[adam@tabris.net: GNATS Login]

[adam@tabris.net: GNATS Login]

[Christopher Faylor]

This has probably been discussed before, but is there any way to avoid
the problem described below?  It is annoying to have to login before
viewing a PR.

It would be nice if you were just logged into guest automatically.

cgf

----- Forwarded message from Adam Schrotenboer <adam@tabris.net> -----

From: Adam Schrotenboer <adam@tabris.net>
To: gcc@gcc.gnu.org
Subject: GNATS Login was Re: Linux Kernel, KDE & GCC 3.0.3
Date: Fri, 14 Dec 2001 21:26:41 -0500
Organization: Dome-S-Isle Data
In-Reply-To: <20011214105513.A5977@mediaone.net>

Something that has been getting on my nerves is that it is impossible to (at 
least thru a direct link) read a PR w/o having a GNATS login.

Any plans to make the GNATS db available for public access?

On Friday 14 December 2001 10:55, Craig Rodrigues wrote:
> On Fri, Dec 14, 2001 at 02:10:32AM -0500, Adam Schrotenboer wrote:
> > Have the KDE issues been resolved in 3.0.3? also, what of the Kernel 2.4
> > 8139too.c ICE?
>
> The 8139too.c ICE should be fixed:
> http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4529&database=gcc
>
> I'm not sure about the GLIBC and KDE issues.

-- 
tabris
Courage is resistance of fear, mastery of fear, not absence of fear.

----- End forwarded message -----

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor wrote:
> This has probably been discussed before, but is there any way to avoid
> the problem described below?  It is annoying to have to login before
> viewing a PR.

I'm not sure what his complaint is... of course anyone can view the
PRs by using the guest/guest login which is documented on the login page.
In fact, any arbitrary name/password will give you read-only access
unless someone has changed the configurations since I set them up.

If you've logged in, doesn't gnatsweb store the login info in a
cookie, and subsequent accesses get the name/password out of the cookie?

I agree that it'd be sweet if gnatsweb could assume a guest/guest login
if no gnatsweb cookie is presented (and people would have to click on a
'login' button to log in with a write-access acct), but I don't see the
current situation as especially unbearable.  Maybe this user has cookies
disabled.

Disclaimer - I haven't looked at gnatsweb in at least a year.  I don't
think this core behavior has changed since I last looked, but it's possible.

J


> 
> It would be nice if you were just logged into guest automatically.
> 
> cgf
> 
> ----- Forwarded message from Adam Schrotenboer <adam@tabris.net> -----
> 
> From: Adam Schrotenboer <adam@tabris.net>
> To: gcc@gcc.gnu.org
> Subject: GNATS Login was Re: Linux Kernel, KDE & GCC 3.0.3
> Date: Fri, 14 Dec 2001 21:26:41 -0500
> Organization: Dome-S-Isle Data
> In-Reply-To: <20011214105513.A5977@mediaone.net>
> 
> Something that has been getting on my nerves is that it is impossible to (at 
> least thru a direct link) read a PR w/o having a GNATS login.
> 
> Any plans to make the GNATS db available for public access?
> 
> On Friday 14 December 2001 10:55, Craig Rodrigues wrote:
> > On Fri, Dec 14, 2001 at 02:10:32AM -0500, Adam Schrotenboer wrote:
> > > Have the KDE issues been resolved in 3.0.3? also, what of the Kernel 2.4
> > > 8139too.c ICE?
> >
> > The 8139too.c ICE should be fixed:
> > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4529&database=gcc
> >
> > I'm not sure about the GLIBC and KDE issues.
> 
> -- 
> tabris
> Courage is resistance of fear, mastery of fear, not absence of fear.
> 
> ----- End forwarded message -----

Re: [adam@tabris.net: GNATS Login]

[Christopher Faylor]

On Fri, Dec 14, 2001 at 07:44:15PM -0800, Jason Molenda wrote:
>On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor wrote:
>>This has probably been discussed before, but is there any way to avoid
>>the problem described below?  It is annoying to have to login before
>>viewing a PR.
>
>I'm not sure what his complaint is...  of course anyone can view the
>PRs by using the guest/guest login which is documented on the login
>page.  In fact, any arbitrary name/password will give you read-only
>access unless someone has changed the configurations since I set them
>up.

He's complaining that he has to *first* login and *then* view the page.
He can't just click on the URL and have the page displayed automatically.

>If you've logged in, doesn't gnatsweb store the login info in a cookie,
>and subsequent accesses get the name/password out of the cookie?
>
>I agree that it'd be sweet if gnatsweb could assume a guest/guest login
>if no gnatsweb cookie is presented (and people would have to click on a
>'login' button to log in with a write-access acct), but I don't see the
>current situation as especially unbearable.  Maybe this user has
>cookies disabled.

No one said it was unbearable.  It just isn't user friendly.  I just
tried this with a browser where I hadn't logged into gnats and got a
login screen rather than the bug.  It's obvious what needs to be done at
this point.  It's not obvious why we require it.  That's all.

I was just asking if it was possible to have things changed so that the
default was to login as guest, which would bypass this minor problem.

cgf

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:55:22PM -0500, Christopher Faylor wrote:

> 
> I was just asking if it was possible to have things changed so that the
> default was to login as guest, which would bypass this minor problem.
> 

My bad, I didn't get it.

gnatsweb doesn't do this right now, but I doubt it would be especially
difficult to hack it to assume a guest/guest login if no cookie is
presented.  Doesn't using gnats require the use of cookies?  A
change like this might make it usable for cookie-adverse people, at
least for read-only ooperations.

J

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 08:08:37PM -0800, Jason Molenda wrote:

> gnatsweb doesn't do this right now, but I doubt it would be especially
> difficult to hack it to assume a guest/guest login if no cookie is
> presented.  Doesn't using gnats require the use of cookies?  A
> change like this might make it usable for cookie-adverse people, at
> least for read-only ooperations.
> 

Some thought would have to be given to how a person with write access can
log in.  Maybe if someone goes specifically to the login screen, we'd not
force a guest/guest cookie on them.  Not insurmountable, but it's one more
small complication that'd have to be considered.

J

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 08:08:37PM -0800, Jason Molenda wrote:

> gnatsweb doesn't do this right now, but I doubt it would be especially
> difficult to hack it to assume a guest/guest login if no cookie is
> presented.  Doesn't using gnats require the use of cookies?  A
> change like this might make it usable for cookie-adverse people, at
> least for read-only ooperations.
> 

Some thought would have to be given to how a person with write access can
log in.  Maybe if someone goes specifically to the login screen, we'd not
force a guest/guest cookie on them.  Not insurmountable, but it's one more
small complication that'd have to be considered.

J

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

[-- Attachment #1: Type: text/plain, Size: 732 bytes --]

On Fri, Dec 14, 2001 at 08:08:37PM -0800, Jason Molenda wrote:

> gnatsweb doesn't do this right now, but I doubt it would be especially
> difficult to hack it to assume a guest/guest login if no cookie is
> presented.  

It wasn't.  I only tested this a little bit, but I bet you could
get away with this patch.  I didn't look at the sourceware case
where the database is not predetermined -- I think it would still
work as long as the database= line is included in the URL, but I
haven't really thought about this.

Incidentally, everything works OK without cookies as long as you're
satisfied with read-only access.

I haven't applied this to the gcc gnatsweb, I don't want to mess
with other people's stuff in this case.

Jason

[-- Attachment #2: pa.txt --]
[-- Type: text/plain, Size: 973 bytes --]

--- gnatsweb.pl	Fri Dec 14 12:02:52 2001
+++ gnatswebber.pl	Fri Dec 14 21:45:49 2001
@@ -3569,12 +3569,18 @@ sub main
   elsif(!$global_prefs{'database'}
         || !$db_prefs{'user'} || !$db_prefs{'password'})
   {
-    # We don't have username/password/database; give login page then
-    # redirect to the url they really want (self_url).
-    login_page($q->self_url());
-    exit();
+    #GCC-LOCAL begin: No cookie?  They're a read-only guest/guest.
+    $global_prefs{'database'} = "gcc";
+    $db_prefs{'user'} = "guest";
+    $db_prefs{'password'} = "guest";
+
+#    # We don't have username/password/database; give login page then
+#    # redirect to the url they really want (self_url).
+#    login_page($q->self_url());
+#    exit();
   }
-  elsif($cmd eq 'submit')
+  if($cmd eq 'submit')
+    #GCC-LOCAL end.
   {
     # User is submitting a new PR.  Store cookie because email address may
     # have changed.  This facilitates entering bugs the next time.

Re: [adam@tabris.net: GNATS Login]

[Gerald Pfeifer]

On Fri, 14 Dec 2001, Jason Molenda wrote:
> I haven't applied this to the gcc gnatsweb, I don't want to mess
> with other people's stuff in this case.

Don't hesitate to mess^Wtoy around with this. ;-)

(My GCC time budget has been seriously limited lately, and your help
is certainly most welcome!)

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

Re: [adam@tabris.net: GNATS Login]

[Gerald Pfeifer]

On Fri, 14 Dec 2001, Jason Molenda wrote:
> I haven't applied this to the gcc gnatsweb, I don't want to mess
> with other people's stuff in this case.

Don't hesitate to mess^Wtoy around with this. ;-)

(My GCC time budget has been seriously limited lately, and your help
is certainly most welcome!)

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

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 08:08:37PM -0800, Jason Molenda wrote:

> gnatsweb doesn't do this right now, but I doubt it would be especially
> difficult to hack it to assume a guest/guest login if no cookie is
> presented.  

It wasn't.  I only tested this a little bit, but I bet you could
get away with this patch.  I didn't look at the sourceware case
where the database is not predetermined -- I think it would still
work as long as the database= line is included in the URL, but I
haven't really thought about this.

Incidentally, everything works OK without cookies as long as you're
satisfied with read-only access.

I haven't applied this to the gcc gnatsweb, I don't want to mess
with other people's stuff in this case.

Jason

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:55:22PM -0500, Christopher Faylor wrote:

> 
> I was just asking if it was possible to have things changed so that the
> default was to login as guest, which would bypass this minor problem.
> 

My bad, I didn't get it.

gnatsweb doesn't do this right now, but I doubt it would be especially
difficult to hack it to assume a guest/guest login if no cookie is
presented.  Doesn't using gnats require the use of cookies?  A
change like this might make it usable for cookie-adverse people, at
least for read-only ooperations.

J

Re: [adam@tabris.net: GNATS Login]

[Christopher Faylor]

On Fri, Dec 14, 2001 at 07:44:15PM -0800, Jason Molenda wrote:
>On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor wrote:
>>This has probably been discussed before, but is there any way to avoid
>>the problem described below?  It is annoying to have to login before
>>viewing a PR.
>
>I'm not sure what his complaint is...  of course anyone can view the
>PRs by using the guest/guest login which is documented on the login
>page.  In fact, any arbitrary name/password will give you read-only
>access unless someone has changed the configurations since I set them
>up.

He's complaining that he has to *first* login and *then* view the page.
He can't just click on the URL and have the page displayed automatically.

>If you've logged in, doesn't gnatsweb store the login info in a cookie,
>and subsequent accesses get the name/password out of the cookie?
>
>I agree that it'd be sweet if gnatsweb could assume a guest/guest login
>if no gnatsweb cookie is presented (and people would have to click on a
>'login' button to log in with a write-access acct), but I don't see the
>current situation as especially unbearable.  Maybe this user has
>cookies disabled.

No one said it was unbearable.  It just isn't user friendly.  I just
tried this with a browser where I hadn't logged into gnats and got a
login screen rather than the bug.  It's obvious what needs to be done at
this point.  It's not obvious why we require it.  That's all.

I was just asking if it was possible to have things changed so that the
default was to login as guest, which would bypass this minor problem.

cgf

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor wrote:
> This has probably been discussed before, but is there any way to avoid
> the problem described below?  It is annoying to have to login before
> viewing a PR.

I'm not sure what his complaint is... of course anyone can view the
PRs by using the guest/guest login which is documented on the login page.
In fact, any arbitrary name/password will give you read-only access
unless someone has changed the configurations since I set them up.

If you've logged in, doesn't gnatsweb store the login info in a
cookie, and subsequent accesses get the name/password out of the cookie?

I agree that it'd be sweet if gnatsweb could assume a guest/guest login
if no gnatsweb cookie is presented (and people would have to click on a
'login' button to log in with a write-access acct), but I don't see the
current situation as especially unbearable.  Maybe this user has cookies
disabled.

Disclaimer - I haven't looked at gnatsweb in at least a year.  I don't
think this core behavior has changed since I last looked, but it's possible.

J


> 
> It would be nice if you were just logged into guest automatically.
> 
> cgf
> 
> ----- Forwarded message from Adam Schrotenboer <adam@tabris.net> -----
> 
> From: Adam Schrotenboer <adam@tabris.net>
> To: gcc@gcc.gnu.org
> Subject: GNATS Login was Re: Linux Kernel, KDE & GCC 3.0.3
> Date: Fri, 14 Dec 2001 21:26:41 -0500
> Organization: Dome-S-Isle Data
> In-Reply-To: <20011214105513.A5977@mediaone.net>
> 
> Something that has been getting on my nerves is that it is impossible to (at 
> least thru a direct link) read a PR w/o having a GNATS login.
> 
> Any plans to make the GNATS db available for public access?
> 
> On Friday 14 December 2001 10:55, Craig Rodrigues wrote:
> > On Fri, Dec 14, 2001 at 02:10:32AM -0500, Adam Schrotenboer wrote:
> > > Have the KDE issues been resolved in 3.0.3? also, what of the Kernel 2.4
> > > 8139too.c ICE?
> >
> > The 8139too.c ICE should be fixed:
> > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4529&database=gcc
> >
> > I'm not sure about the GLIBC and KDE issues.
> 
> -- 
> tabris
> Courage is resistance of fear, mastery of fear, not absence of fear.
> 
> ----- End forwarded message -----

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor quoted someone:

> 
> Something that has been getting on my nerves is that it is impossible to (at 
> least thru a direct link) read a PR w/o having a GNATS login.
> 

I think this person's real complaint is that the gcc gnatsweb page is
not saying "log in as guest/guest to get guest access" - it probably got
dropped with Gerald's latest gnatsweb update.

I agree with the points you were making about how it should behave, but
I think this complaint is a more basic one; it appears that you need
an account to view the PRs.

J

Re: [adam@tabris.net: GNATS Login]

[Christopher Faylor]

On Fri, Dec 14, 2001 at 09:28:48PM -0800, Jason Molenda wrote:
>On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor quoted someone:
>> Something that has been getting on my nerves is that it is impossible to (at 
>> least thru a direct link) read a PR w/o having a GNATS login.
>> 
>
>I think this person's real complaint is that the gcc gnatsweb page is
>not saying "log in as guest/guest to get guest access" - it probably got
>dropped with Gerald's latest gnatsweb update.
>
>I agree with the points you were making about how it should behave, but
>I think this complaint is a more basic one; it appears that you need
>an account to view the PRs.

Ah.  You're probably right.  I just assumed that it would be obvious that
you should login as guest but my assumption is based on past experience
rather than any on-screen clues.

cgf

Re: [adam@tabris.net: GNATS Login]

[Christopher Faylor]

On Fri, Dec 14, 2001 at 09:28:48PM -0800, Jason Molenda wrote:
>On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor quoted someone:
>> Something that has been getting on my nerves is that it is impossible to (at 
>> least thru a direct link) read a PR w/o having a GNATS login.
>> 
>
>I think this person's real complaint is that the gcc gnatsweb page is
>not saying "log in as guest/guest to get guest access" - it probably got
>dropped with Gerald's latest gnatsweb update.
>
>I agree with the points you were making about how it should behave, but
>I think this complaint is a more basic one; it appears that you need
>an account to view the PRs.

Ah.  You're probably right.  I just assumed that it would be obvious that
you should login as guest but my assumption is based on past experience
rather than any on-screen clues.

cgf

Re: [adam@tabris.net: GNATS Login]

[Gerald Pfeifer]

On Fri, 14 Dec 2001, Jason Molenda wrote:
> I think this person's real complaint is that the gcc gnatsweb page is
> not saying "log in as guest/guest to get guest access" - it probably got
> dropped with Gerald's latest gnatsweb update.

It seems we lost this quite a bit earlier (probably as part of some
earlier gnatsweb merger). :(

I just committed this patch to at least prominently mention the guest
login.

Gerald

Index: gnatsweb.pl
===================================================================
RCS file: /cvs/gcc/wwwdocs/cgi-bin/gnatsweb.pl,v
retrieving revision 1.44
diff -u -3 -p -r1.44 gnatsweb.pl
--- gnatsweb.pl	2001/12/14 20:02:50	1.44
+++ gnatsweb.pl	2001/12/15 09:55:40
@@ -3340,6 +3340,10 @@ sub login_page

   # Print the login form.
   print $q->start_form(),
+  #GCC-LOCAL begin.
+        "<p>Use username `<em>guest</em>' and password `<em>guest</em>' ",
+        "for read-only and bug reporting access.</p>",
+  #GCC-LOCAL end.
         "<table>",
         "<tr>\n<td>User Name:</td>\n<td>",
         $q->textfield(-name=>'user',

Re: [adam@tabris.net: GNATS Login]

[Gerald Pfeifer]

On Fri, 14 Dec 2001, Jason Molenda wrote:
> I think this person's real complaint is that the gcc gnatsweb page is
> not saying "log in as guest/guest to get guest access" - it probably got
> dropped with Gerald's latest gnatsweb update.

It seems we lost this quite a bit earlier (probably as part of some
earlier gnatsweb merger). :(

I just committed this patch to at least prominently mention the guest
login.

Gerald

Index: gnatsweb.pl
===================================================================
RCS file: /cvs/gcc/wwwdocs/cgi-bin/gnatsweb.pl,v
retrieving revision 1.44
diff -u -3 -p -r1.44 gnatsweb.pl
--- gnatsweb.pl	2001/12/14 20:02:50	1.44
+++ gnatsweb.pl	2001/12/15 09:55:40
@@ -3340,6 +3340,10 @@ sub login_page

   # Print the login form.
   print $q->start_form(),
+  #GCC-LOCAL begin.
+        "<p>Use username `<em>guest</em>' and password `<em>guest</em>' ",
+        "for read-only and bug reporting access.</p>",
+  #GCC-LOCAL end.
         "<table>",
         "<tr>\n<td>User Name:</td>\n<td>",
         $q->textfield(-name=>'user',

Re: [adam@tabris.net: GNATS Login]

[Jason Molenda]

On Fri, Dec 14, 2001 at 10:31:28PM -0500, Christopher Faylor quoted someone:

> 
> Something that has been getting on my nerves is that it is impossible to (at 
> least thru a direct link) read a PR w/o having a GNATS login.
> 

I think this person's real complaint is that the gcc gnatsweb page is
not saying "log in as guest/guest to get guest access" - it probably got
dropped with Gerald's latest gnatsweb update.

I agree with the points you were making about how it should behave, but
I think this complaint is a more basic one; it appears that you need
an account to view the PRs.

J

[adam@tabris.net: GNATS Login]

[Christopher Faylor]

This has probably been discussed before, but is there any way to avoid
the problem described below?  It is annoying to have to login before
viewing a PR.

It would be nice if you were just logged into guest automatically.

cgf

----- Forwarded message from Adam Schrotenboer <adam@tabris.net> -----

From: Adam Schrotenboer <adam@tabris.net>
To: gcc@gcc.gnu.org
Subject: GNATS Login was Re: Linux Kernel, KDE & GCC 3.0.3
Date: Fri, 14 Dec 2001 21:26:41 -0500
Organization: Dome-S-Isle Data
In-Reply-To: <20011214105513.A5977@mediaone.net>

Something that has been getting on my nerves is that it is impossible to (at 
least thru a direct link) read a PR w/o having a GNATS login.

Any plans to make the GNATS db available for public access?

On Friday 14 December 2001 10:55, Craig Rodrigues wrote:
> On Fri, Dec 14, 2001 at 02:10:32AM -0500, Adam Schrotenboer wrote:
> > Have the KDE issues been resolved in 3.0.3? also, what of the Kernel 2.4
> > 8139too.c ICE?
>
> The 8139too.c ICE should be fixed:
> http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4529&database=gcc
>
> I'm not sure about the GLIBC and KDE issues.

-- 
tabris
Courage is resistance of fear, mastery of fear, not absence of fear.

----- End forwarded message -----