bug#38796: 26.3; `view-lossage': Use a variable for the lossage limit

[Drew Adams]

> I thought it sounded quite natural to use the same mechanism to switch
> lossage logging off.  `(update-lossage-size 0)' seems like a natural
> thing to do to have no lossage.

FWIW, I said the same thing:

https://debbugs.gnu.org/cgi/bugreport.cgi?bug=38796#28

It's not unusual or weird for a zero-valued size
parameter to have such an effect. 

> >     Users can still safely modify the lossage limit with the
> >     command `update-lossage-size'.  For convenience, add
> >     a function `lossage-size' to return the current limit.
> 
> Hm...  other functions use the convention of having a zero-parameter
> function return the data, and having a parameter to set it.  So
> (lossage-size 500) could set it and (lossage-size) could return the
> current size?

FWIW, most `set-*' functions are commands.
And a command is what we should shoot for here.