I've prepared repepro 3.5.2-2 with some fixes from 3.6.0 (which is in
experimental, so this could go via unstable).
The debdiff is as follows:
diff -u reprepro-3.5.2/debian/changelog reprepro-3.5.2/debian/changelog
--- reprepro-3.5.2/debian/changelog
+++ reprepro-3.5.2/debian/changelog
@@ -1,3 +1,12 @@
+reprepro (3.5.2-2) unstable; urgency=low
+
+ * backport some bugfixes of 3.6.0 to 3.5.2:
+ - no longer double-free with --export=never
+ - no longer trigger assert() in wrong argument of list arguments
+ - no longer trigger assert() in restorefiler
+
+ -- Bernhard R. Link <[EMAIL PROTECTED]> Wed, 06 Aug 2008 17:10:04 +0200
+
reprepro (3.5.2-1) unstable; urgency=medium
* new bugfix release
only in patch2:
unchanged:
--- reprepro-3.5.2.orig/distribution.c
+++ reprepro-3.5.2/distribution.c
@@ -792,7 +792,7 @@
if( when == EXPORT_NEVER ) {
if( verbose > 10 )
fprintf(stderr, "Not exporting anything as
--export=never specified\n");
- return distribution_freelist(distributions);
+ return RET_NOTHING;
}
for( d=distributions; d != NULL; d = d->next ) {
only in patch2:
unchanged:
--- reprepro-3.5.2.orig/main.c
+++ reprepro-3.5.2/main.c
@@ -1330,7 +1330,7 @@
struct distribution *destination;
retvalue result, r;
- assert( argc == 3 );
+ assert( argc == 4 );
result = distribution_get(alldistributions, argv[1], true,
&destination);
assert( result != RET_NOTHING );
@@ -2417,7 +2417,7 @@
{"removesrc", A_D(removesrc),
2, 3, "removesrc <codename> <source-package-names>
[<source-version>]"},
{"list", A_ROBact(list),
- 2, -1, "[-C <component>] [-A <architecture>] [-T <type>] list
<codename> <package-name>"},
+ 2, 2, "[-C <component>] [-A <architecture>] [-T <type>] list
<codename> <package-name>"},
{"listfilter", A_ROBact(listfilter),
2, 2, "[-C <component>] [-A <architecture>] [-T <type>]
listfilter <codename> <term to describe which packages to list>"},
{"removefilter", A_Dact(removefilter),
Is there a chance to get a freeze exception for that?
(Note that assert (and glibc's variant in memory handling) cause
reprepro not to touch a repository again until manually repaired, so
this is not only a cosmetic problem).
Hochachtungsvoll,
Bernhard R. Link
--
"Never contain programs so few bugs, as when no debugging tools are available!"
Niklaus Wirth
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
