Re: lynx-dev Update for djgpp + _fmode

[Gisle Vanem]

address@hidden said:

> > You mean ";". Yes, Borland warn about that, but can be switched off.
> is that a separate error, or part of the error-level?  (I'd rather stick
> with the empty semicolon -- or else we can add a magic symbol such as
> LY_EMPTY_STATEMENT to cover both cases...)

To correct my earlier statement; Borland only warns if an empty macro is
used together with e.g. an if-statement. Like:

  main()
  {
    if (a == b)
       SetFmode();
  }

Gives warning `Code has no effect in function main' for the if-statement.
The Borland optimizer won't strip the if-statement though, but gcc 2.7.2
does (even with -O1).

The warning can be switched off with `#pragma warn -eff' or cmd-line
switch `-w-eff'.

Defining SetFmode() as `((void)0)' doesn't produce any warning.

Gisle V.