Date: Wed, 23 May 2001 23:47:48 +0300 From: Peter Pentchev <roam@orbitel.bg> To: "David W. Chapman Jr." <dwcjr@inethouston.net> Cc: "Scot W. Hetzel" <hetzels@westbend.net>, freebsd-ports@FreeBSD.org Subject: Re: ports/27563: postfix-current is broken Message-ID: <20010523234748.D535@ringworld.oblivion.bg> In-Reply-To: <004101c0e3c4$73e11580$931576d8@inethouston.net>; from dwcjr@inethouston.net on Wed, May 23, 2001 at 03:10:46PM -0500 References: <200105231920.f4NJK3W05339@freefall.freebsd.org> <20010523230339.B535@ringworld.oblivion.bg> <004101c0e3c4$73e11580$931576d8@inethouston.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 23, 2001 at 03:10:46PM -0500, David W. Chapman Jr. wrote: > > > I just did a grep thru the strings returned from the libsasl.a library > and > > > it returned the MySQL & LDAP functions that it is using. Theses > functions > > > will not occur in a SASL library that hasn't been compiled with support > for > > > MySQL and/or LDAP. So a check for MySQL and/or LDAP would need: > > > > > > strings PREFIX/lib/libsasl.a | grep mysql_verify_password > > > strings PREFIX/lib/libsasl.a | grep ldap_verify_password > > > > Just a minor comment: this would be better done with nm(1) or objdump(1), > > not strings(1). nm(1)'s -u option seems especially relevant. > > > > Still, I think it would be better to leave this to the user - let them > > define WITH_MYSQL or WITH_LDAP or whatever is appropriate. If those > > defines are the same that are used in the dependent ports (in this case, > > SASL), you get the added benefit that a build on a clean machine would > > keep those in the environment when building the dependent ports, and > > provide a one-step full build. > > > > I hope this last sentence did not come out too unclear - all I meant was, > > if SASL is not installed, and you do a make -DWITH_MYSQL in the Postfix > > port directory, the automagically-invoked 'make' for the SASL port > inherits > > this WITH_MYSQL and DTRT. In this scenario, even if Postfix does not > honor > > the WITH_MYSQL flag, you would have to define it anyway for the build > > of the dependent port, so it makes no sense for the Postfix port to use > > a different, additional, non-standard, > not-really-certain-to-be-error-proof > > mechanism to detect the need for -lmysqlclient and such. > > It is a little more complex than that. We need to see how sasl is compiled > because however it is compiled, postfix must be compiled or it will not > work. SASL defaults to using mysql while postfix doesn't, so by default > they are incompatible. On the other side of the sword, postfix can use > mysql for 2 different things along with ldap, ie postfix has some things of > its own that it can use mysql for without sasl. but the bottom line is if > you compile sasl with krb or mysql, postfix won't compile without them > either. Maybe I partly didn't understand what you were saying, let me know > if that's the case. Yes, you did understand what I was saying, or at least you did, if what you think I was saying is 'leave it to the user to give the proper config options/defines'. This is precisely what I mean, and what I still think, even as you're saying that the options are critical. Give the user a prominent warning in a pre-everything or pre-configure or something stage, or in the configuration dialog if there is one, and then don't blame yourself for their lack of basic reading comprehension skills :) Sure, this could be viewed as 'give them enough rope to hang themselves', but, well, I really don't see how this could be done in a more elegant way (and I do consider running nm(1) on libraries a less than elegant solution). G'luck, Peter -- What would this sentence be like if pi were 3? 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?20010523234748.D535>