skip ldconfig optimization

http://lists.pld-linux.org/mailman/pipermail/pld-devel-en/2016-May/024910.html

diff --git a/rpm.spec b/rpm.spec
index 2d6d3bba1dc8e9c452c459e80c4842926908aa70..1c6d227f8bf0938ef1beb79365a1f3ee903e8802 100644 (file)
--- a/rpm.spec
+++ b/rpm.spec
@@ -51,7 +51,7 @@ Summary(ru.UTF-8):    Менеджер пакетов от RPM
 Summary(uk.UTF-8):     Менеджер пакетів від RPM
 Name:          rpm
 Version:       5.4.15
-Release:       33
+Release:       34
 License:       LGPL v2.1
 Group:         Base
 # http://rpm5.org/files/rpm/rpm-5.4/rpm-5.4.15-0.20140824.src.rpm
@@ -182,6 +182,7 @@ Patch92:    fast_python_deps.patch
 Patch93:       python2_explicit.patch
 Patch94:       do_not_write_before_macro_buffer.patch
 Patch95:       rpm-python-spec-header.patch
+Patch96:       skip-ldconfig-optimization.patch
 
 # Patches imported from Mandriva
 
@@ -1022,6 +1023,7 @@ cd -
 %patch93 -p1
 %patch94 -p1
 %patch95 -p1
+%patch96 -p1
 
 %patch1050 -p1
 

diff --git a/skip-ldconfig-optimization.patch b/skip-ldconfig-optimization.patch
new file mode 100644 (file)
index 0000000..917a8fc
--- /dev/null
+++ b/skip-ldconfig-optimization.patch
@@ -0,0 +1,74 @@
+http://lists.pld-linux.org/mailman/pipermail/pld-devel-en/2016-May/024910.html
+
+patch from jbj answer:
+http://lists.pld-linux.org/mailman/pipermail/pld-devel-en/2016-May/024913.html
+
+On May 31, 2016, at 2:52 AM, Elan Ruusamäe wrote:
+
+> On 31.05.2016 09:45, Elan Ruusamäe wrote:
+>> [root@57c8cccdc671 bin]# poldek -u procps 
+> oh, and it's 100% reproducible:
+>
+
+Yes. And it has been discussed many times over the last 15y,
+several times on this list.
+
+Short answer:
+       Disable the mechanism with attached patch to lib/psm.c
+               - OR --
+       Don't use
+               %scriptlet -p /sbin/ldconfig
+       in packages that are known to invoke an executable that depends on
+       a new library.
+
+Longer answer:
+       Libraries often have scriptlets like this:
+               %post -p /sbin/ldconfig
+               %postun -p /sbin/ldconfig
+       (or the equivalent within a scriptlet body).
+
+       On upgrade, with install before erase, this results in 2 executions of /sbin/ldconfig.
+
+       One of those invocations is redundant when a library is being replaced, the
+       install %post is sufficient to make the library avaialble for loading.
+
+       The redundant invocation of /sbin/ldconfig  was deemed a performance flaw @redhat,
+       where I was asked to implement the current optimization.
+
+       So the RPM optimization detects when /sbin/ldconfig has just been run, and skips the next
+               %scriptlet -p /sbin/ldconfig
+       (when written exactly like that) if /sbin/ldconfig has been executed.
+
+       Issues are only seen when a scriptlet contains an executable that needs
+       a changed API: this is usually a very rare occurrence
+
+       There are other flaws with the optimization when dependency loops break the sorted order
+       (and/or --noorder is used). , and the fix of running /sbin/ldconfig manually is rather simple
+       (even if the problem is obscure).
+
+       There is another flaw (as in your report), when the very last package
+       has a pending skipped /sbin/ldconfig invocation.
+
+       The fix of running /sbin/ldconfig manually is rather simple
+       (even if the problem is obscure).
+
+hth
+
+73 de Jeff
+===================================================================
+RCS file: /v/rpm/cvs/rpm/lib/psm.c,v
+retrieving revision 2.399.2.20
+diff -p -u -w -r2.399.2.20 psm.c
+--- ./lib/psm.c        11 May 2015 21:10:16 -0000      2.399.2.20
++++ ./lib/psm.c        31 May 2016 15:19:20 -0000
+@@ -800,7 +800,7 @@ static rpmRC runEmbeddedScript(rpmpsm ps
+ static int ldconfig_done = 0;
+ 
+ /*@unchecked@*/ /*@observer@*/ /*@null@*/
+-static const char * ldconfig_path = "/sbin/ldconfig";
++static const char * ldconfig_path = NULL;
+ 
+ /**
+  * Run scriptlet with args.
+
+