Re: [9fans] PPP screwup

Re: [9fans] PPP screwup

[presotto]

I put a new update containing our current /sys/src/9/port/devns16552.c.
Someone (dan?) please try it and make sure it still compiles and works
on distributed systems and fixes the PPP problem.

Re: [9fans] PPP screwup

[Dan Cross]

In article <20010910124555.DD9E019A47@mail.cse.psu.edu> you write:
>I put a new update containing our current /sys/src/9/port/devns16552.c.

Thanks, Dave.

>Someone (dan?) please try it and make sure it still compiles and works
>on distributed systems and fixes the PPP problem.

I snagged it and applied it to my system (the guys left for the plane
to the demo about an hour ago; yikes!).  It seems to be working fine;
I backed out my change to ip/ppp and we tested it for about an hour
or two with no failures before they took off.  I guess I'll know better
either way after the demo.  :-)

	- Dan C.

Re: [9fans] PPP screwup

[Martin Harriss]

presotto@plan9.bell-labs.com wrote:
>
> I put a new update containing our current /sys/src/9/port/devns16552.c.
> Someone (dan?) please try it and make sure it still compiles and works
> on distributed systems and fixes the PPP problem.

For the record, the text of the link to this update on the distribution
updates page is labeled "[June 5]" (rather than "[September 10th]" which
I presume would be correct.)

Martin

Re: [9fans] PPP screwup

[presotto]

thanks

Re: [9fans] PPP screwup

[Dan Cross]

In article <20010910220806.8C22B199D5@mail.cse.psu.edu> you write:
>Thanks much.  I'm terribly sorry about that.  I should have thought of it
>when you first complained since the same thing was happening to me in
>Maine earlier this summer.

That's quite all right; thanks for putting out the fix so quickly!

	- Dan C.

Re: [9fans] PPP screwup

[presotto]

Thanks much.  I'm terribly sorry about that.  I should have thought of it
when you first complained since the same thing was happening to me in
Maine earlier this summer.

[9fans] PPP screwup

[presotto]

Sorry about that.  I thought that fix had gotten out.  The problem is
the serial driver.  It had a bug that caused it sometimes to return 0 length on
reads even when ther was no EOF.  It was a stupid check.

The fix was small, but we've since rewritten the driver a few times.  I'll
put the current version out there on the update page.