lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev lynx 2.8.2dev 15 trace bug ?


From: Klaus Weide
Subject: Re: lynx-dev lynx 2.8.2dev 15 trace bug ?
Date: Fri, 5 Feb 1999 03:17:53 -0600 (CST)

On Fri, 5 Feb 1999, Larry W. Virden wrote:

> I ran
> 
> lynx -trace URL
> 
> where the URL was some arbitrary site.  When I did a Q to exit lynx, I 
> got:
> delete anchor:55/1,1,0 
> http://www.watercolor.com/projectsites/ImportCD/website/icd_html/icd_body_home.cfm
> 
> written to the screen.

Yeah, it's some trace output I recently added, it is mentioned in
CHANGES.

> I think that somewhere there is a trace line going to stdout or stderr
> instead of to the trace file.

It uses the same CTRACE(tfp, ...) as is used everywhere else, but
apparently it happens so late in exiting that the trace log file has
already been closed.  That wasn't intended, but I decided to regard it
as a feature not a bug... (you may disagree.)

> Another peculiar thing that I saw was that before using trace, lynx was
> reporting Bad HTML ... turn on trace to get more info.  When I turned on
> trace, I got a 9000+ line file - without a line saying "error".  I did
> find a "Bad HTML" just now as I was typing.  It would be nice if warnings and
> errors that are going to be reported to the user of lynx might be tagged
> with a special string indicating in the trace where they are...  Would
> there be 1 or 2 places in the code where I could determine that ? I would
> hate to have to change many tracing lines...

Most should go through HTAlert(), HTInfoMsg(), HTUserMsg(), HTUserMsg2().
Those functions use some string prefix, see HTAlert.c.

   Klaus

reply via email to

[Prev in Thread] Current Thread [Next in Thread]