Re: lynx-dev Building 2 copies of Lynx: EXP_NESTED_TABLES

[Thomas Dickey]

On Fri, Mar 02, 2001 at 12:59:36AM -0300, Frédéric L. W. Meunier wrote:
> Hi. I'm building 2 copies of Lynx and would like to know if I'm
> doing anything wrong:
> 
> First copy is built without --enable-nested-tables.
> 
> To build the second copy I modify objdir/lynx_cfg.h and change
> 
> /* #undef EXP_NESTED_TABLES */  /*
> CF_ARG_ENABLE(nested-tables) */
> 
> to
> 
> #define EXP_NESTED_TABLES 1 /* CF_ARG_ENABLE(nested-tables) */
> 
> Then I rm objdir/src/{GridText,HTML,TRSTable}.o and do a make.
> I noticed LYMain.o LYShowInfo.o are also rebuilt.
> 
> Is this enough ? I just want to save time and resources.

the userdefs.h dependency is ad hoc - is not necessarily complete since
that file is included from HTUtils.h, but was set up in makefile.in to
reflect the most obvious dependencies.  (I would simply do a 'make clean'
first anyway, though there is a 'make depend' rule which also is useful).
 
> BTW, I reported the warnings you get from string.h (not only
> building Lynx) to the libc-alpha mailing-list, and Andreas
> Jaeger answered that gcc 3.0 will fix them.
> 
> But what about the flood of warnings you get building with
> --with-ssl ?

I filter that out by putting a 'void' in the headers where there are
empty parameter lists.  (That isn't necessarily 100% correct, but it
suffices for what I need).
 
> warning: function declaration isn't a prototype
> 
> Should OpenSSL be fixed, or it's also gcc ?

OpenSSL.  gcc is correct (at least in this detail).
At least OpenSSL will build with an ANSI C compiler.
OpenSSH does not.

-- 
Thomas E. Dickey <address@hidden>
http://dickey.his.com
ftp://dickey.his.com

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden