Advisory: ======================== This update fixes detecting various bugs which should improve the stability of perl written programs: - Avoid panic when last value of search-and-replace is tainted and UTF-8 (RT#134409) - Fix overloading for binary and octal floats (RT#125557) - Fix handling undefined array members in Dumpvalue (RT#134441) - Fix taint mode documentation regarding @INC - Fix handling a layer argument in Tie::StdHandle::BINMODE() (RT#132475) - Fix an unintended upgrade to UTF-8 in the middle of a transliteration - Fix a race in File::stat() tests (GH#17234) - Fix GCC 10 version detection (GH#17295) - Fix a memory leak when compiling a regular expression with a non-word class (GH#17218) - Fix a memory leak when matching a UTF-8 regular expression (RT#134390) - Fix a detection for futimes (RT#134432) - Fix propagating non-string variables in an exception value (RT#134291) - Include trailing zero in scalars holding trie data (RT#134207) - Fix a use after free in /(?{...})/ (RT#134208) - Fix a use after free in debugging output of a collation - Fix file mode of a perl-example.stp example - Fix a NULL pointer dereference in PerlIOVia_pushed() - Fix a crash when setting $@ on unwinding a call stack (RT#134266) - Fix a documentation about a future API change - Do not panic when evaluating non-ASCII bare words (RT#134061) - Fix a crash in SIGALARM handler when waiting on a child process to be closed (RT#122112) - Fix a crash with a negative precision in sprintf function (RT#134008) - Prevent from wrapping a width in a numeric format string (RT#133913) - Fix subroutine protypes to track reference aliases (RT#134072) - Improve retrieving a scalar value of a variable modified in a signal handler (RT#134035) - Fix changing packet destination sent from a UDP IO::Socket object (RT#133936) - Fix a stack underflow in readline() if passed an empty array as an argument (#RT133989) - Fix %%{^CAPTURE_ALL} to be an alias for %%- variable (RT#131867) - Fix %%{^CAPTURE} value when used after @{^CAPTURE} (RT#134193) - Fix a test for a crash in SIGALARM handler when waiting on a child process to be closed (RT#122112) - Fix a crash on an uninitialized warning when processing a multideref node (RT#134275) - Preserve append mode when opening anonymous files (RT#134221) ======================== Updated packages in core/updates_testing: ======================== perl-5.28.2-1.mga7 perl-base-5.28.2-1.mga7 perl-devel-5.28.2-1.mga7 perl-doc-5.28.2-1.mga7
Mageia7, x86_64 These packages having been in place a few hours. The list of perl dependent packages is quite long. $ urpmq --whatrequires perl | sort -u | wc -l 2105 It includes a few games, iceape and gurpmi. Installed iceape using gurpmi and launched it. The trace did not show any direct access to perl but for gurpmi there were numerous references to perl5 and perl-tk. Played frozen-bubble. Installed boomaga. $ urpmq --requires-recursive boomaga | grep perl groff-perl perl perl-File-HomeDir perl-File-Sync perl-File-Which perl-MDK-Common perl-base No problems so far.
CC: (none) => tarazed25Whiteboard: (none) => MGA7-64-OK
Not OK. This is built as 5.28.2-1.mga7 which we already have in release And this confirms it that patches were added but no rel or subrel was bumped: http://svnweb.mageia.org/packages/updates/7/perl/current/SPECS/perl.spec?r1=1470079&r2=1470078&pathrev=1470079
Keywords: (none) => feedbackWhiteboard: MGA7-64-OK => (none)CC: (none) => tmb
Updated, release was bumped to 2 (clearer than using sub as here, cauldron's perl has been updated to 5.30)
Source RPM: perl-5.28.2-1.mga7 => perl-5.28.2-2.mga7Status: NEW => ASSIGNED
So to be clear, the rpms to test are now: perl-5.28.2-2.mga7 perl-base-5.28.2-2.mga7 perl-devel-5.28.2-2.mga7 perl-doc-5.28.2-2.mga7
Keywords: feedback => (none)
Keywords: (none) => advisory
64-bit Plasma system, i5-2500, integrated graphics. The following 3 packages are going to be installed: - perl-5.28.2-2.mga7.x86_64 - perl-base-5.28.2-2.mga7.x86_64 - perl-doc-5.28.2-2.mga7.noarch Packages installed cleanly. Following Len's lead, I used gurpmi to install Frozen Bubble, and dependencies. I played it afterward, and it worked OK. (I like the Android version on my tablet better, though...) If those tests are adequate, then the newer packages look OK here. With the number of bugs this is supposed to address, I'm going to send it on it's way. Validating, too.
Whiteboard: (none) => MGA7-64-OKKeywords: (none) => validated_updateCC: (none) => andrewsfarm, sysadmin-bugs
An update for this issue has been pushed to the Mageia Updates repository. https://advisories.mageia.org/MGAA-2020-0027.html
Resolution: (none) => FIXEDStatus: ASSIGNED => RESOLVED