Re: LYNX-DEV CHANGES 2.8.1dev.5

[David Combs]

> From address@hidden Sun Mar 29 19:33:47 1998
> Date: Sun, 29 Mar 1998 22:02:20 -0500 (EST)
> From: Wayne Buttles <address@hidden>
> To: address@hidden
> Subject: Re: LYNX-DEV CHANGES 2.8.1dev.5
> In-Reply-To: <address@hidden>
> MIME-Version: 1.0
> Content-Type: TEXT/PLAIN; charset=US-ASCII
> Sender: address@hidden
> Precedence: bulk
> Reply-To: address@hidden
> 
> 
> On Sun, 29 Mar 1998, T.E.Dickey wrote:
> 
> > > > > The log file stuff doesn't work for BORLANDC's compile :-( 
> > The real fix is to replace all of those stderr references with the actual
> > logfile pointer (not something that you'd want to do along with some other
> > changes, since there's a thousand or so to change).
> 
> Is it just a search and replace?  Textpad for MS Windows can do a easy
> search and replace for a string in all documents within a directory.  Yea,
> I know there are a dozen ways to do it in Unix, but this is a quick and
> easy way for us lower minions. ;-)
> 
> 
> 

Do it this way: egrep -w stderr everywhere > t.loc

Then, go throught that t.loc, and for each one
you DO NOT want to change, change the source
stderr to stdbananas or something else unusual.

Then, in one fell swoop, change ALL the rest.

Then, in one fell swoop, change all the stdbananas 
back to stderr.

emacs makes it even easier -- that is if you do it
one file at a time, via mx-occur, for then you
can EASILY change the stderrs you want to stdbananas,
via a simple ^C^C in the *Occur* buffer.

Both schemes above allow a delay between deciding
on what to change to stdbananas and the desired
change of stderr to something else, eg outfile.

Or you COULD be bolder, in emacs, and merely answer
y or n for EACH change, saying n for those you don't
want to change, leaving out the stdbananas passes.

---

For those who don't use these tools, I hope this
gets some at least to investigate them.

By the way, that -w in the egrep works only for
GNU egrep.  Another feature (very nice) that that egrep
has is -A<number of lines AFTER the found pattern, to print>
and -B<ditto, but BEFORE> -- and it shows all the overlapping
regions correctly and easily to see, just like you'd want.
That -A and -B gives you some CONTEXT for what you are egrepping
for.

Unix tools are clearly the way to go, as unix was designed
BY and FOR programmers who do programming 10 hours a day,
whereas MS-stuff was done for "users" most of whom NEVER
program anything, and mostly have never even seen an ascii
file, at least not AS one.  Very different set of tools
for a very different purpose.