lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: LYNX-DEV BSD makefile support 4


From: Matthew Kelly
Subject: Re: LYNX-DEV BSD makefile support 4
Date: Thu, 15 May 1997 13:18:16 -0400 (EDT)

On Thu, 15 May 1997, Michael Sokolov wrote:
>    Roger Hill <address@hidden> writes:
> >Isn't this discussion pointless?

Thats my question!  I have to admit I haven't yet tried using the
autoconfigure process ... But, the bottom line is, does it work?  Does it
produce a proper executable for your OS?  Why are we worried about how it
determines settings that work to compile lynx?

To satiate Michael's quest for purity, would it be easy to have an option
to the autoconfigure script which just produces the makefile instead of
making it also (I'm assuming thats how it works, but you get the idea),
which then he could take and vi with whatever west coast BSD pure
equivalent of the probably tainted vi the rest of us use, and purify it of
all makefile impurities.  Sorry if I'm short, but I really don't think
this whole thing is to building lynx as "+ +" is to system security :)

Matt

>    As I have said in previous messages, I didn't mean to start an East
> Coast vs. West Coast flame war, I just want to help people.
>    I see you saying that the Makefile system is deprecated and you urge
> everyone to move to the auto-configure script. However, not everyone would
> like that. A UNIX guru who knows every line in his/her OS's source code
> would certainly prefer a Makefile that he/she can review and/or edit,
> rather than a mysterious 160 KB (!) shell script. A Makefile is completely
> predictable: one can see what files are compiled and what options are used,
> and one can change that if necessary. With an autoconfigure monster script,
> one runs it and has no idea what that alien ship is going to do.
>    You can't just throw out the traditional Makefile, since many people
> depend on it. The only way I can compile Lynx is to use the target I wrote
> a couple of days ago, and the whole point of my argument is to make it
> available to others. Of course, I could post it on my homepage, but that's
> counterproductive: imagine what would happen if everyone starts doing that.
> Such publications should be centralized.
>    When I loaded that 160 KB monster in an editor and tried to decipher it,
> I did notice references to BSDI, FreeBSD, NetBSD, and OpenBSD. But these
> are just special cases! The general case, i.e., the pure 4.4BSD tape from
> UC Berkeley, is missing. You may think that no one uses it, but there is an
> example to the contrary right in front of you: I'm going to install pure
> BSD in a couple of weeks.
>    Think of it this way. The target I have written is specifically
> optimized for pure 4.4BSD and BSDI. It has been tested on these systems and
> is known to work beautifully. For users of these systems, it will give
> optimum results, since it's specifically optimized and tested. The old BSDI
> target, on the other hand, didn't work in the optimum fashion. Besides
> SVR4, it defined NO_FILIO_H and NO_UTMP, even though these features ARE
> available on all BSD systems. This means that the existing target compiles
> an executable that is not optimized for its target system and doesn't use
> all of its features. In particular, needlessly defining NO_UTMP would make
> the inside-domain and outside-domain divisions unavailable when they could
> be available with correct compilation options.
>    As for the developmental code, I did get the latest one
> (lynx2.7.1ac-0.23.zip), but it's no different from the previous ones. Yes,
> the autoconfigure monster has been added, but the traditional Makefile is
> still broken.
>    Why can't you just add my target to the Makefile? It does the same thing
> as the existing one, except that it sets the compilation options correctly.
> If you don't want to update the code that you consider deprecated, why
> can't you assign this task to someone who doesn't think that way (e.g. me)?
>    
>    Sincerely,
>    Michael Sokolov
>    Phone: 216-646-1864
>    ARPA Internet SMTP mail: address@hidden
> ;
> ; To UNSUBSCRIBE:  Send a mail message to address@hidden
> ;                  with "unsubscribe lynx-dev" (without the
> ;                  quotation marks) on a line by itself.
> ;
> 

-------------------------------------------------------------------------
Matthew Kelly
address@hidden

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

reply via email to

[Prev in Thread] Current Thread [Next in Thread]