Web problem

Web problem

[Jeffrey A Law]

I've got no idea how to fix this....


------- Forwarded Message

From:     Richard Stallman <rms@gnu.org>
To:       gcc-sc@goof.com
Date:     Tue, 5 Dec 2000 06:25:10 -0700 (MST)
Subject:  [mark@kuchel.net: gnats forwarding]

Would the right person please DTRT?

- ------- Start of forwarded message -------
X-Authentication-Warning: new-smtp2.ihug.com.au: Host IDENT:mark@p110-tnt1.mel.
ihug.com.au [203.173.160.110] claimed to be bombers
From: Mark Kuchel <mark@kuchel.net>
Date: Mon, 4 Dec 2000 22:49:24 +1100
Subject: gnats forwarding 
To: webmasters@gnu.org
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Length: 1579

Trying to access the gnats database usi	ng konqueror (kde 2.0), from 
the link from the gcc homepage, gcc.gnu.org (down the bottom on the 
left).  The page didn't forward on as expected, (it says hold on... 
forwarding), but rather stops dead.   Submitting this to the KDE 
people, they said that this was a bug on your end that you guys should 
fix.  Below should be their edited response.  

Mark Kuchel

- - ----------  Forwarded Message  ----------
Subject: Bug#16445 acknowledged by developer (konqueror not forwarding )
On Monday 04 December 2000 02:49, Daniel Naber wrote:
> On 2000-12-04 01:49, mark@kuchel.net wrote:
> > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?database=gcc&user=guest&pass
> >word= guest&cmd=login
>
> That server sends
> Refresh: 0; URL= http://gcc.gnu.org/cgi-bin/gnatsweb.pl
> in the header of the http (not HTML). IMHO that's not even part of
> HTTP/1.1.

Hello,

That entry belongs in the actual html page and not in the HTTP header.
BTW, it does not work with Netscape either for the exact reasons stated
above.  Please inform the web master about your problem.

Thanks for the report, though we were unable to help you out :(

Regards,
Dawit A.

- - -------------------------------------------------------

- - -- 
- - --------------------------------------------------------------------
| Mark Kuchel            |                                         |
| mark@kuchel.net        | www.kuchel.net                          |
| Phone: 03 9387 4668    | Mobile: 0438 650 150                    |
- - --------------------------------------------------------------------
- ------- End of forwarded message -------



------- End of Forwarded Message

Re: Web problem

[Phil Edwards]

> > On 2000-12-04 01:49, mark@kuchel.net wrote:
> > > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?database=gcc&user=guest&pass
> > >word= guest&cmd=login

"refresh" isn't the only problem with that version of gnatsweb.

  http://sources.redhat.com/ml/overseers/2000-q4/msg00174.html

If you following the referenced URLs and read through the (short) thread,
you'll see a note that gnatsweb 2.7 (beta?) fixes at least one serious
problem.  I seem to recall getting 2.7 to look at and finding that it
requires a newer version of GNATS, maybe.  I think it fixes the refresh
stuff too, but I don't remember.


Phil

-- 
pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.

Re: Web problem

[Tom Tromey]

Jeff> I've got no idea how to fix this....

gnatsweb uses cookies and this redirect trick to work.  IMNSHO it is
an ugly approach, but the gnatsweb author is quite adamant about it.

I suggest forwarding it as a bug report to <bug-gnats@gnu.org>.

However, there might not be anybody maintaining gnats.  So, we'd have
to fix it ourselves.  Unfortunately I don't understand the problem
well enough to fix it.

KDE-dude> That entry belongs in the actual html page and not in the
KDE-dude> HTTP header.  BTW, it does not work with Netscape either for
KDE-dude> the exact reasons stated above.

First, it does work with netscape, at least whatever netscape I am
using.  I use gnats almost every day.

Second, would it suffice to move the header from the http reply
section to be the actual first line of the body?  That seems weird to
me.

Tom

Re: Web problem

[Ian Lance Taylor]

   From: Tom Tromey <tromey@redhat.com>
   Date: 05 Dec 2000 11:03:28 -0700

   KDE-dude> That entry belongs in the actual html page and not in the
   KDE-dude> HTTP header.  BTW, it does not work with Netscape either for
   KDE-dude> the exact reasons stated above.

   First, it does work with netscape, at least whatever netscape I am
   using.  I use gnats almost every day.

   Second, would it suffice to move the header from the http reply
   section to be the actual first line of the body?  That seems weird to
   me.

I assume they are suggesting using a META tag.  For example, here is
the start of the source of www.cnn.com, which automatically refreshes
itself every 30 minutes (something I have taken advantage of during
this election silly season).

<HTML>
<HEAD>
        <TITLE>CNN.com</TITLE>
        <META http-equiv="REFRESH" content="1800; URL= http://www.cnn.com/ ?">

Needless to say, this will probably only work with a browser which
supports the Refresh: header.  None of this is standard as far as I
know.

Ian

Re: Web problem

[Gerald Pfeifer]

On Tue, 5 Dec 2000, Phil Edwards wrote:
>   http://sources.redhat.com/ml/overseers/2000-q4/msg00174.html
>
> If you following the referenced URLs and read through the (short) thread,
> you'll see a note that gnatsweb 2.7 (beta?) fixes at least one serious
> problem.

gnatsweb 2.6 is the last version I've been able to find and some weeks
ago Tom confirmed that we have the most current version installed.

Tom, do we have 2.6 or already some 2.7 (beta)? Does anyone have a URL
for gnatsweb (different from < http://sources.redhat.com/gnats/ >?)

On 5 Dec 2000, Tom Tromey wrote:
> I suggest forwarding it as a bug report to <bug-gnats@gnu.org>.

/dev/null :-(

> However, there might not be anybody maintaining gnats.  So, we'd have
> to fix it ourselves.

GNATS is essentially unmaintained. gnatsweb is definitely unmaintained:
I sent a patch which I had developed for gcc.gnu.org to the original
author, the address above and twice also to the current list a sources
-- no response whatsoever. :-(

> Second, would it suffice to move the header from the http reply
> section to be the actual first line of the body?  That seems weird to
> me.

I think that should become (or copied as?) a <META> tag in the HTML
<HEAD>.

Any volunteers interested in maintaining gnatsweb? I'm nearly fed up
sufficiently to put a new version with my original patch and any patch
we will come up with here up on my server and announce it as a new offical
release.

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Web problem

[Jason Molenda]

Hey wait, I work for Yahoo, I can pretend I'm an expert or something. ;-)

> > That server sends
> > Refresh: 0; URL= http://gcc.gnu.org/cgi-bin/gnatsweb.pl
> > in the header of the http (not HTML). IMHO that's not even part of
> > HTTP/1.1.
> 
> Hello,
> 
> That entry belongs in the actual html page and not in the HTTP header.
> BTW, it does not work with Netscape either for the exact reasons stated
> above.  Please inform the web master about your problem.


The whole refresh mechanism is a netscape extension.  I'd be
surprised if netscape didn't support it.

It is true that HTTP 1.0 and 1.1 do not include anything about a
Refresh:  header.  And HTML 4.0 doesn't include anything about
supporting a <META HTTP-EQIV="refresh">, either (it just mentions
the existence of the META tag and the HTTP-EQUIV bit, nothing about
'Refresh').

Perl's CGI.pm specifically instructs developers to put the Refresh
specification in the HTTP header, so the gnatsweb author wasn't going
out on a limb here.

Although Konquerer may be standards conforming, that doesn't mean
much on the web.  If they want a pleasant user experience, they're
going to need to support this.

Jason

Re: Web problem

[Tom Tromey]

Jason> Although Konquerer may be standards conforming, that doesn't
Jason> mean much on the web.  If they want a pleasant user experience,
Jason> they're going to need to support this.

I'm not going to look at changing gnatsweb then.  My understanding is
that the refresh hack was needed because of the whole cookie-based
scheme it uses.  Changing that might be a major undertaking.

It would be nice to move gnatsweb away from cookies and towards using
normal URLs.  You still can't refer to certain things from the outside
very easily.  However I doubt I'll be working on that.

Tom

Re: Web problem

[Gerald Pfeifer]

On 13 Dec 2000, Tom Tromey wrote:
>> Although Konquerer may be standards conforming, that doesn't
>> mean much on the web.  If they want a pleasant user experience,
>> they're going to need to support this.
> I'm not going to look at changing gnatsweb then.

Well, the problem is, also lynx does not work with our gnatsweb interface
-- and this is more of a problem than Konquerer, because lynx Users often
do not have an adequate alternativ.

> It would be nice to move gnatsweb away from cookies and towards using
> normal URLs.  You still can't refer to certain things from the outside
> very easily.

Yes, that definitely would be nice.

I have started a ``discussion'' on the gnats mailing list, but it really
doesn't seem like anyone is going to actively maintain this anytime soon.
:-(

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Web problem

[Jeffrey A Law]

I've got no idea how to fix this....


------- Forwarded Message

From:     Richard Stallman <rms@gnu.org>
To:       gcc-sc@goof.com
Date:     Tue, 5 Dec 2000 06:25:10 -0700 (MST)
Subject:  [mark@kuchel.net: gnats forwarding]

Would the right person please DTRT?

- ------- Start of forwarded message -------
X-Authentication-Warning: new-smtp2.ihug.com.au: Host IDENT:mark@p110-tnt1.mel.
ihug.com.au [203.173.160.110] claimed to be bombers
From: Mark Kuchel <mark@kuchel.net>
Date: Mon, 4 Dec 2000 22:49:24 +1100
Subject: gnats forwarding 
To: webmasters@gnu.org
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Length: 1579

Trying to access the gnats database usi	ng konqueror (kde 2.0), from 
the link from the gcc homepage, gcc.gnu.org (down the bottom on the 
left).  The page didn't forward on as expected, (it says hold on... 
forwarding), but rather stops dead.   Submitting this to the KDE 
people, they said that this was a bug on your end that you guys should 
fix.  Below should be their edited response.  

Mark Kuchel

- - ----------  Forwarded Message  ----------
Subject: Bug#16445 acknowledged by developer (konqueror not forwarding )
On Monday 04 December 2000 02:49, Daniel Naber wrote:
> On 2000-12-04 01:49, mark@kuchel.net wrote:
> > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?database=gcc&user=guest&pass
> >word= guest&cmd=login
>
> That server sends
> Refresh: 0; URL= http://gcc.gnu.org/cgi-bin/gnatsweb.pl
> in the header of the http (not HTML). IMHO that's not even part of
> HTTP/1.1.

Hello,

That entry belongs in the actual html page and not in the HTTP header.
BTW, it does not work with Netscape either for the exact reasons stated
above.  Please inform the web master about your problem.

Thanks for the report, though we were unable to help you out :(

Regards,
Dawit A.

- - -------------------------------------------------------

- - -- 
- - --------------------------------------------------------------------
| Mark Kuchel            |                                         |
| mark@kuchel.net        | www.kuchel.net                          |
| Phone: 03 9387 4668    | Mobile: 0438 650 150                    |
- - --------------------------------------------------------------------
- ------- End of forwarded message -------



------- End of Forwarded Message

Re: Web problem

[Gerald Pfeifer]

On 13 Dec 2000, Tom Tromey wrote:
>> Although Konquerer may be standards conforming, that doesn't
>> mean much on the web.  If they want a pleasant user experience,
>> they're going to need to support this.
> I'm not going to look at changing gnatsweb then.

Well, the problem is, also lynx does not work with our gnatsweb interface
-- and this is more of a problem than Konquerer, because lynx Users often
do not have an adequate alternativ.

> It would be nice to move gnatsweb away from cookies and towards using
> normal URLs.  You still can't refer to certain things from the outside
> very easily.

Yes, that definitely would be nice.

I have started a ``discussion'' on the gnats mailing list, but it really
doesn't seem like anyone is going to actively maintain this anytime soon.
:-(

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Web problem

[Ian Lance Taylor]

   From: Tom Tromey <tromey@redhat.com>
   Date: 05 Dec 2000 11:03:28 -0700

   KDE-dude> That entry belongs in the actual html page and not in the
   KDE-dude> HTTP header.  BTW, it does not work with Netscape either for
   KDE-dude> the exact reasons stated above.

   First, it does work with netscape, at least whatever netscape I am
   using.  I use gnats almost every day.

   Second, would it suffice to move the header from the http reply
   section to be the actual first line of the body?  That seems weird to
   me.

I assume they are suggesting using a META tag.  For example, here is
the start of the source of www.cnn.com, which automatically refreshes
itself every 30 minutes (something I have taken advantage of during
this election silly season).

<HTML>
<HEAD>
        <TITLE>CNN.com</TITLE>
        <META http-equiv="REFRESH" content="1800; URL= http://www.cnn.com/ ?">

Needless to say, this will probably only work with a browser which
supports the Refresh: header.  None of this is standard as far as I
know.

Ian

Re: Web problem

[Phil Edwards]

> > On 2000-12-04 01:49, mark@kuchel.net wrote:
> > > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?database=gcc&user=guest&pass
> > >word= guest&cmd=login

"refresh" isn't the only problem with that version of gnatsweb.

  http://sources.redhat.com/ml/overseers/2000-q4/msg00174.html

If you following the referenced URLs and read through the (short) thread,
you'll see a note that gnatsweb 2.7 (beta?) fixes at least one serious
problem.  I seem to recall getting 2.7 to look at and finding that it
requires a newer version of GNATS, maybe.  I think it fixes the refresh
stuff too, but I don't remember.


Phil

-- 
pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.

Re: Web problem

[Tom Tromey]

Jason> Although Konquerer may be standards conforming, that doesn't
Jason> mean much on the web.  If they want a pleasant user experience,
Jason> they're going to need to support this.

I'm not going to look at changing gnatsweb then.  My understanding is
that the refresh hack was needed because of the whole cookie-based
scheme it uses.  Changing that might be a major undertaking.

It would be nice to move gnatsweb away from cookies and towards using
normal URLs.  You still can't refer to certain things from the outside
very easily.  However I doubt I'll be working on that.

Tom

Re: Web problem

[Jason Molenda]

Hey wait, I work for Yahoo, I can pretend I'm an expert or something. ;-)

> > That server sends
> > Refresh: 0; URL= http://gcc.gnu.org/cgi-bin/gnatsweb.pl
> > in the header of the http (not HTML). IMHO that's not even part of
> > HTTP/1.1.
> 
> Hello,
> 
> That entry belongs in the actual html page and not in the HTTP header.
> BTW, it does not work with Netscape either for the exact reasons stated
> above.  Please inform the web master about your problem.


The whole refresh mechanism is a netscape extension.  I'd be
surprised if netscape didn't support it.

It is true that HTTP 1.0 and 1.1 do not include anything about a
Refresh:  header.  And HTML 4.0 doesn't include anything about
supporting a <META HTTP-EQIV="refresh">, either (it just mentions
the existence of the META tag and the HTTP-EQUIV bit, nothing about
'Refresh').

Perl's CGI.pm specifically instructs developers to put the Refresh
specification in the HTTP header, so the gnatsweb author wasn't going
out on a limb here.

Although Konquerer may be standards conforming, that doesn't mean
much on the web.  If they want a pleasant user experience, they're
going to need to support this.

Jason

Re: Web problem

[Tom Tromey]

Jeff> I've got no idea how to fix this....

gnatsweb uses cookies and this redirect trick to work.  IMNSHO it is
an ugly approach, but the gnatsweb author is quite adamant about it.

I suggest forwarding it as a bug report to <bug-gnats@gnu.org>.

However, there might not be anybody maintaining gnats.  So, we'd have
to fix it ourselves.  Unfortunately I don't understand the problem
well enough to fix it.

KDE-dude> That entry belongs in the actual html page and not in the
KDE-dude> HTTP header.  BTW, it does not work with Netscape either for
KDE-dude> the exact reasons stated above.

First, it does work with netscape, at least whatever netscape I am
using.  I use gnats almost every day.

Second, would it suffice to move the header from the http reply
section to be the actual first line of the body?  That seems weird to
me.

Tom

Re: Web problem

[Gerald Pfeifer]

On Tue, 5 Dec 2000, Phil Edwards wrote:
>   http://sources.redhat.com/ml/overseers/2000-q4/msg00174.html
>
> If you following the referenced URLs and read through the (short) thread,
> you'll see a note that gnatsweb 2.7 (beta?) fixes at least one serious
> problem.

gnatsweb 2.6 is the last version I've been able to find and some weeks
ago Tom confirmed that we have the most current version installed.

Tom, do we have 2.6 or already some 2.7 (beta)? Does anyone have a URL
for gnatsweb (different from < http://sources.redhat.com/gnats/ >?)

On 5 Dec 2000, Tom Tromey wrote:
> I suggest forwarding it as a bug report to <bug-gnats@gnu.org>.

/dev/null :-(

> However, there might not be anybody maintaining gnats.  So, we'd have
> to fix it ourselves.

GNATS is essentially unmaintained. gnatsweb is definitely unmaintained:
I sent a patch which I had developed for gcc.gnu.org to the original
author, the address above and twice also to the current list a sources
-- no response whatsoever. :-(

> Second, would it suffice to move the header from the http reply
> section to be the actual first line of the body?  That seems weird to
> me.

I think that should become (or copied as?) a <META> tag in the HTML
<HEAD>.

Any volunteers interested in maintaining gnatsweb? I'm nearly fed up
sufficiently to put a new version with my original patch and any patch
we will come up with here up on my server and announce it as a new offical
release.

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/