On Feb 25, 2011, at 10:10 AM, pinto.e...@gmail.com wrote: > I know perfectly your Opinion. My self was, perhaps 4 years ago, against to > disabling tout court this feature. for this i have put a blueprint for this > and Some time ago i have committed this change on HEAD. Probably i have > missed your thought on these change . I have belenix and openindiana in mind > in doing this, and the patch only permit to disable, but the default is > enforcing the deps. Have i to revert however ? No problem at all for me. What > do you think ?
These days the only anwer that remains to be tried is
Fork every distro on the planet and "fix" the packaging.
Seriously: I don't know what to do with "No time" and *shrug* answers
from package monkeys and distro vendors and "standards" organizations
that refuse to change.
The implementation was an RFE for an open source project that was discussed
and voted on and implemented with all the usual enablers/disablers necessary
for the transition. The whole issue is
What SHOULD the default settings for 2 bits be?
which leads to 4 cases and endless discussion ensues and (in this case)
a 3rd or 4th mechanism to set those 2 bits in best "Have it your own way!"
fashion
(aside)
Please don't take my comments personally. Your current
checkin is better AutoFu narrowly targeted at specific goal,
unlike macors (which are dependent on configuration) or RPM_VENDOR_WHATEVER
(which is compile time, not run-time, and implicitlt assumes that
vendors are actually capable of make engineering, not "business",
decisions).
My work was done years ago. Yet there's this endless fascination with
bruned out car wrecks like Suggests: and "LSB package format" (nee RPMv3)
and here with parent directory dependencies.
Huge waste of time imho. When all is said and done, a developer chooses what
to implement and moves on.
73 de Jeff
> -----Original Message-----
> From: Jeff Johnson
> Sent: 25/02/2011, 15:22
> To: rpm-devel@rpm5.org
> Subject: Re: [CVS] RPM: rpm-5_4: rpm/ CHANGES configure.ac rpm/lib/ depends.c
>
>
> Nothing wrong with your AutoFu whatsoever.
>
> I'm not sure that AutoFu is the best solution here.
>
> The implementation was done like 7 years ago, and the RFE
> from Enrico for a clear policy wrto directory inclusion
> in *.rpm behavior goes back even further.
>
> Adding "Have it your own way!" AutoFu is just gonna
> prolong the transition forevermore.
>
> Ditto macros for persistence, and vendors for distros, and LSB for standards:
>
> All this "choice" makes it impossible to identify or implement or test
> "correct" RPM behavior.
>
> All the behavior is contextually sensitive.
>
> No matter what:
>
> Sooner or later you just gotta stop being lazy and actually
> fix your packages and determine a consistent policy for directories
> in *.rpm packages. "No time." == "The dog ate my homework."
>
> 73 de Jeff
>
>
> On Feb 25, 2011, at 4:26 AM, Pinto Elia wrote:
>
>> RPM Package Manager, CVS Repository
>> http://rpm5.org/cvs/
>> ____________________________________________________________________________
>>
>> Server: rpm5.org Name: Pinto Elia
>> Root: /v/rpm/cvs Email: devzero2...@rpm5.org
>> Module: rpm Date: 25-Feb-2011 10:26:35
>> Branch: rpm-5_4 Handle: 2011022509263401
>>
>> Modified files: (Branch: rpm-5_4)
>> rpm CHANGES configure.ac
>> rpm/lib depends.c
>>
>> Log:
>> permit in configure phase to enable the
>> optional-dirname-and-symlink-deps used by Mandriva and Ark today
>> (ref
>> https://blueprints.launchpad.net/rpm/+spec/rpm-split-vendor-config-in-autofu)
>>
>> Summary:
>> Revision Changes Path
>> 1.3501.2.63 +3 -0 rpm/CHANGES
>> 2.472.2.8 +13 -0 rpm/configure.ac
>> 1.461.4.5 +4 -4 rpm/lib/depends.c
>> ____________________________________________________________________________
>>
>> patch -p0 <<'@@ .'
>> Index: rpm/CHANGES
>> ============================================================================
>> $ cvs diff -u -r1.3501.2.62 -r1.3501.2.63 CHANGES
>> --- rpm/CHANGES 24 Feb 2011 19:33:15 -0000 1.3501.2.62
>> +++ rpm/CHANGES 25 Feb 2011 09:26:34 -0000 1.3501.2.63
>> @@ -1,4 +1,7 @@
>> 5.4.0 -> 5.4.1:
>> + - devzero2000: permit in configure phase to enable the
>> + optional-dirname-and-symlink-deps used by Mandriva and Ark today
>> + (ref
>> https://blueprints.launchpad.net/rpm/+spec/rpm-split-vendor-config-in-autofu)
>> - proyvind: use expanded %{___NVRA} for 'rpm --last'.
>> - jbj: valgrind: fix: typ in macros building w/o <valgrind/valgrind.h>.
>> - proyvind: set mi->mi_count in rpmmiCount() when db cursor exists
>> rather
>> @@ .
>> patch -p0 <<'@@ .'
>> Index: rpm/configure.ac
>> ============================================================================
>> $ cvs diff -u -r2.472.2.7 -r2.472.2.8 configure.ac
>> --- rpm/configure.ac 9 Jan 2011 03:55:21 -0000 2.472.2.7
>> +++ rpm/configure.ac 25 Feb 2011 09:26:34 -0000 2.472.2.8
>> @@ -351,6 +351,19 @@
>> [ENABLE_BUILD_LAFILES],
>> [test ".$do_build_lafiles" = .yes])
>>
>> +dnl # optional sysmlink and parentdir dependency
>> +AC_ARG_ENABLE(
>> + [dirname-and-symlink-deps],
>> + AS_HELP_STRING([--disable-dirname-and-symlink-deps],[permit to disable
>> via macro the automatic dirname and symlink dependecy]),
>> + [ do_optional_dirname_and_symlink_deps=$enableval ],
>> + [ do_optional_dirname_and_symlink_deps=yes ])
>> +
>> +AS_IF([test "x$do_optional_dirname_and_symlink_deps" != xyes],
>> + [
>> + AC_DEFINE([RPM_OPTIONAL_DIRNAME_AND_SYMLINK_DEPS],[1],[permit to disable
>> via macro the automatic dirname and symlink dependecy])
>> + ])
>> +
>> +
>> dnl # FIXME: partly Linux-specific only!?
>> dnl # (but at least it usually doesn't harm on other platforms)
>> CFLAGS="$CFLAGS -D_GNU_SOURCE -D_REENTRANT"
>> @@ .
>> patch -p0 <<'@@ .'
>> Index: rpm/lib/depends.c
>> ============================================================================
>> $ cvs diff -u -r1.461.4.4 -r1.461.4.5 depends.c
>> --- rpm/lib/depends.c 8 Jan 2011 23:10:10 -0000 1.461.4.4
>> +++ rpm/lib/depends.c 25 Feb 2011 09:26:35 -0000 1.461.4.5
>> @@ -1567,7 +1567,7 @@
>> int terminate = 2; /* XXX terminate if rc >= terminate */
>> int rc;
>> int ourrc = 0;
>> -#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) /*
>> optional-dirname-and-symlink-deps */
>> +#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) ||
>> defined(RPM_OPTIONAL_DIRNAME_AND_SYMLINK_DEPS) /*
>> optional-dirname-and-symlink-deps */
>> int dirname_deps;
>> int symlink_deps;
>> #endif
>> @@ -1646,7 +1646,7 @@
>> }
>> }
>>
>> -#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) /*
>> optional-dirname-and-symlink-deps */
>> +#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) ||
>> defined(RPM_OPTIONAL_DIRNAME_AND_SYMLINK_DEPS) /*
>> optional-dirname-and-symlink-deps */
>> dirname_deps =
>> rpmExpandNumeric("%{?_check_dirname_deps}%{?!_check_dirname_deps:1}");
>> if (dirname_deps) {
>> #endif
>> @@ -1690,7 +1690,7 @@
>> /*@switchbreak@*/ break;
>> }
>> }
>> -#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) /*
>> optional-dirname-and-symlink-deps */
>> +#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) ||
>> defined(RPM_OPTIONAL_DIRNAME_AND_SYMLINK_DEPS) /*
>> optional-dirname-and-symlink-deps */
>> }
>>
>> symlink_deps =
>> rpmExpandNumeric("%{?_check_symlink_deps}%{?!_check_symlink_deps:1}");
>> @@ -1738,7 +1738,7 @@
>> /*@switchbreak@*/ break;
>> }
>> }
>> -#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) /*
>> optional-dirname-and-symlink-deps */
>> +#if defined(RPM_VENDOR_MANDRIVA) || defined(RPM_VENDOR_ARK) ||
>> defined(RPM_OPTIONAL_DIRNAME_AND_SYMLINK_DEPS) /*
>> optional-dirname-and-symlink-deps */
>> }
>> #endif
>>
>> @@ .
>> ______________________________________________________________________
>> RPM Package Manager http://rpm5.org
>> CVS Sources Repository rpm-...@rpm5.org
>
>
> ______________________________________________________________________
> RPM Package Manager http://rpm5.org
> Developer Communication List rpm-devel@rpm5.org
smime.p7s
Description: S/MIME cryptographic signature
