Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 6 May 2000 15:55:42 -0400
From:      Will Andrews <andrews@technologist.com>
To:        Satoshi - Ports Wraith - Asami <asami@FreeBSD.ORG>
Cc:        Kris Kennaway <kris@FreeBSD.ORG>, ports@FreeBSD.ORG
Subject:   Re: ports projects
Message-ID:  <20000506155542.J55274@argon.blackdawn.com>
In-Reply-To: <vqck8hb97nm.fsf@silvia.hip.berkeley.edu>; from asami@FreeBSD.ORG on Wed, May 03, 2000 at 04:48:13AM -0700
References:  <Pine.BSF.4.21.0005021629410.44965-100000@freefall.freebsd.org> <vqck8hb97nm.fsf@silvia.hip.berkeley.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 03, 2000 at 04:48:13AM -0700, Satoshi - Ports Wraith - Asami wrote:
> Yes, that will probably work, but it's also a lot of work to do!

Users can opt to regenerate this database at night and specify a regex of
ports (i.e. only these categories and/or only these ports under those
categories, and so forth). Such a database may save a considerable amount
of time downloading files.

But then there's the problem of laptop computers.  Mine hooks up to a
FE->2xT1 school connection AND a FE->26400bps connection at home.

In general, I don't think this idea is really (universally) feasible,
although my laptop is probably one of the rarer cases of bandwidth
differentials. I've made a conclusion, and have a better idea.

My proposal (rotating MASTER_SITES) would be a better general solution. Of
course, I also think that MASTER_SITE_N (where N = distfile N) would be
necessary in order to implement this. Also, if we were to implement
something like that, a MASTER_SITE_CVS seems like a good idea too (since
the ports that use cvs in our tree are notorious for having large numbers
of distfiles). Of course, we would only need MASTER_SITE_N when a single
set of MASTER_SITES won't work.

How does this proposal sound ? I think the code for writing this would be
fairly simple (and quite logical IMO) and therefore the investment of time
in writing it would be much less of a waste than the time people save on
download times.

Although I do think that in some cases, some MASTER_SITES should only be
used purely for backup purposes; as such, a MASTER_SITE_N_BACKUP or
MASTER_SITES_BACKUP variable can be used to specify sites that should be
used only if forced to due to primary sites being out or something similar.
For example, let's say that a certain port has 2 primary sites and 2 backup
sites. The ports mechanism will, by default, try the 2 primary sites
several times (a number N which can be defined by the user in make.conf or
something similar), then it will fall back to the backup sites (which can
be disabled completely by a boolean variable).

And so it goes (I think I just lost my train of thought there)..

> Honestly, I don't really mind if a port doesn't respect some of the
> above as long as it doesn't use "-g" or "-O0" or something equally
> inane as their compilation option.  If the user wants to debug the
> port, they have to go into the source directory anyway.

Or if they want to HELP debug the program, they can send debuggable core
dumps to the author(s). Or simply use gdb to read the dump and send the
author the results of their backtrace/look/etc./etc.

In any case, the general meaninglessness of "-g" or "-O0" is fairly correct
for the average user, and as such the default CFLAGS in any program's
Makefile (as set by the port) should be something like "-O -pipe", although
I think "-O2 -pipe -Wall" is OK for most programs too.

-- 
Will Andrews <andrews@technologist.com>
GCS/E/S @d- s+:+>+:- a--->+++ C++ UB++++ P+ L- E--- W+++ !N !o ?K w---
?O M+ V-- PS+ PE++ Y+ PGP+>+++ t++ 5 X++ R+ tv+ b++>++++ DI+++ D+ 
G++>+++ e->++++ h! r-->+++ y?


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000506155542.J55274>