Re: Lout newbie

[P.J. Kersten]

Valeriy E. Ushakov wrote:
> 
> On Fri, Nov 28, 1997 at 10:35:42AM +0100, P.J. Kersten wrote:
> 
> > I compiled Lout with Gnu-win32's gcc.
> 
> Can you please drop me a note about your experience with compiling
> Lout with gnuwin.  I have gnuwin installed but don't have enough time
> to try it out.  I don't want to make Win32 Lout distribution dependent
> on cygwin.dll, but people that use gnuwin might prefer to compile Lout
> for themselves using gcc.
> 
> PS: As my VC++ 5.0 CD finally made its way through customs, I'll
>     probably compile 3.11 and try to make an installation wizard with
>     InatallShield 5 Free Edition that comes with VC 5.0 Professional.
>     Expect it by the end of next week (well, I'll try).
> 
> SY, Uwe
> --
> address@hidden                         |       Zu Grunde kommen
> http://www.ptc.spbu.ru/~uwe/            |       Ist zu Grunde gehen

Essentially there is nothing to it; it compiles out-of-the-box.
The only thing to look after is to set USELOC=0. Both OSUNIX=1 and
OSDOS=1 will compile. I choose OSUNIX=1.
Be prepared to ignore a _lot_ warnings about signed/unsigned comparisons
or compile with ASSERT_ON=0. There are a few freaking warnings though;
'expression is always one due to the limited range of the datatype' or
something similar in the lexical analyser. It backtraces to zz_length[]
as far I could see in my limited time. However, I could find no strange
behaviour (the manuals lout'ed just fine).
The makefile needs a little patch to (un)install the *.exe's right.
This is straightforward though. I recall there is a scripted version 
for gnu-win32's install.exe which can manage '.exe's, but the makefile
uses 'cp' and 'chmod' quite extensively, so a patch is still needed.

Hope this helps.
SY,
-- 
Pieter J. Kersten
Erasmus Universiteit Rotterdam
dir. I&A/I&S, WH2-10
Postbus 1738
3000 DR  Rotterdam
Nederland
Tel +31-10-4082599
Fax +31-10-4527236