb(self):
       ...
           #100420:zoomq   appended for usage SyntaxHighlighter
           if "::brush:" in result[1]:
               #print result[1][2:]
               result[0] = '<PRE class="%s">'%result[1][2:]
               result[1] = ""
               #print result[0]
           #TODO maybe use if not TAGS['blockVerbClose']

that all!
so when delpoy SyntaxHighlighter in site,can base t2t export HTML,
effect syntax color;-)

demo:
http://blog.zoomquiet.org/pyblosxom/techic/PyBlosxom/relivin-pyblosxom-2010-04-1
8-23-23.html#toc6R01YQ0JaV
t2t raw is:
http://blog.zoomquiet.org/pybnodes/techic/PyBlosxom/relivin-pyblosxom-2010-04-18
-23-23.t2t

What steps will reproduce the problem?
1. Create a txt2tags document with:
- list 1
- list 2
 - list 2b
- list 3


+ list 1
+ list 2
 + list 2B
 + list 2C
+ list 3

2. convert it to pmWiki markup.
3.

What is the expected output? What do you see instead?

According to http://www.pmwiki.org/wiki/PmWiki/TextFormattingRules , 
pmWiki should display it:

* list 1
* list 2
** list 2b
* list 3

# list 1
# list 2
## list 2B
## list 2C
# list 3


Instead we got:

* list 1
* list 2
 * list 2b
* list 3

# list 1
# list 2
 # list 2B
 # list 2C
# list 3


What version of the product are you using? On what operating system?
2.6b (current svn version)

Please provide any additional information below.

It seems to be the repeated bullets (like ##) is the preferred version 
(presented by default in http://www.pmwiki.org/wiki/PmWiki/BasicEditing#lists 
), although the version with whitespace is allowed too:
http://www.pmwiki.org/wiki/PmWiki/TextFormattingRules#WhitespaceRules

I can't tell which version is better, but at least the version with repeated 
bullet is more secure for filling some web forms which may remove the trailing 
whitespace.

See attached patch if you wish to correct this.

Jonatan Liljedahl wrote me:

lout already does numbering of sectionsitself, so if txt2tags does it too the 
headings goes like "2.1 2.1 my section".

The tar file contains some garbage files. Not a serious bug, but it's noisy
when packaging :-)

$ tar ztf txt2tags-2.5.tgz |grep "\._"
txt2tags-2.5/._ChangeLog
txt2tags-2.5/doc/._manpage.t2t
txt2tags-2.5/doc/._userguide.pdf
txt2tags-2.5/extras/._css-reference.txt
txt2tags-2.5/extras/._ctags
txt2tags-2.5/extras/._gensite
txt2tags-2.5/extras/._gvim-menu.vim
txt2tags-2.5/extras/._nanorc
txt2tags-2.5/extras/._README
txt2tags-2.5/extras/._txt2tags.php
txt2tags-2.5/extras/._txt2tags.vim
txt2tags-2.5/extras/._txt2tagsrc
txt2tags-2.5/po/._fi.po
txt2tags-2.5/._README
txt2tags-2.5/samples/img/._t2tpowered-black.png
txt2tags-2.5/samples/._sample.t2t
txt2tags-2.5/test/crossing/ok/._table.html
txt2tags-2.5/test/marks/._inline.t2t
txt2tags-2.5/test/marks/._line.t2t
txt2tags-2.5/test/marks/ok/._inline.html
txt2tags-2.5/test/marks/ok/._line.html
txt2tags-2.5/test/marks/ok/._quote.html
txt2tags-2.5/test/marks/ok/._table.html
txt2tags-2.5/test/marks/._quote.t2t
txt2tags-2.5/test/marks/._table.t2t
txt2tags-2.5/test/nesting/._list.t2t
txt2tags-2.5/test/nesting/ok/._list.html
txt2tags-2.5/test/nesting/ok/._quote.html
txt2tags-2.5/test/nesting/ok/._table.html
txt2tags-2.5/test/nesting/._quote.t2t
txt2tags-2.5/test/options/ok/._encoding-4.tex
txt2tags-2.5/test/options/ok/._style-3.tex
txt2tags-2.5/test/options/ok/._style-4.tex
txt2tags-2.5/test/options/ok/._style-5.tex
txt2tags-2.5/._TODO
txt2tags-2.5/._txt2tags

I've made a patch for adding the creole target, you'll find it as attachment to 
this "feature request".

I hope it's correct and accurate, so it could enter the svn code.

Keep in mind the creole markup is quite limited compared to txt2tags (no 
strikeout, no underlined text), but some wiki engine are compatible with it so 
it may be a useful inclusion.

I've made a patch for adding the creole target, you'll find it as attachment to 
this "feature request".

I hope it's correct and accurate, so it could enter the svn code.

Keep in mind the creole markup is quite limited compared to txt2tags (no 
strikeout, no underlined text), but some wiki engine are compatible with it so 
it may be a useful inclusion.

I think it would make a great addition if txt2tags would support sub- and
superscript. It is essential for many users, especially in scientific
environments.

What steps will reproduce the problem?
1. Use -t art with non-ASCII 7 bits chars.
2. Rendering is buggy for boxed and underlined titles.

As the ASCII Art target is in fact a more important feature than I was thinking 
at first, I think this should be fixed (hacked ?) before release.

I just happened to know txt2tags, and like it.

The "Tips and tricks" on your home site is very helpful, so I did a Chinese
translation. You can find the translated page and source here:
http://www.shine-it.net/txt2tags .

However, one thing is driving me crazy. 

If you look at the aforementioned site, you will notice that the examples
in Tip #1 are misaligned when you browse it in Firefox on Windows platform.
Firefox on Linux is working fine, so as to IE7 on Windows.

I guess it could be something relevant with the browser. Any ideas???

BTW. You can put my translation to txt2tags Home site if you like.

Cheers, 

Tony Gu