Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-11-01 Thread Stig S. Bakken
On Wed, 2002-10-09 at 09:35, Yasuo Ohgaki wrote:
 Colin Viebrock wrote:
  I really think the best solution (not perfect, but best) is to specify
  some fonts so the pages look nice, and hard code in the ISO-8859-1 font
 
 hard code in the ISO-8859-1 font means assuming ISO 8859-1 and
 use ISO 8859-1 type face by converting chars to entities?
 
 Take a look at this page, for example.
 
 http://czyborra.com/charsets/iso8859.html
 
 Do you see converting text to entity assuming ISO 8859-1
 breaks not only multibyte encoding but also other ISO 8859
 encoding?
 
 I'm well aware of that I'm suggesting to make phpinfo()
 non XHTML, since it is more useful if it is not confirm
 XHTML perfectly.
 
 If use of HTML entities are preferred, only text that
 needs HTML entities should use entities. e.g. Names.
 Isn't using entities for names or like enough?
 
 Please no automatic entity conversion assuming ISO 8859-1.
 Thanks.

A bit late, but I'd like to throw in my .02EUR.

When dealing with multiple languages, the only reasonable charset to
support internally is Unicode, encoded in utf-8.

Now, while MSIE supports utf-8, it doesn't sent the Accept-Encoding
header.  NS4 is AFAIK the only browser that explicitly announces being
able to handle utf-8 in the request, but it's not a big issue to figure
out if the user agent is from a browser that can deal with utf-8.

The real solution would be using iconv on the output buffer to change
the utf-8 to whatever charset the browser prefers, or if iconv is not
available, try converting to 8859-1 and replace characters that don't
fit to ?.  IMHO this is the only thing that will work for everyone.

 - Stig

-- 
Stig Sæther Bakken, Fast Search  Transfer ASA, Trondheim, Norway
http://pear.php.net/wishlist.php/ssb


--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-11-01 Thread Marcus Boerger
At 01:40 02.11.2002, Stig S. Bakken wrote:

On Wed, 2002-10-09 at 09:35, Yasuo Ohgaki wrote:
 Colin Viebrock wrote:
  I really think the best solution (not perfect, but best) is to specify
  some fonts so the pages look nice, and hard code in the ISO-8859-1 font

 hard code in the ISO-8859-1 font means assuming ISO 8859-1 and
 use ISO 8859-1 type face by converting chars to entities?

 Take a look at this page, for example.

 http://czyborra.com/charsets/iso8859.html

 Do you see converting text to entity assuming ISO 8859-1
 breaks not only multibyte encoding but also other ISO 8859
 encoding?

 I'm well aware of that I'm suggesting to make phpinfo()
 non XHTML, since it is more useful if it is not confirm
 XHTML perfectly.

 If use of HTML entities are preferred, only text that
 needs HTML entities should use entities. e.g. Names.
 Isn't using entities for names or like enough?

 Please no automatic entity conversion assuming ISO 8859-1.
 Thanks.

A bit late, but I'd like to throw in my .02EUR.

When dealing with multiple languages, the only reasonable charset to
support internally is Unicode, encoded in utf-8.

Now, while MSIE supports utf-8, it doesn't sent the Accept-Encoding
header.  NS4 is AFAIK the only browser that explicitly announces being
able to handle utf-8 in the request, but it's not a big issue to figure
out if the user agent is from a browser that can deal with utf-8.

The real solution would be using iconv on the output buffer to change
the utf-8 to whatever charset the browser prefers, or if iconv is not
available, try converting to 8859-1 and replace characters that don't
fit to ?.  IMHO this is the only thing that will work for everyone.


Since mbstring is better integrated in php i HOPE that will be done by
mbstring (and it is not even dependant on another library).

marcus


--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-10 Thread Moriyoshi Koizumi

Hi,

Although I have no font problem differently to yasuo, (Actually I have got 
Arial Unicode MT) but I've experienced character encoding problem since 
your recent patch on ext/standard/info.c.

How come you concluded that part is unnecesary? 

 I really think the best solution (not perfect, but best) is to specify
 some fonts so the pages look nice, and hard code in the ISO-8859-1 font.

I think it would not be the best solution, but just the last one of best 
solutions we could choose.

AFAIK, IE and Mozilla automatically substitute the font when a character 
being displayed is not a part of the font which has been determined best 
for the specified encoding after HTML(or XML) parsing, even if 
that character is described in HTML entity form. I think there is few 
browsers that may have CSS font/encoding related problem.

But I don't see XHTML conformance could be a good reason to discard the 
support for encodings other than ISO-8859-1... Hmm, I could say there are 
actually many users in Japan who use phpinfo() to check if the proper 
value have been submitted by the browser in the proper encoding... Yet I 
have to say it's always problematic because of the lack of commonly 
accepted specification for the encoding(not about url-encoding) used in 
url-embedded (namely I mean GET method) queries.


Regards,

Moriyoshi Koizumi



-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-09 Thread Derick Rethans

On Wed, 9 Oct 2002, Yasuo Ohgaki wrote:

 Colin Viebrock wrote:
  This is getting a little more complicated than I think is necessary.
  There are two issues here, I think:
  
  a) Fonts.  Some people didn't like Arial, so I reverted to letter the
  browser decide.  Some people didn't like that, and they'd like the font
  specifications back in.
 
 This will simply break output under some browser.
 It's more important to show info, but show a little nicely on some
 systems while printing garages on some systems.
 
 If anyone prefer any specific font, they should use their own CSS.

For once I agree with yasuo here :) phpinfo() doesn't need to look 
pretty, it's for _debug_ output.

Derick

--

---
 Derick Rethans   http://derickrethans.nl/ 
 JDI Media Solutions
--[ if you hold a unix shell to your ear, do you hear the c? ]-


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-09 Thread Wez Furlong

On 10/09/02, Yasuo Ohgaki [EMAIL PROTECTED] wrote:
 AFAIK, there is no Arial that includes CJK.

Arial Unicode MS (that 22MB TTF you'll find in \windows\fonts) seems
to do a good job of including virtually all characters known to man.

BTW: Colin - there is a bug report about phpinfo() - you are not
efree()ing your htmlentities()-ed strings and that causes a load avg.
of 2 when PHP shuts down and cleans up and warns about it 380+ times :-)
The bug report is titled something like php causes a load avg of 2.

--Wez.



-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-09 Thread Melvyn Sopacua

At 08:34 9-10-2002, Derick Rethans wrote:

For once I agree with yasuo here :) phpinfo() doesn't need to look
pretty, it's for _debug_ output.

Which means it needs to be as clear as possible. Formatting it for readibility
is a plus - you don't want output that's hard to read, as you're already 
debugging.

But if this breaks stuff for a certain charset/encoding/fontspec 
combination, than
it should be removed, for the very same reason, as it's not readable.

[ I just wish, W3C and browser vendors would get on the same page for once, 
sigh ]


Met vriendelijke groeten / With kind regards,

Webmaster IDG.nl
Melvyn Sopacua


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-09 Thread Yasuo Ohgaki

Wez Furlong wrote:
 On 10/09/02, Yasuo Ohgaki [EMAIL PROTECTED] wrote:
 
AFAIK, there is no Arial that includes CJK.
 
 
 Arial Unicode MS (that 22MB TTF you'll find in \windows\fonts) seems
 to do a good job of including virtually all characters known to man.

Hmm...
I don't have it under my W2k.
MS Japan should provide it :)

--
Yasuo Ohgaki


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Colin Viebrock

   Log:
   don't define fonts ... use the browser defaults
 
 Reason being?

Yasuo Ohgaki's post earlier today said:

  First problem, CSS should not contain specific font
  name. Otherwise, characters may be broken under some
  browsers when font does not have type faces needed.

Although, to be honest, I think his problem is more with the character
set that is being sent with phpinfo pages.  I'm guessing he has a
Japanese version of Arial, that's missing some of the characters like
eactue; etc..  Since the page is probably outputting
charset=shift-JIS or whatever, the page looks broken.

I'm waiting to hear a response from him on this issue, which might
affect my changes to the css page.


 The definitions end in generic font families. There's no shame in
 defining a __readable__ __preference__.
 Especially Andale Mono, since it is one of the few monotype fonts,
 that has a readable difference between the l and a 1 and the O and 0.
 
 I've seen more people fiddling with their fonts, because they're pretty
 than because they're readable. Docs should aim to be readable.

Agreed.  Which is why the default fonts on people's browsers should be
set to be readable already, no?

Also, some people have emailed suggesting I remove them because of the
CSS styling that they apply themselves to the phpinfo() output, using
their own font defs.


 What's more problematic in the css code below, is the 75% default on td/th.
 
 If you nest that, the inner td, is 75% of 75% of the body.

But there is no nesting of tables in phpinfo()'s output.

- Colin


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




Re: [PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Melvyn Sopacua

At 00:28 9-10-2002, Colin Viebrock wrote:

Log:
don't define fonts ... use the browser defaults
 
  Reason being?

Yasuo Ohgaki's post earlier today said:

   First problem, CSS should not contain specific font
   name. Otherwise, characters may be broken under some
   browsers when font does not have type faces needed.

I've had this discussion with respect to the euro-sign also. Apparently,
there are 2 different standards, which define different places for the
euro sign in fonts - one on the 'international currency' glyph, and
the 'microsoft' way, in unicode, with an OS-hack for non-unicode fonts.
[1]

I lack the specifics right now, but can look them up if you're interested.

Although, to be honest, I think his problem is more with the character
set that is being sent with phpinfo pages.  I'm guessing he has a
Japanese version of Arial, that's missing some of the characters like
eactue; etc..  Since the page is probably outputting
charset=shift-JIS or whatever, the page looks broken.

Agreed, as [1] is a problem with the font implementation and should be
taken up with the vendor of either the font or the OS. 'The web' cannot
be held responsible for how the browser interprets euro; (or a charset
header and bytes from a different one for that matter).

I'm waiting to hear a response from him on this issue, which might
affect my changes to the css page.


  The definitions end in generic font families. There's no shame in
  defining a __readable__ __preference__.
  Especially Andale Mono, since it is one of the few monotype fonts,
  that has a readable difference between the l and a 1 and the O and 0.
 
  I've seen more people fiddling with their fonts, because they're pretty
  than because they're readable. Docs should aim to be readable.

Agreed.  Which is why the default fonts on people's browsers should be
set to be readable already, no?

For a sysadmin yes. For a designer (large group of new users, since articles
on flash+php seem to emerge everywhere) that's questionable :-)

The CSS mechanism was however designed to express preferences and fallback.

Also, some people have emailed suggesting I remove them because of the
CSS styling that they apply themselves to the phpinfo() output, using
their own font defs.

Ehm - that is what user-defined stylesheets on the client are for. And settings
like 'always use my fonts'.

But anyway - not worth a long discussion - just a tiny bit worried about 
bogusing
bug reports like 'phpinfo page suddenly looks weird'.
And it's already much better than deprecated font-tags :-)

  What's more problematic in the css code below, is the 75% default on td/th.
 
  If you nest that, the inner td, is 75% of 75% of the body.

But there is no nesting of tables in phpinfo()'s output.

yet :-)


- Colin


--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php


Met vriendelijke groeten / With kind regards,

Webmaster IDG.nl
Melvyn Sopacua


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Yasuo Ohgaki

Colin Viebrock wrote:
 I'm waiting to hear a response from him on this issue, which might
 affect my changes to the css page.
 

You forgot to disable automatic char to entities conversion

e.g. Followings cannot be Japanese text obviously
Ccedil;macr; 9middot;icirc; 9AElig;uuml; 
middot;icirc;Iacute;Euml;AElig;uuml;
This is EUC-JP text output from phpinfo(), btw.

--
Yasuo Ohgaki


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Yasuo Ohgaki

Colin Viebrock wrote:
 Yasuo Ohgaki's post earlier today said:
First problem, CSS should not contain specific font
name. Otherwise, characters may be broken under some
browsers when font does not have type faces needed.

 
 Although, to be honest, I think his problem is more with the character
 set that is being sent with phpinfo pages.  I'm guessing he has a

It is well known font issue. Some browser won't try to use other
fonts when font specified is missing type faces. If you really would
like to list specific font, list CJK and other fonts first then western
fonts.

Anyway, you have been resolved this.

 Japanese version of Arial, that's missing some of the characters like
 eactue; etc..  Since the page is probably outputting
 charset=shift-JIS or whatever, the page looks broken.

AFAIK, there is no Arial that includes CJK.

--
Yasuo Ohgaki


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Colin Viebrock

This is getting a little more complicated than I think is necessary.
There are two issues here, I think:

a) Fonts.  Some people didn't like Arial, so I reverted to letter the
browser decide.  Some people didn't like that, and they'd like the font
specifications back in.

I'm going to add the font specifications back in again.  I like Arial,
personally, but I will put Verdana as the first font.  This is in
keeping with the look of php.net anyway, so I can justify it.  There is
nothing stopping anyone from specifying use my own fonts in their
browsers.

b) Charsets.  I don't pretend to know what the charset deciding code in
ext/standard/info.c is trying to do.  I guess it's looking to see if
your browser supports multi-byte charsets, or what encoding you prefer
... and then defaulting to US-ASCII if nothing else works.  Seemed
reasonable to me.

I suppose, what Yasuo has pointed out, is that some (many? all?)
multibyte fonts or versions of those fonts don't support the HTML entity
encoding that is used on some of the phpinfo() pages ... mainly people's
names in the credits pages, but also any  or  or  that show up in
other data.  These need to be encoded *somehow* if the page is to be
XHTML compatible, and I'm not about to implement utf8_encode() in C.  :)

I don't know what the solution to this is.  My instinct is to ask why
don't we just hardcode the pages to use US-ASCII, or ISO-8859-1?  Would
this not solve the charset issues?

- Colin


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Yasuo Ohgaki

Colin Viebrock wrote:
 This is getting a little more complicated than I think is necessary.
 There are two issues here, I think:
 
 a) Fonts.  Some people didn't like Arial, so I reverted to letter the
 browser decide.  Some people didn't like that, and they'd like the font
 specifications back in.

This will simply break output under some browser.
It's more important to show info, but show a little nicely on some
systems while printing garages on some systems.

If anyone prefer any specific font, they should use their own CSS.

 I don't know what the solution to this is.  My instinct is to ask why
 don't we just hardcode the pages to use US-ASCII, or ISO-8859-1?  Would
 this not solve the charset issues?

No.

Why you bother?

Letting browser decide is more feasible and most importantly,
it works.

If you are not certain how to make phpinfo() work under most
environment, I can help.

--
Yasuo Ohgaki



-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Colin Viebrock

 a) Fonts.  Some people didn't like Arial, so I reverted to letter the
 browser decide.  Some people didn't like that, and they'd like the font
 specifications back in.
 
 This will simply break output under some browser.
 It's more important to show info, but show a little nicely on some
 systems while printing garages on some systems.

But the old phpinfo() specified fonts, so are you saying it was broken
then too?


 If anyone prefer any specific font, they should use their own CSS.

But I bet 99% of the people will have versions of either Arial or
Verdana or Helvetica that *will* work.  Why not make life easier for the
majority, since the 1% is going to have to specify their own font anyway?


 I don't know what the solution to this is.  My instinct is to ask why
 don't we just hardcode the pages to use US-ASCII, or ISO-8859-1?  Would
 this not solve the charset issues?
 
 No.
 
 Why you bother?
 
 Letting browser decide is more feasible and most importantly,
 it works.

Okay ... so don't output the charset meta tag at all?


 If you are not certain how to make phpinfo() work under most
 environment, I can help.

If you can make it work under most environment *and* look good (i.e. as
much like it used to, sans-serif fonts and all), then great!  CVS is a
wonderful thing. :)

- Colin


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Yasuo Ohgaki

Colin Viebrock wrote:
a) Fonts.  Some people didn't like Arial, so I reverted to letter the
browser decide.  Some people didn't like that, and they'd like the font
specifications back in.

This will simply break output under some browser.
It's more important to show info, but show a little nicely on some
systems while printing garages on some systems.
 
 
 But the old phpinfo() specified fonts, so are you saying it was broken
 then too?

Yes and No. It works withh my browser, but I just happen to know
some browsers do not like it. Let's be more conservative.

I don't know what the solution to this is.  My instinct is to ask why
don't we just hardcode the pages to use US-ASCII, or ISO-8859-1?  Would
this not solve the charset issues?

No.

Why you bother?

Letting browser decide is more feasible and most importantly,
it works.
 
 
 Okay ... so don't output the charset meta tag at all?

That's better. And you must not try to convert chars to
entities. This is the worst thing... Text cannot be read
even with HTML sources.

--
Yasuo Ohgaki


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php




[PHP-DEV] Re: [PHP-CVS] cvs: php4 /ext/standard css.c

2002-10-08 Thread Colin Viebrock

 Yes and No. It works withh my browser, but I just happen to know
 some browsers do not like it. Let's be more conservative.

I'd rather have a nice looking page that works on 99% of the browsers,
than a not-so-nice looking one that works on 100%.


 Okay ... so don't output the charset meta tag at all?
 
 That's better. And you must not try to convert chars to
 entities. This is the worst thing... Text cannot be read
 even with HTML sources.

But the characters need to be converted.  It isn't valid XHTML if we
don't specify a charset.

Testing on Mozilla, if I have a document that contains:

Hello Theacute;o, welcome to my Straße!

This only renders completely properly if the charset is ISO-8859-1.
HTML entities render in UTF-8 and US-ASCII, but the ß does not.  Even
using the character set your emails to me come in (ISO-2022-JP)
renders the HTML entities correctly, but fails on the un-encoded character.

I really think the best solution (not perfect, but best) is to specify
some fonts so the pages look nice, and hard code in the ISO-8859-1 font.

- Colin


-- 
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php