Re: [ethereal-dev] static vs dynamic column width

[Guy Harris <guy@xxxxxxxxxx>]

(I don't know who's on one of "ethereal-users" or "ethereal-dev" but not
on both, so I'm CCing both lists for this....)

> I was hoping that the proposal of going back to static width columns
> would die out after the messages reporting problems with the dynamic
> column code and the corrections were posted.  However, it didn't, and
> we've now reverted to the static column width. 

Yup.  Resizing columns dynamically doesn't work very well with "-S"; at
least some people, it appears, do *NOT* want the sizes changing while
the capture is going on - Guy Swarc said, in the "ethereal-users"
thread:

> I still don't like the idea of resizing columns each time new records
> are read with -S.  If I am trying to examine some of the packets
> captured so far without stopping the capture process, I don't want the
> screen flickering and changing on me.  That makes it very hard to
> examine the data.

and Laurent Deniel agreed.

> I'd like to propose that we keep the ability to size the column based
> on the longest string within that column.  This is VERY useful for long
> "Info" strings - what I had been doing before was just setting the
> display width to the maximum size of the character string (which is MUCH
> larger than the other columns (in my modified version, which nobody has
> expressed an interest in). It's very nice to have the horizontal scroll
> bar reflect the actual width of the window based on the data in the
> columns.

Yes, it's a pain that the fields other than "Info" are wider than they
need be, so that "Info" is narrower than it could be.

Unfortunately, it appears it's also a pain to have the fields changing
out from under somebody running a capture with "-S".

Laurent Deniel suggested, in the "ethereal-users" thread:

> Resizing on user actions is ok but auto-resizing while capturing with -S
> option without any user actions is a bit confusing. The only moment where 
> it might be allowed to auto-resize is at the end of the capture session
> (when the user presses the stop button).

Would it be acceptable to all concerned to have the current static
widths be in effect as long as a "-S" capture is in progress, with the
columns resized as appropriate when the capture is stopped?

He also said:

> And auto-rezing while reading 
> a captured file is not necessary (but the appropriate widths may be saved 
> in order to resize at the end of file).

I assume by "auto-resizing while reading a captured file" he means
"changing the column sizes as each packet is read in" - I don't *think*
it was doing that when it auto-sized the columns; was it?