Date: Sat, 19 Aug 1995 12:13:27 +0200 From: "Julian Stacey <jhs@freebsd.org>" <jhs@vector.eikon.e-technik.tu-muenchen.de> To: asami@cs.berkeley.edu (Satoshi Asami) Cc: gary@palmer.demon.co.uk, ports@FreeBSD.org Subject: Re: bsd.ports.mk checksum Message-ID: <199508191013.MAA06715@vector.eikon.e-technik.tu-muenchen.de> In-Reply-To: Your message of "Sat, 19 Aug 1995 01:15:09 PDT." <199508190815.BAA05947@silvia.HIP.Berkeley.EDU>
next in thread | previous in thread | raw e-mail | index | archive | help
> Hey calm down. Your argument doesn't make much sense. You mean you > have all the sources extracted, but don't have the distfiles lying > around? EXACTLY ! I wrote you that already. Never heard of a CD-ROM ? It's used for distributing bulk computer data, but vendors don't distribute one drive per media disk ;-) When one has extracted & compiled 800M of ports, & starts to maintain ports off the net, & regularly compiles ports, if one takes the 200M of distfiles on the cd-rom off line, to look at other cd-roms, md5 fails on `make all'. > Then all I can say is that you are a distinct minority. No, You are the minority ! Just 3 people have expressed an opinion on this, (listed chronologically): - Julian S wants it removed - Gary P wrote : I'm not sure how this ever became the case, it certainly was never my intention to do this with the checksum mechanism, as I know from experience that doing a MD5 checksum on (say) emacs, when the distfile is on a CDROM, is painful. - Satoshi wants it to stay. That's 2:1 for removal so far. Though I have 3 Gig, I still don't want to waste 200M/250M keeping distfiles I've already compiled on line, others may not want to either. It's unfortunate that you force us to choose between wasting 200M, or abandoning automatic use of md5 on new ports. > This check is > added to save users from strange errors, It merely detect if the distfile has been mangled, gone missing or been changed, but once the user has a compiled source tree, this is not of vital interest, & the user can do a period check for this with a make checksum, if they're truly paranoid (Hey even _I'm_ not that paranoid on checks ;-). > and I am not going to remove it. I hope you don't stay this inflexible, (it discourages one from regularly recompiling the tree & reporting errors, when you make it neccessary to waste an extra 200/250M ). Perhaps pride of authorship & spare disc capacity blind you to the nuisance it imposes on others ? Was this extension of md5 usage discussed before it was CVS commited ? I hope it wasn't just slipped through by a minority of one ;-) ( perhaps after no one happened to notice the change during a test phase on a test system which luxuriates in all distfiles on line, & is thus unrepresentative ? ) > If the setenv works for you, then let's all be happy. :) No ! How can we be happy with a horrible cludge that exists merely to avoid a problem you impose ? Try asking people if they appreciate you forcing them to waste 200M/250M permanently keeping all distfiles on line ! As you raised the subject of minority, let's invite more opinions, meantime may I remind you that currently, To Please The Majority (2:1) you should: remove the md5 invocation from make all, & just leave it on make extract Thanks Julian S
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508191013.MAA06715>