Re: LYNX-DEV _Lots_ more on INSTALLATION and SIMPLEDIR

[Bill Schiavo]

Hello everyone.  It's me again.  Are you curious what ***was*** is??  I
should have written it as ***W. A. S. *** Those are my initials.  Okay now
on to my reason for writing. 


On Wed, 8 Oct 1997, Nelson Henry Eric wrote:

> > > You only need to stubedit cc1.exe.  Cc1plus.exe is the c++ compiler.
> > > Lynx doesn't use c++, as far as I know.  You're right Doug,  Sorry about
> 
> "http://163.51.110.11/lynxdev/INSTALLATION.gz"; was changed slightly in
> this respect.
> 
> > Remember that the batch file I supplied was only a sample.  I have cleaned 
> > it
> 
> Unless other members of the list express a real desire for it, I personally
> think having it archived on lynx-dev is more appropriate than putting it in
> with the distribution documents.  It looked good at the time, and it looks
> even better now.  However, I tend to agree with Doug that anyone who could
> compile Lynx from source code on 386DOS would have the capability to write
> such a script, i.e., it's not core to the installation process.
>
Many of us can write simple batch files but why should we when it has
already been done. 

> > I have included the above changes.  Also, If we are going to include a 
> > makefil
> > to build the chartrans code, it isn't necessary to build the makeuctb 
> > program 
> > the top-level makefile (makefile.dos).  As I said, that program along with 
> > the
> > translation of the chartrans tables can be done stand alone.
> 
> Hmmm.  Things really could use some polishing up here.  Since the top-level
> makefile is already in place, wouldn't it be simpler to just remove the
> makeuctb target from your makefile proposed for src/chrtrans?  Or would it
> be better to go the other way and put everything in the top-level makefile?
>
I can see that you don't really understand the process at all.  The
chartrans code must be compiled before ucdomap.c since that module uses
the header files that contain the character translation maps.  Without
these header files, ucdomap.c will bomb.  If the makeuctb dependency is
removed from the top level makefile and if you have a separate makefile to
do the chartrans code, you'll keep things much simpler.  
> > Henry, for you to infer that my makefile doesn't work isn't called for here.
> > You are certainly invited to write one of your own.  The one I provided 
> > works
> > very well on my system and isn't meant to be the final version.
>                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> Well, I did get the reaction I wanted, which is commitment.  I asked you
> explicitly to NOT write to me privately, and TO keep the discussion public
> on lynx-dev.  The very reason for this is because I am not qualified in any
> way to judge whether or not a makefile might work, let alone write one.
It's a good thing you admitted it because I would have told you so if you
didn't.
> 
> A new release of Lynx is imminent.  It would be nice to have a "final version"
> in place before that happens.
To that end, I have provided one that I feel will do the job quite well.
It is below:

#
# Makefile for the makeuctb and unicode tables
# for use with DJGPP.
#
# Type make to build makeuctb  and all character translation maps.
# Type make fontmap to build makeuctb and translation map iso8859-1.
# Type make makeuctb.exe to build makeuctb only.
# Type make clean to remove makeuctb and character translation maps.
# Type make distclean to remove makeuctb, character translation maps
# and .bak files.
#
CFLAGS = $(MCFLAGS)

CC = gcc
MCFLAGS = -O3 -DEXP_CHARTRANS -DDOSPATH -DNO_TTYTYP \
-I../../WWW/library/implement -I../../djgpp/tcplib/include \
-I../../djgpp/tcplib/include/tcp

.SUFFIXES: .tbl
#
# This file contains the font map for the default (hardware) font
#

FONTMAP_INC = iso01_un.h

CHRTR=

TABLES= $(CHRTR)iso01_un.h \
 $(CHRTR)iso02_un.h \
 $(CHRTR)def7_uni.h \
 $(CHRTR)iso03_un.h \
 $(CHRTR)iso04_un.h \
 $(CHRTR)iso05_un.h \
 $(CHRTR)iso06_un.h \
 $(CHRTR)iso07_un.h \
 $(CHRTR)iso08_un.h \
 $(CHRTR)iso09_un.h \
 $(CHRTR)iso10_un.h \
 $(CHRTR)koi8r_un.h \
 $(CHRTR)cp437_un.h \
 $(CHRTR)cp850_un.h \
 $(CHRTR)cp852_un.h \
 $(CHRTR)cp866_un.h \
 $(CHRTR)cp1250_u.h \
 $(CHRTR)cp1251_u.h \
 $(CHRTR)cp1252_u.h \
 $(CHRTR)viscii_u.h \
 $(CHRTR)utf8_uni.h \
 $(CHRTR)rfc_suni.h \
 $(CHRTR)mnemonic.h \
 $(CHRTR)mnem_sun.h

default: $(TABLES)

fontmap: $(FONTMAP_INC)

makeuctb.exe: makeuctb.c UCkd.h
        $(CC) $(CFLAGS) -o makeuctb.exe makeuctb.c
        strip makeuctb.exe

.tbl.h:
        ./makeuctb $*.tbl > $@

iso01_un.h: iso01_un.tbl makeuctb.exe
iso02_un.h: iso02_un.tbl makeuctb.exe
def7_uni.h: def7_uni.tbl makeuctb.exe
iso03_un.h: iso03_un.tbl makeuctb.exe
iso04_un.h: iso04_un.tbl makeuctb.exe
iso05_un.h: iso05_un.tbl makeuctb.exe
iso06_un.h: iso06_un.tbl makeuctb.exe
iso07_un.h: iso07_un.tbl makeuctb.exe
iso08_un.h: iso08_un.tbl makeuctb.exe
iso09_un.h: iso09_un.tbl makeuctb.exe
iso10_un.h: iso10_un.tbl makeuctb.exe
koi8r_un.h: koi8r_un.tbl makeuctb.exe
cp437_un.h: cp437_un.tbl makeuctb.exe
cp850_un.h: cp850_un.tbl makeuctb.exe
cp852_un.h: cp852_un.tbl makeuctb.exe
cp1250_u.h: cp1250_u.tbl makeuctb.exe
cp1251_u.h: cp1251_u.tbl makeuctb.exe
cp1252_u.h: cp1252_u.tbl makeuctb.exe
utf8_uni.h: utf8_uni.tbl makeuctb.exe
mnemonic.h: mnemonic.tbl makeuctb.exe
mnem_sun.h: mnem_sun.tbl makeuctb.exe
rfc_suni.h: rfc_suni.tbl makeuctb.exe

clean:
        rm -f makeuctb.exe makeuctb *.o *un.h *u.h *c.h *i.h

distclean: clean
        -rm -f *.bak

> 
> You and anyone else are more than welcome to look over your makefile proposal
> and the editing changes I made to it.  Make suggestions, criticize, bash away!
That also goes for my new and finalized version of my proposed makefile.
> As I said yesterday, you can find it, in its entirety, as a diff -c in
> "http://163.51.110.11/lynxdev/SIMPLEDIR.gz";.  (Grep for your name.)  The
> reason for the diff -c format is because I thought this might be easier for
> Klaus who has so much other stuff to do.
> 
> And, BTW, before anyone gets too hot, I hope you all know that there is no
> guarantee at all that any of this will get into the main-stream distribution.
> 
> __Henry
> ;
> ; To UNSUBSCRIBE:  Send a mail message to address@hidden
> ;                  with "unsubscribe lynx-dev" (without the
> ;                  quotation marks) on a line by itself.
> ;
> 

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;