Shelton/C# should be able to match my HTM_TXT.EXE .

[Jeff_Relf]

Hi Tom, You showed: <<
private const string PHONE_LIST =
"495.1000__424.1111___(206)564-5555_1.800.325.3333";

static void Main( string[] args ) {
foreach (string phoneNumber in Regex.Split (PHONE_LIST, "_+")) {
Console.WriteLine (phoneNumber); } }

Output:
495.1000
424.1111
[phone]
1.800.325.3333 >>

Thanks Tom, that's very interesting,
but not enough to switch me away from LoopTo(),
RegEx simply isn't as flexible.

#define LoopTo( StopCond ) \
while ( Ch && ( Ch = ( uchar ) * ++ P ) \
&& ! ( Ch2 = ( uchar ) P [ 1 ], StopCond ) )

It's a very simple matter to convert HTML to plain text,
following these rules:

These are valid HTML tags: &#x20; <! Comment --> <Alpha> </Alpha>
But, due to the leading space, < Alpha> is not.
Things like &Unknown are sent through untranslated, for obvious reasons.

Pass HTM_TXT.EXE a .HTML file and it spits out a .TXT file.
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.EXE
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.CPP
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.VCPROJ

If RegEx is as powerful as you say,
you should be able to produce something that works at least as well,
and which is just as readable, or more, to me.

[Richard Blewett [DevelopMentor]]

Can you take this OT stuff of the .NET groups please

Regards

Richard Blewett - DevelopMentor
http://www.dotnetconsult.co.uk/weblog
http://www.dotnetconsult.co.uk

[Richard Blewett [DevelopMentor]]

Can you take this OT stuff of the .NET groups please

Regards

Richard Blewett - DevelopMentor
http://www.dotnetconsult.co.uk/weblog
http://www.dotnetconsult.co.uk

[Stefan Simek]

Jeff_Relf wrote:

Hi Stefan_Simek, The title of the post you replied to was: <<
I judge code by what it does. >>

Can you show me links like my AA.HTM, HTM_TXT.EXE and AA.TXT ?

This is the input, View_Source --> File --> Save_Page_As
http://www.Cotse.NET/users/jeffrelf/AA.HTM http://www.triaxis.sk/temp/AA.HTM This is what HTM_TXT.EXE outputs:
http://www.Cotse.NET/users/jeffrelf/AA.TXT http://www.triaxis.sk/temp/AA.TXT
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.EXE
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.CPP
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.VCPROJ http://www.triaxis.sk/temp/HTM_TXT.ZIP

[Jeff_Relf]

Re: http://www.triaxis.sk/temp/AA.TXT
Well, Howdy_Do_Dee, Stefan_Simek... well done !

But you're not preserving whitespace like I do,
you're simply collapsing whitespace,
so it doesn't handle the <pre> tag.

For example, HTM_TXT.EXE handles <pre> by turning this:
http://www.Cotse.NET/users/jeffrelf/index.htm
( View_Source --> File --> Save_Page_As )
into this:
http://www.Cotse.NET/users/jeffrelf/index.TXT

What I do is remove lines that have nothing but whitespace and tags
but leave all lines that had no tags, even if they're just blank lines.
I doubt that your RegEx could do that.

On a much more minor point,
your  » and  © are each using two 7-bit char UTF encoding,
which can be further decoded to single 8-bit chars.

Given that Ch is the name of the first 7-bit char and Ch2 is the second,
UTF decoding goes something like this:
if ( Ch == 194 || Ch == 195 ) && ( Ch2 & 0xC0 ) == 0x80 )
return ( Ch & 3 ) << 6 | Ch2 & 0x3F ;

Re: http://www.triaxis.sk/temp/HTM_TXT.ZIP

That's a lot more files/directories than I showed you,
which, in my opinion, makes your project much less readable/maintainable.

Re: Double_clicking: C:\__\Stefan_Simek\bin\Release\htm_txt.EXE

I got a message saying that I didn't have the .NET's 2.0.5 framework.
I'm running on a Win_XP system I bought about 6 months ago at Office_Depot.
I have MS_Office_XP and Visual_Studio_NET_2003 installed on it,
with whatever the default install did.

Re: C:\__\Stefan_Simek\Program.cs

Although I'm sure it's not true, my copy of Visual_Studio_NET_2003 tells me
there's a bunch of syntax errors in it, including mismatched braces {}.

Re:
Regex.Replace( input
, @"(?'entity'&((\w+)|(#[0-9]+)|(#x[0-9a-fA-F]+));?)|
(?'tag'<[/?]?\w+(([^\][^""])|(\""[^\][^""]*?\""))*?>)
|(?'comment'<!--.*?-->)",
delegate ( Match m ) { // convert entities
if ( m.Groups["entity"].Success ) {
if ( m.Value == "&nbsp;") return " ";
return System.Web.HttpUtility.HtmlDecode( m.Value ); }
if ( m.Groups[ "tag" ].Success && m.Value.ToLower() == "<br>" )
return "\n";
// clear the rest
return ""; }, RegexOptions.Singleline );

Hmm... return System.Web.HttpUtility.HtmlDecode( m.Value ); ?
That's quite bizarre, nothing like C... but it works, I see.

[Stefan Simek]

Jeff_Relf wrote:

Re: http://www.triaxis.sk/temp/AA.TXT
Well, Howdy_Do_Dee, Stefan_Simek... well done !
Thx ;)

But you're not preserving whitespace like I do,
you're simply collapsing whitespace,
so it doesn't handle the <pre> tag.
??

For example, HTM_TXT.EXE handles <pre> by turning this:
http://www.Cotse.NET/users/jeffrelf/index.htm
( View_Source --> File --> Save_Page_As )
into this:
http://www.Cotse.NET/users/jeffrelf/index.TXT

What I do is remove lines that have nothing but whitespace and tags
but leave all lines that had no tags, even if they're just blank lines.
I doubt that your RegEx could do that.
Well, my conversion of your index.htm is byte-to-byte equivalent except
for two empty lines at the end of the document. And I see no special
treatment of the <pre> tag in your code, unless I'm blind...

On a much more minor point,
your  » and  © are each using two 7-bit char UTF encoding,
which can be further decoded to single 8-bit chars.
The output encoding is UTF-8 by default in .NET. You can change it to
anything else by changing the line

using (StreamWriter sw = new StreamWriter(args[1]))

to

using (StreamWriter sw = new StreamWriter(args[1], false,
Encoding.GetEncoding(1250_or_whatever)))

Given that Ch is the name of the first 7-bit char and Ch2 is the second,
UTF decoding goes something like this:
if ( Ch == 194 || Ch == 195 ) && ( Ch2 & 0xC0 ) == 0x80 )
return ( Ch & 3 ) << 6 | Ch2 & 0x3F ;

Re: http://www.triaxis.sk/temp/HTM_TXT.ZIP

That's a lot more files/directories than I showed you,
which, in my opinion, makes your project much less readable/maintainable.
???
I thought that VS6.0 for example generated .dsw, .dsp, .ncb and more
files as well. The *only* file required is the Program.cs. I've provided
a link to the .NET 1.1 source, build command and exe at the end.

Re: Double_clicking: C:\__\Stefan_Simek\bin\Release\htm_txt.EXE

I got a message saying that I didn't have the .NET's 2.0.5 framework.
I'm running on a Win_XP system I bought about 6 months ago at Office_Depot.
I have MS_Office_XP and Visual_Studio_NET_2003 installed on it,
with whatever the default install did.
Well, as the error message says, it requires the .NET framework 2.0.5
(beta2).

Re: C:\__\Stefan_Simek\Program.cs

Although I'm sure it's not true, my copy of Visual_Studio_NET_2003 tells me
there's a bunch of syntax errors in it, including mismatched braces {}.
Because anonymous delegates were not supported back in 1.1.

Re:
Regex.Replace( input
, @"(?'entity'&((\w+)|(#[0-9]+)|(#x[0-9a-fA-F]+));?)|
(?'tag'<[/?]?\w+(([^\][^""])|(\""[^\][^""]*?\""))*?>)
|(?'comment'<!--.*?-->)",
delegate ( Match m ) { // convert entities
if ( m.Groups["entity"].Success ) {
if ( m.Value == "&nbsp;") return " ";
return System.Web.HttpUtility.HtmlDecode( m.Value ); }
if ( m.Groups[ "tag" ].Success && m.Value.ToLower() == "<br>" )
return "\n";
// clear the rest
return ""; }, RegexOptions.Singleline );

Hmm... return System.Web.HttpUtility.HtmlDecode( m.Value ); ?
I see no reason for writing my own entity parser as long as there's one
provided by the framework.
That's quite bizarre, nothing like C... but it works, I see.
Sure it's not like C. It's been a few years since 1978, and the ways of
programing have evolved by now...

See the following for a .NET 1.1 version, whith added command line
checking and exception handling:
http://www.kascomp.sk/tmp/htm_txt.cs
http://www.kascomp.sk/tmp/htm_txt.exe
http://www.kascomp.sk/tmp/build.bat

[Jeff_Relf]

Hi Stefan_Simek,

Your HTM_TXT.EXE preserves whitespace reasonably well,
as your Index.TXT was exactly the same, byte for byte, as mine.
As I tried to say before, my code does not look for the <pre> tag,
it just always preserves whitespace.

Re: using (StreamWriter sw = new StreamWriter(args[1], false
, Encoding.GetEncoding(1250))) sw.Write(output);

Well done !

Re: This .NET 1.1 stuff:
http://www.kascomp.sk/tmp/htm_txt.cs
http://www.kascomp.sk/tmp/htm_txt.exe
http://www.kascomp.sk/tmp/build.bat

Now that's much more like it, well done,
I liked that much more than the .ZIP you showed before.

Re: Your .NET framework 2.0.5 ( beta2 ) code with anonymous delegates,

That was very interesting, but it complicated the installation.

Re: System.Web.HttpUtility.HtmlDecode( m.Value );,

You wrote: << I see no reason for writing my own entity parser
as long as there's one provided by the framework. >>

You have a point there, but I like playing with the lower_level stuff.

You wrote: << Sure it's not like C. It's been a few years since 1978,
and the ways of programing have evolved by now... >>

C# is a mutation, and a rather recent one at that,
it doesn't meet my needs... no #define.

[Exciting Possibility]

Jeff_Relf wrote:

Hi Stefan_Simek,

Your HTM_TXT.EXE preserves whitespace reasonably well,
Am I missing something here?

As a test I saved this page:

http://www.howstuffworks.com/search.php

as a text file.

I compiled htm_txt.cs in mono and ran it against search.php.html, it
produced a text file search.php.txt This output file has only one line
in it -- what appears to be the title.

Both the html source and txt output are attached.


as your Index.TXT was exactly the same, byte for byte, as mine.
As I tried to say before, my code does not look for the <pre> tag,
it just always preserves whitespace.

Re: using (StreamWriter sw = new StreamWriter(args[1], false
, Encoding.GetEncoding(1250))) sw.Write(output);

Well done !

Re: This .NET 1.1 stuff:
http://www.kascomp.sk/tmp/htm_txt.cs
http://www.kascomp.sk/tmp/htm_txt.exe
http://www.kascomp.sk/tmp/build.bat

Now that's much more like it, well done,
I liked that much more than the .ZIP you showed before.

Re: Your .NET framework 2.0.5 ( beta2 ) code with anonymous delegates,

That was very interesting, but it complicated the installation.

Re: System.Web.HttpUtility.HtmlDecode( m.Value );,

You wrote: << I see no reason for writing my own entity parser
as long as there's one provided by the framework. >>

You have a point there, but I like playing with the lower_level stuff.

You wrote: << Sure it's not like C. It's been a few years since 1978,
and the ways of programing have evolved by now... >>

C# is a mutation, and a rather recent one at that,
it doesn't meet my needs... no #define.

HowStuffWorks - Search

[Jeff_Relf]

Hi John, Re: Your attempt to convert
http://www.howstuffworks.com/search.php

You can't just do a Save_Page_As, you Must do a View_Source first.

This is what the .HTM file should look like:
http://www.Cotse.NET/users/jeffrelf/BB.HTM
and my HTM_TXT.EXE translates it to this:
http://www.Cotse.NET/users/jeffrelf/BB.TXT

Notice that my HTM_TXT is faithful to the raw whitespace and the <BR> tag,
leaving/printing as many as was called for,
while Stefan_Simek's HTM_TXT won't allow more than two blank lines in a row.

But I had to radically modify my HTM_TXT to properly handle multilined tags:
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.EXE
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.CPP
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.VCPROJ

And, because HTM_TXT is merely a demo of code from X.CPP
( my custom e-mail client and newsreader ) these files were also affected:
http://www.Cotse.NET/users/jeffrelf/X.EXE
http://www.Cotse.NET/users/jeffrelf/X.CPP
http://www.Cotse.NET/users/jeffrelf/X.VCPROJ

[Jeff_Relf]

By the way John, Re:
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.CPP
http://www.kascomp.sk/tmp/htm_txt.cs

Both HTM_TXT.CPP and htm_txt.cs could be called obfuscated.
In fact, I consider my own code to be more readable,
( perhaps because I wrote it ).

Any code that does useful work is going to take time to understand.

[Journey To The Center of The Earth]

Jeff_Relf wrote:

Any code that does useful work is going to take time to understand.

Very true...which is why you should not dismiss all .NET code as being
written by "script kiddies".

[Journey To The Center of The Earth]

Jeff_Relf wrote:

But I had to radically modify my HTM_TXT to properly handle multilined tags: And, because HTM_TXT is merely a demo of code from X.CPP

I see, so you admit that all you do is write sample code...but
completely unextensible.

The c# folks just proved how inflexible c++ code is because you have to
set up a static starting point and then produce a result.

Whereas us c# people have been able to dynamically change our code very
quickly as you move the goal posts around to suit yourself.

[Tom Shelton]

In article <2s********************@speakeasy.net>, Journey To The Center of The Earth wrote:

Jeff_Relf wrote:

But I had to radically modify my HTM_TXT to properly handle multilined tags:

And, because HTM_TXT is merely a demo of code from X.CPP

I see, so you admit that all you do is write sample code...but
completely unextensible.

The c# folks just proved how inflexible c++ code is because you have to
set up a static starting point and then produce a result.

Whereas us c# people have been able to dynamically change our code very
quickly as you move the goal posts around to suit yourself.

That's why I gave up... Relf will never be satisfied. My attempt met
every one of the criteria that he laid out in the original post - except
for entity translation, and the only reason I didn't handle that was
because I was unclear on what he wanted to have happen (and I did ask
for clarification, but never did see a response). But, when he saw that
I did it in like 3-4 lines of code - suddenly, several more criteria are
added.

It was the same thing with the phone number parsing... With every post
of an answer he had to change the format.

Relf doesn't want to use C#. That's fine, we all make choices and he
has made his. I have no intention of producing another line of code for
him.

--
Tom Shelton

[Jeff_Relf]

Hi Tom_Shelton ( and Bellow ),
Re: The Piss_Poor job you did of converting HTML to plain text,

You thought you could match my HTM_TXT.CPP while your dishes dryed,
....how naive !

Not even Stefan_Simek could faithfully perserve blank lines
or honor all <br> tags.

Try converting BB.HTM Tom:
http://www.Cotse.NET/users/jeffrelf/BB.HTM
Make it look like this:
http://www.Cotse.NET/users/jeffrelf/BB.TXT

http://www.Cotse.NET/users/jeffrelf/HTM_TXT.EXE
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.CPP
http://www.Cotse.NET/users/jeffrelf/HTM_TXT.VCPROJ

The best Simek could do was
to leave a lot of blank lines where the tags used to be
and then consolidate mulitple blank lines into one.

i.e. Simek's htm_txt.cs produces blank lines where I do Not want them,
and omits blank lines where I Do want them.

Does it surprize you that I didn't give you my full specs at first ?
It shouldn't... I didn't want to overwhelm you any more than I already was.

You told Bailo: << That's why I gave up... Relf will never be satisfied.
My attempt met every one of the criteria that he laid out
in the original post - except for entity translation... >>

Unlike Kelsey and Simek, you never figured out how to download an HTML page
( you must do a View_Source before the Save_Page_As ).

You wrote: << It was the same thing with the phone number parsing,
...With every post of an answer he had to change the format. >>

Right... you couldn't understand what my code was doing,
and, therefore, what I required of it.
Nor did you have the patience to have me explain it.
....No surprises there... huh ?

You concluded: << Relf doesn't want to use C#.
That's fine, we all make choices and he has made his.
I have no intention of producing another line of code for him. >>

Using COM and other bloatware is fine in a pinch ( ¡ ¡ ¡ Slo-o-ow ),
but it's not the hallmark of a serious coder.
....You're a Half_Assed coder, Shelton... end of story.

#define is too dangerous for C# kiddies such as you.

#define LOOP while ( 1 )

#define Loop( N ) int J = - 1, LLL = N ; while ( ++ J < LLL )

#define LoopTo( StopCond ) \
while ( Ch && ( Ch = ( uchar ) * ++ P ) \
&& ! ( Ch2 = ( uchar ) P [ 1 ], StopCond ) )

#define LoopXx( Xx ) Xx##P P = 0, B ; int J = -1 ; \
Xx##A BB = Xx.BB, EE = Xx.PP + 1, PP = BB - 1 ; \
if ( BB ) while ( ++ J, B = P = * ++ PP, PP < EE )

[Jeff_Relf]

Hi Bellow, You told me: <<
I see, so you admit that all you do is write sample code
...but completely unextensible. >>

HTM_TXT.EXE demonstrates 47 lines of code from X.CPP,
and X is used daily be me, as it's the best e-mail client newsreader
I've never known... by miles.

And you know it's totally flexible, as you see how I change it all the time.