Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1818857 - Build python-debug with -O0 to avoid test_gdb failures
Summary: Build python-debug with -O0 to avoid test_gdb failures
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: python3.9
Version: rawhide
Hardware: Unspecified
OS: Unspecified
low
unspecified
Target Milestone: ---
Assignee: Charalampos Stratakis
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: ARMTracker 1926283
TreeView+ depends on / blocked
 
Reported: 2020-03-30 14:44 UTC by Charalampos Stratakis
Modified: 2021-02-08 15:03 UTC (History)
13 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
: 1926283 (view as bug list)
Environment:
Last Closed: 2021-01-29 13:53:09 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Charalampos Stratakis 2020-03-30 14:44:24 UTC
When compiling python3, test_gdb fails due to regex mismatch on the expected output, on s390x and arm 32 bits.

Traceback:

======================================================================
FAIL: test_basic_command (test.test_gdb.PyPrintTests)
Verify that the "py-print" command works
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 924, in test_basic_command
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n\'args\' not found\n' did not match ".*\\nlocal 'args' = \\(1, 2, 3\\)\\n.*"
======================================================================
FAIL: test_print_after_up (test.test_gdb.PyPrintTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 933, in test_print_after_up
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n#10 (frame information optimized out)\n\'c\' not found\n\'b\' not found\n\'a\' not found\n' did not match ".*\\nlocal 'c' = 3\\nlocal 'b' = 2\\nlocal 'a' = 1\\n.*"
======================================================================
FAIL: test_printing_builtin (test.test_gdb.PyPrintTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 949, in test_printing_builtin
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n\'len\' not found\n' did not match ".*\\nbuiltin 'len' = <built-in method len of module object at remote 0x-?[0-9a-f]+>\\n.*"
======================================================================
FAIL: test_printing_global (test.test_gdb.PyPrintTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 941, in test_printing_global
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n\'__name__\' not found\n' did not match ".*\\nglobal '__name__' = '__main__'\\n.*"
======================================================================
FAIL: test_basic_command (test.test_gdb.PyLocalsTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 958, in test_basic_command
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n' did not match '.*\\nargs = \\(1, 2, 3\\)\\n.*'
======================================================================
FAIL: test_locals_after_up (test.test_gdb.PyLocalsTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 967, in test_locals_after_up
    self.assertMultilineMatches(bt,
  File "/builddir/build/BUILD/Python-3.8.2/Lib/test/test_gdb.py", line 279, in assertMultilineMatches
    self.fail(msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 (builtin_id) pending.\n[Thread debugging using libthread_db enabled]\nUsing host libthread_db library "/lib/libthread_db.so.1".\n\nBreakpoint 1, builtin_id (self=, v=42) at /builddir/build/BUILD/Python-3.8.2/Python/bltinmodule.c:1171\n1171\t{\n#4 (frame information optimized out)\n#10 (frame information optimized out)\n' did not match '.*\\na = 1\\nb = 2\\nc = 3\\n.*'
----------------------------------------------------------------------
Ran 46 tests in 144.929s

Comment 1 Victor Stinner 2020-03-31 17:17:20 UTC
I validated with success that backporting my https://bugs.python.org/issue40019 change to 3.8 fix test_gdb on s390x and armv7hl.

See my PR https://src.fedoraproject.org/rpms/python3/pull-request/182 created to test that.

I backported my fix upstream from master to Python 3.7 and 3.8: https://bugs.python.org/issue40019.

This change only fix test_gdb. To get a fully working python-gdb.py, IMHO the best is to disable all compiler optimization using -O0 optimization level.

Comment 2 Charalampos Stratakis 2020-05-27 12:25:15 UTC
We need to not skip the test in the SPEC

Comment 3 Victor Stinner 2020-05-27 12:29:50 UTC
Python 3.8.3 has my fix https://github.com/python/cpython/commit/4ced9a7611ddfd923bd8f72aa61121d0e5aeb8fc

We can try to reenable test_gdb python3.9 package, and maybe also python 3.8 package.

Comment 4 Victor Stinner 2020-05-27 12:36:10 UTC
I created https://src.fedoraproject.org/rpms/python3.8/pull-request/1 to reenable test_gdb in python3.8. Let's see how it goes in CIs.

Comment 6 Miro Hrončok 2020-06-24 13:03:44 UTC
Let's build debug with -O0.

Comment 7 Petr Viktorin 2020-07-08 12:23:42 UTC
*** Bug 1846390 has been marked as a duplicate of this bug. ***

Comment 8 Victor Stinner 2020-08-10 14:48:19 UTC
FYI I modified test_gdb to skip it on gdb 9.2 and newer to work around a gdb issue:

* https://bugs.python.org/issue41473
* https://bugzilla.redhat.com/show_bug.cgi?id=1866884

"AArch64: sometimes, gdb fails to load symbols of a dynamic library with a pending breakpoint"

Fedora Rawhide is affected by the gdb issue.

Comment 9 Ben Cotton 2020-08-11 13:15:50 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 33 development cycle.
Changing version to 33.

Comment 10 Miro Hrončok 2020-08-20 16:16:14 UTC
So if we simply change -Og with -O0, we get warnings like this:

In file included from /usr/include/bits/libc-header-start.h:33,
                 from /usr/include/limits.h:26,
                 from /usr/lib/gcc/x86_64-redhat-linux/10/include/limits.h:195,
                 from /usr/lib/gcc/x86_64-redhat-linux/10/include/syslimits.h:7,
                 from /usr/lib/gcc/x86_64-redhat-linux/10/include/limits.h:34,
                 from /builddir/build/BUILD/Python-3.9.0rc1/Include/Python.h:11,
                 from /builddir/build/BUILD/Python-3.9.0rc1/Programs/_freeze_importlib.c:5:
/usr/include/features.h:397:4: warning: #warning _FORTIFY_SOURCE requires compiling with optimization (-O) [-Wcpp]
  397 | #  warning _FORTIFY_SOURCE requires compiling with optimization (-O)
      |    ^~~~~~~


If we add -Wp,-D_FORTIFY_SOURCE=0, we have warnings like this:

<command-line>: warning: "_FORTIFY_SOURCE" redefined
<command-line>: note: this is the location of the previous definition
<command-line>: warning: "_FORTIFY_SOURCE" redefined
<command-line>: note: this is the location of the previous definition
<command-line>: warning: "_FORTIFY_SOURCE" redefined
<command-line>: note: this is the location of the previous definition


More over, even if we replace the FORTIFY_SOURCE value with some sed magic, annobin warns as well:

annobin: /builddir/build/BUILD/Python-3.9.0rc1/Programs/_freeze_importlib.c: Warning: -D_FORTIFY_SOURCE defined as 0


I suppose we will always get a warning of some sort anyway, one way it to use -Wno-cpp with the debug build. WDYT?

Comment 11 Victor Stinner 2020-08-20 16:27:53 UTC
> More over, even if we replace the FORTIFY_SOURCE value with some sed magic, annobin warns as well: (...)

Is it possible to tell annobin to ignore -D_FORTIFY_SOURCE?

Comment 12 Miro Hrončok 2020-08-20 16:36:17 UTC
I don't know.

Comment 13 Petr Viktorin 2020-10-13 09:07:11 UTC
Here is a PR: https://src.fedoraproject.org/rpms/python3.9/pull-request/30

Comment 14 Petr Viktorin 2020-11-24 10:07:24 UTC
test_gdb is skipped upstream. Compiling with -Og doesn't fix it even in versions where it isn't skipped.
Restoring test_gdb should be done upstream before we do this.

Comment 15 Victor Stinner 2020-12-04 15:30:31 UTC
https://bugzilla.redhat.com/show_bug.cgi?id=1866884 has been fixed (gdb 10.1).

I reenabled test_gdb on gdb 9.2+ on Python upstream: https://github.com/python/cpython/commit/066394018a8463643cc63d933493f0afa99d72cc

I move the issue statuc back to NEW even if there is a (closed) PR: https://src.fedoraproject.org/rpms/python3.9/pull-request/30

Comment 16 Victor Stinner 2021-01-29 15:49:39 UTC
Woooot, thanks, that's a major enhancement to make gdb usable on python3.9-debug! No make "optimized out" issue which prevent to inspect anything ;-)


Note You need to log in before you can comment on or make changes to this bug.