lynx-dev Re: "new anchor" message

[Clair Wilson]

Klaus

1. This has happened on 2 releases of lynx :  release 2.5, and release 2.8

   I compiled release 2.5 in July of 1996, and it ran successfully ever
   since. I compiled lynx 2.8 a few weeks ago after this problem started.
   I tried to fix the problem by downloading and installing what appeared
   to be the latest release of lynx.

   The problem started after I installed fixes to AIX 3.2.5 in order to
make the operating system y2k compliant. All the fixes seemed to work. AIX
is now y2k compliant. However, the only thing which seems to have been
adversely affected by the y2k fixes is lynx. Everything else works
perfectly, including all other web access.

2. Since I did not yet apply the y2k fixes to the other AIX 3.2.5 system
   on the other rs600, lynx is still working on that system.

3. I have upgraded both system from 2.5 to 2.8. The level appears to have
nothing to do with this problem, since lynx works on both where the y2k
fix was not applied, and lynx fails on both where the y2k fix WAS applied.

4. I looked for both strace and truss, but found neither on my system. 
   Are these UNIX traces or LYNX traces ?

5. I DO have a system call trace utility, however. It is the aix trace
   command.
   I will run it and let you know what the results are.

Thank You
 
Mr. Clair Wilson
Internet Administrator
Mainframe System Programmer
Information Services
Robert Packer Hospital
Sayre, Pa.

address@hidden

phone 717-882-5389
fax   717-882-5330

On Tue, 2 Feb 1999, Klaus Weide wrote:

> On Tue, 2 Feb 1999, Clair Wilson wrote:
> 
> > Does "new anchor" have anything to do with the problem ?
> > 
> > We have one rs6000 upon which lynx is functional. I traced lynx there in
> > order to look for trace differences.
> 
> Are you really comparing the same versions?
> 
> > I don't see the "new anchor" message coming out on the lynx which is
> > functioning properly. However, it comes out every time on the lynx which
> > is failing. So here is the "new anchor" message.
> 
> If one is an old version, the trace output may have changed.
> I don't think this is significant.
> 
> Actually the trace *should* say "New anchor ..." if this is on
> startup, as it does.  There's nothing wrong with that.
> The later, different status number you pointed to is just a
> result of something failing earlier, it has no direct relation
> to sytem error numbers.
> 
> I don't really trust that your method of capturing trace output
> is reliable.  The 2.8 should direct trace output by default int
> a file Lynx.trace, that should be more reliable than cut-and-pase
> from the screen.  (If you don't get the file, try the -tlog flag).
> 
> We are getting further away from where the problem most likely
> lies.  Thanks for the traces, but unfortunately they only show
> that they don't show enough.  Lynx probably isn't really generating
> enough trace output to pinpoint the problem.
> 
> I asked you a number of question in another mail which you didn't answer -
> please do so.  Most important - Did you compile Lynx yourself, can you
> recompile it?  Also, do you have a system call trace utility, like strace
> or truss on other Unixes?
> 
>    Klaus
> 
>