Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread Adam Di Carlo 


I suppose I wasn't clear.

I honestly can't remember if the created empty ld.so.conf thing gave
me a hit return to continue type prompt or not.  I'll pay more
attention next time.

My general contention is that such messages are too low-level and
there's no reason to present such things to the user.  It will make a
luser think that something is wrong, because an empty file had to be
made, when in fact, it's normal.  Either we could suppress the message
or make it more clear it's just SOP (standard operating procedure).

The screen is looking a lot better -- I wonder if we could do it as a
scrolling screen rather than just one message at a time...

-- 
.Adam Di [EMAIL PROTECTED]URL:http://www.onshored.com/


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread David Whedon 

Thu, May 24, 2001 at 03:16:36AM -0400 wrote:
 
 On install base (debootstrap) step, I am getting a hang while it
 tells me to acknowledge that it created an empty ld.so.conf and other
 things.  I would prefer if we not make the user have to hit return for
 that...
 
 -- 

That's by design, but I'm happy to change it.  debootstrap considers these
messages to be warnings, so we were displaying them.  Now those messages will be
displayed just like the informational messages.  I considered making the choice
of behavior conditional on bootargs.quiet.  Another option would be to change
what debootstrap thinks is a warning.  If either of those sound better, speak
up.

David


I just comitted the following:

diff -u -r1.34 extract_base.c
--- extract_base.c  2001/05/15 06:47:13 1.34
+++ extract_base.c  2001/05/24 15:08:26
@@ -128,15 +128,10 @@
if (ptr[1] == ':')
{
switch (ptr[0]){
+   case 'W':
case 'I':
{
pleaseWaitBox(ptr+3);
-   break;
-   }
-   case 'W':
-   {
-   problemBox(ptr+3, _(Warning));
-   pleaseWaitBox(_(Continuing));
break;
}
case 'E':


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread Adam Di Carlo 

David Whedon [EMAIL PROTECTED] writes:

 Thu, May 24, 2001 at 03:16:36AM -0400 wrote:
  
  On install base (debootstrap) step, I am getting a hang while it
  tells me to acknowledge that it created an empty ld.so.conf and other
  things.  I would prefer if we not make the user have to hit return for
  that...

 That's by design, but I'm happy to change it.  debootstrap considers these
 messages to be warnings, so we were displaying them. 

Ok -- it's a debootstrap bug then, or else debootstrap is complaining
about packages in base which are not as they should be?

 Now those messages will be
 displayed just like the informational messages. 

I'm not sure if that's really the right choice -- shouldn't we get it
fixed in debootstrap instead?

 I considered making the choice
 of behavior conditional on bootargs.quiet.  Another option would be to change
 what debootstrap thinks is a warning.  If either of those sound better, speak
 up.

Fix debootstrap -- we do need some way of stopping for user notice --
e.g., disk full.


-- 
.Adam Di [EMAIL PROTECTED]URL:http://www.onshored.com/


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread Richard Hirst 

On Thu, May 24, 2001 at 08:14:49AM -0700, David Whedon wrote:
 Thu, May 24, 2001 at 03:16:36AM -0400 wrote:
  
  On install base (debootstrap) step, I am getting a hang while it
  tells me to acknowledge that it created an empty ld.so.conf and other
  things.  I would prefer if we not make the user have to hit return for
  that...
  
  -- 
 
 That's by design, but I'm happy to change it.  debootstrap considers these
 messages to be warnings, so we were displaying them.  Now those messages will be
 displayed just like the informational messages.  I considered making the choice
 of behavior conditional on bootargs.quiet.  Another option would be to change
 what debootstrap thinks is a warning.  If either of those sound better, speak
 up.

I'd say fixing debootstrap to downgrade certain things from Warning
to Info is the right thing to do, rather than disable the Warning
feature entirely.

A worse problem, IMHO, is that the warning box gets scrolled off the
screen by all the Unpacking ..., Setting up messages that
scroll up my screen staircase fashion.  Maybe that is a feature of
a serial console install, and output goes somewhere else on a
graphical install.

Richard

 David
 
 
 I just comitted the following:
 
 diff -u -r1.34 extract_base.c
 --- extract_base.c  2001/05/15 06:47:13 1.34
 +++ extract_base.c  2001/05/24 15:08:26
 @@ -128,15 +128,10 @@
 if (ptr[1] == ':')
 {
 switch (ptr[0]){
 +   case 'W':
 case 'I':
 {
 pleaseWaitBox(ptr+3);
 -   break;
 -   }
 -   case 'W':
 -   {
 -   problemBox(ptr+3, _(Warning));
 -   pleaseWaitBox(_(Continuing));
 break;
 }
 case 'E':
 
 
 -- 
 To UNSUBSCRIBE, email to [EMAIL PROTECTED]
 with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
 


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread Matt Kraai 

On Thu, May 24, 2001 at 04:33:34PM +0100, Richard Hirst wrote:
 A worse problem, IMHO, is that the warning box gets scrolled off the
 screen by all the Unpacking ..., Setting up messages that
 scroll up my screen staircase fashion.  Maybe that is a feature of
 a serial console install, and output goes somewhere else on a
 graphical install.

Could you file a bug against the boot-floppies about this.  This
sounds pretty gross.

Does anyone know what happens with the stdout/stderr messages
from other programs when installing via a serial console?

Matt


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrap/b-f interaction -- quieten

2001-05-24 Thread David Whedon 

Hi Anthony,

I don't know if you've seen this thread, it started here:
http://lists.debian.org/debian-boot-0105/msg00719.html

I figure you can provide some input.

Thu, May 24, 2001 at 11:35:42AM -0400 wrote:
 David Whedon [EMAIL PROTECTED] writes:
 
  Thu, May 24, 2001 at 03:16:36AM -0400 wrote:
   
   On install base (debootstrap) step, I am getting a hang while it
   tells me to acknowledge that it created an empty ld.so.conf and other
   things.  I would prefer if we not make the user have to hit return for
   that...
 
  That's by design, but I'm happy to change it.  debootstrap considers these
  messages to be warnings, so we were displaying them. 
 
 Ok -- it's a debootstrap bug then, or 
I don't think it is a debootstrap bug.

 else debootstrap is complaining
 about packages in base which are not as they should be?
Anthony? care to comment?

 
  Now those messages will be
  displayed just like the informational messages. 
 
 I'm not sure if that's really the right choice -- shouldn't we get it
 fixed in debootstrap instead?


It really depends on how we are going to define 'warning'.  As it was ajt had
decided that some messages were warnings but it looked like boot-floppies didn't
want to see them.  It sounds as though debootstrap and boot-floppies have to
agree on what constitutes a warning.

 
  I considered making the choice
  of behavior conditional on bootargs.quiet.  Another option would be to change
  what debootstrap thinks is a warning.  If either of those sound better, speak
  up.
 
 Fix debootstrap -- we do need some way of stopping for user notice --
 e.g., disk full.

debootstrap can still send errors, when we get them we bail.  


David


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]