GDB 7.6.90 available for testing

GDB 7.6.90 available for testing

[Joel Brobecker]

Hello,

I have just finished creating the gdb-7.6.90 pre-release.
It is available for download at the following location:

    ftp://sourceware.org/pub/gdb/snapshots/branch/gdb-7.6.90.tar.bz2

A gzip'ed version is also available: gdb-7.6.90.tar.gz.

Please give it a test if you can and report any problems you might find.

On behalf of all the GDB contributors, thank you!
-- 
Joel Brobecker

Re: GDB 7.6.90 available for testing

[Reini Urban]

On 01/08/2014 04:14 AM, Joel Brobecker wrote:
>      ftp://sourceware.org/pub/gdb/snapshots/branch/gdb-7.6.90.tar.bz2

I got those FAIL regressions from 7.6.50:

FAIL: gdb.base/annota1.exp: run until main breakpoint (timeout)
FAIL: gdb.base/attach-pie-noexec.exp: info shared
FAIL: gdb.base/attach-pie-noexec.exp: info shared
FAIL: gdb.base/completion.exp: complete 'file ./gdb.base/complet' (timeout)
FAIL: gdb.base/completion.exp: complete 'info func marke' (timeout)
FAIL: gdb.base/completion.exp: complete 'set follow-fork-mode' (timeout)
FAIL: gdb.base/completion.exp: field completion with invalid field (got 
interactive prompt)
FAIL: gdb.base/corefile.exp: core-file warning-free
FAIL: gdb.base/longjmp.exp: next over longjmp(1)
FAIL: gdb.base/longjmp.exp: next over call_longjmp (2)
FAIL: gdb.base/morestack.exp: continue
FAIL: gdb.base/restore.exp: caller5 calls callee1; return restored l1 to 
32492
FAIL: gdb.base/restore.exp: caller5 calls callee2; return restored l1 to 
32492
FAIL: gdb.base/restore.exp: caller5 calls callee3; return restored l1 to 
32492
FAIL: gdb.base/restore.exp: caller5 calls callee4; return restored l1 to 
32492
FAIL: gdb.base/restore.exp: caller5 calls callee5; return restored l1 to 
32492
FAIL: gdb.base/store.exp: var doublest l; print old l, expecting -1
FAIL: gdb.base/store.exp: var doublest l; print old r, expecting -2
FAIL: gdb.base/store.exp: var doublest l; setting l to 4
FAIL: gdb.base/store.exp: var doublest l; print new l, expecting 4
FAIL: gdb.base/store.exp: var doublest l; print incremented l, expecting 2
FAIL: gdb.base/store.exp: upvar doublest l; print old l, expecting -1
FAIL: gdb.base/store.exp: upvar doublest l; print old r, expecting -2
FAIL: gdb.base/store.exp: upvar doublest l; set l to 4
FAIL: gdb.base/store.exp: upvar doublest l; print new l, expecting 4
FAIL: gdb.base/valgrind-infcall.exp: target remote for vgdb
Running ./gdb.dwarf2/dw2-compressed.exp ...
ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
cannot stat './gdb.dwarf2/file1.txt': No such file or directory
FAIL: gdb.dwarf2/dw2-compressed.exp: list func_cu1
FAIL: gdb.dwarf2/dw2-filename.exp: interpreter-exec mi 
-file-list-exec-source-files
FAIL: gdb.dwarf2/dw2-filename.exp: info sources
ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
cannot stat './gdb.dwarf2/file1.txt': No such file or directory
FAIL: gdb.dwarf2/dw2-intercu.exp: list func_cu1
Running ./gdb.dwarf2/dw2-intermix.exp ...
ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
cannot stat './gdb.dwarf2/file1.txt': No such file or directory
FAIL: gdb.dwarf2/dw2-intermix.exp: list func_cul
ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
cannot stat './gdb.dwarf2/file1.txt': No such file or directory
FAIL: gdb.dwarf2/dw2-producer.exp: list func_cu1
ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
cannot stat './gdb.dwarf2/file1.txt': No such file or directory
FAIL: gdb.dwarf2/mac-fileno.exp: list func_cu1
FAIL: gdb.fortran/array-element.exp: continue to breakpoint once again 
(the program exited)
FAIL: gdb.fortran/array-element.exp: print the second element of array a
FAIL: gdb.java/jmain.exp: compilation jmain.java
FAIL: gdb.java/jmisc.exp: compilation jmisc.java
FAIL: gdb.java/jnpe.exp: compilation jnpe.java
FAIL: gdb.java/jprint.exp: compilation jprint.java
FAIL: gdb.mi/mi-var-rtti.exp: list children of s.public in 
type_update_when_use_rtti
FAIL: gdb.mi/mi-var-rtti.exp: list children of s.ptr (without RTTI) in 
type_update_when_use_rtti
FAIL: gdb.mi/mi-var-rtti.exp: list children of s.ptr.public (without 
RTTI) in type_update_when_use_rtti

./gdb.opencl/* fails all

FAIL: gdb.python/py-prettyprint.exp: print ss
FAIL: gdb.python/py-prettyprint.exp: print ssa[1]
FAIL: gdb.python/py-prettyprint.exp: print ssa
FAIL: gdb.python/py-prettyprint.exp: print arraystruct
FAIL: gdb.python/py-prettyprint.exp: print ns2
FAIL: gdb.python/py-prettyprint.exp: print x
FAIL: gdb.python/py-prettyprint.exp: print cstring
FAIL: gdb.python/py-prettyprint.exp: print estring
FAIL: gdb.python/py-prettyprint.exp: print estring2
FAIL: gdb.python/py-prettyprint.exp: print hint_error
FAIL: gdb.python/py-prettyprint.exp: print c
FAIL: gdb.python/py-prettyprint.exp: print nstype
FAIL: gdb.python/py-prettyprint.exp: print nstype on one line
ERROR: remote_download to host of ./gdb.python/py-prettyprint.py to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.python/py-prettyprint.py: 
cp: cannot stat './gdb.python/py-prettyprint.py': No such file or directory
FAIL: gdb.python/py-prettyprint.exp: print ss
FAIL: gdb.python/py-prettyprint.exp: print ssa[1]
FAIL: gdb.python/py-prettyprint.exp: print ssa
FAIL: gdb.python/py-prettyprint.exp: print arraystruct
FAIL: gdb.python/py-prettyprint.exp: print cps
FAIL: gdb.python/py-prettyprint.exp: print cpss
FAIL: gdb.python/py-prettyprint.exp: print cpssa[0]
FAIL: gdb.python/py-prettyprint.exp: print cpssa[1]
FAIL: gdb.python/py-prettyprint.exp: print cpssa
FAIL: gdb.python/py-prettyprint.exp: print sss
FAIL: gdb.python/py-prettyprint.exp: print ref
FAIL: gdb.python/py-prettyprint.exp: print derived
FAIL: gdb.python/py-prettyprint.exp: print ns
FAIL: gdb.python/py-prettyprint.exp: print ns
FAIL: gdb.python/py-prettyprint.exp: print ns
FAIL: gdb.python/py-prettyprint.exp: print ns2
FAIL: gdb.python/py-prettyprint.exp: print x
FAIL: gdb.python/py-prettyprint.exp: print cstring
FAIL: gdb.python/py-prettyprint.exp: print estring
FAIL: gdb.python/py-prettyprint.exp: print estring2
FAIL: gdb.python/py-prettyprint.exp: print hint_error
FAIL: gdb.python/py-prettyprint.exp: print c
FAIL: gdb.python/py-prettyprint.exp: print nstype
FAIL: gdb.python/py-prettyprint.exp: print nstype on one line
ERROR: remote_download to host of ./gdb.python/py-prettyprint.py to 
/home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.python/py-prettyprint.py: 
cp: cannot stat './gdb.python/py-prettyprint.py': No such file or directory
FAIL: gdb.python/py-prettyprint.exp: info locals
FAIL: gdb.python/py-prettyprint.exp: print ss enabled #1
FAIL: gdb.python/py-prettyprint.exp: python disable_lookup_function ()
FAIL: gdb.python/py-prettyprint.exp: python enable_lookup_function ()
FAIL: gdb.python/py-prettyprint.exp: print ss enabled #2

# of expected passes		24906
# of unexpected failures	6079
# of unexpected successes	1
# of expected failures		56
# of known failures		61
# of unresolved testcases	11
# of untested testcases		25
# of unsupported tests		136

on an uptodate debian testing amd64.
I hope you can replicate that.
-- 
Reini

Re: GDB 7.6.90 available for testing

[Joel Brobecker]

> I got those FAIL regressions from 7.6.50:

Thanks.

If you find regressions compared to 7.6.x, or issues that you would
consider blocking for 7.7, we would be very interested too!

I would never expect the testsuite to produce clean results, because
the results depend too much on the testing environment.

The number of FAILs you are getting is high, but that too, would
depend on the host and target that you tested against, which I could
not find in your message (maybe I missed it).

-- 
Joel

Re: GDB 7.6.90 available for testing

[Tom Tromey]

>>>>> "Reini" == Reini Urban <reini@cpanel.net> writes:

Reini> I got those FAIL regressions from 7.6.50:
[...]

Reini> on an uptodate debian testing amd64.
Reini> I hope you can replicate that.

I checked out and built the 7.7 branch here.

I then spot-checked my results against yours -- I did not check every
test, but a reasonable subset of them.  My build did not have any of the
failures you are seeing.

Offhand I don't know what went wrong in your test run.  I guess you'd
have to look into gdb.log to find out.


Joel, I did notice this -- not a regression, but a new test that fails:

gdb.ada/mi_interface.exp: Create ggg1 varobj: FAIL
gdb.ada/mi_interface.exp: list ggg1's children: FAIL


From gdb.log:

-var-create ggg1 * ggg1
^error,msg="-var-create: unable to create variable object"
(gdb) 

I didn't dig any deeper though.

Tom

Regressions on s390x RHEL-6.5 (was: Re: GDB 7.6.90 available for testing)

[Sergio Durigan Junior]

On Wednesday, January 08 2014, Joel Brobecker wrote:

> A gzip'ed version is also available: gdb-7.6.90.tar.gz.
>
> Please give it a test if you can and report any problems you might find.

Thank you very much for all the effort you put in the release process,
Joel.  I second what Tom already said in another message :-).

Having said that, I am now testing the 7.7 against some architectures of
interest.  I will send e-mails separately for different arch's, to
organize things better.  These are the regressions I found (and have not
investigated yet) on s390x with RHEL-6.5.

  -PASS: gdb.server/server-kill.exp: tstatus
  +FAIL: gdb.server/server-kill.exp: tstatus

  -PASS: gdb.trace/mi-traceframe-changed.exp: tfile: -trace-find frame-number 0
  +FAIL: gdb.trace/mi-traceframe-changed.exp: tfile: -trace-find frame-number 0

There are also some newly introduced failures:

  +FAIL: gdb.dwarf2/dwz.exp: p the_int

  +FAIL: gdb.dwarf2/implptrpiece.exp: print/d p[-1]

  +FAIL: gdb.server/wrapper.exp: continue to marker
  +FAIL: gdb.server/wrapper.exp: print d

  +FAIL: gdb.trace/entry-values.exp: disassemble foo+52,+10

As I said, I still haven't investigated them, but I plan to do so soon
enough (for now I'm still running tests on other arch's).

Thanks,

-- 
Sergio

Re: GDB 7.6.90 available for testing

[Yao Qi]

On 01/09/2014 02:18 AM, Reini Urban wrote:

> FAIL: gdb.base/completion.exp: complete 'file ./gdb.base/complet' (timeout)
> FAIL: gdb.base/completion.exp: complete 'info func marke' (timeout)
> FAIL: gdb.base/completion.exp: complete 'set follow-fork-mode' (timeout)
> FAIL: gdb.base/completion.exp: field completion with invalid field (got 
> interactive prompt)

> FAIL: gdb.base/longjmp.exp: next over longjmp(1)
> FAIL: gdb.base/longjmp.exp: next over call_longjmp (2)

> FAIL: gdb.base/valgrind-infcall.exp: target remote for vgdb

> Running ./gdb.dwarf2/dw2-compressed.exp ...
> ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
> cannot stat './gdb.dwarf2/file1.txt': No such file or directory
> FAIL: gdb.dwarf2/dw2-compressed.exp: list func_cu1

> FAIL: gdb.dwarf2/dw2-filename.exp: interpreter-exec mi 
> -file-list-exec-source-files
> FAIL: gdb.dwarf2/dw2-filename.exp: info sources
> ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
> cannot stat './gdb.dwarf2/file1.txt': No such file or directory
> FAIL: gdb.dwarf2/dw2-intercu.exp: list func_cu1
> Running ./gdb.dwarf2/dw2-intermix.exp ...
> ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
> cannot stat './gdb.dwarf2/file1.txt': No such file or directory
> FAIL: gdb.dwarf2/dw2-intermix.exp: list func_cul
> ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
> cannot stat './gdb.dwarf2/file1.txt': No such file or directory
> FAIL: gdb.dwarf2/dw2-producer.exp: list func_cu1
> ERROR: remote_download to host of ./gdb.dwarf2/file1.txt to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.dwarf2/file1.txt: cp: 
> cannot stat './gdb.dwarf2/file1.txt': No such file or directory
> FAIL: gdb.dwarf2/mac-fileno.exp: list func_cu1
> FAIL: gdb.fortran/array-element.exp: continue to breakpoint once again 
> (the program exited)

> FAIL: gdb.mi/mi-var-rtti.exp: list children of s.public in 
> type_update_when_use_rtti
> FAIL: gdb.mi/mi-var-rtti.exp: list children of s.ptr (without RTTI) in 
> type_update_when_use_rtti
> FAIL: gdb.mi/mi-var-rtti.exp: list children of s.ptr.public (without 
> RTTI) in type_update_when_use_rtti

> FAIL: gdb.python/py-prettyprint.exp: print ss
> FAIL: gdb.python/py-prettyprint.exp: print ssa[1]
> FAIL: gdb.python/py-prettyprint.exp: print ssa
> FAIL: gdb.python/py-prettyprint.exp: print arraystruct
> FAIL: gdb.python/py-prettyprint.exp: print ns2
> FAIL: gdb.python/py-prettyprint.exp: print x
> FAIL: gdb.python/py-prettyprint.exp: print cstring
> FAIL: gdb.python/py-prettyprint.exp: print estring
> FAIL: gdb.python/py-prettyprint.exp: print estring2
> FAIL: gdb.python/py-prettyprint.exp: print hint_error
> FAIL: gdb.python/py-prettyprint.exp: print c
> FAIL: gdb.python/py-prettyprint.exp: print nstype
> FAIL: gdb.python/py-prettyprint.exp: print nstype on one line
> ERROR: remote_download to host of ./gdb.python/py-prettyprint.py to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.python/py-prettyprint.py: 
> cp: cannot stat './gdb.python/py-prettyprint.py': No such file or directory
> FAIL: gdb.python/py-prettyprint.exp: print ss
> FAIL: gdb.python/py-prettyprint.exp: print ssa[1]
> FAIL: gdb.python/py-prettyprint.exp: print ssa
> FAIL: gdb.python/py-prettyprint.exp: print arraystruct
> FAIL: gdb.python/py-prettyprint.exp: print cps
> FAIL: gdb.python/py-prettyprint.exp: print cpss
> FAIL: gdb.python/py-prettyprint.exp: print cpssa[0]
> FAIL: gdb.python/py-prettyprint.exp: print cpssa[1]
> FAIL: gdb.python/py-prettyprint.exp: print cpssa
> FAIL: gdb.python/py-prettyprint.exp: print sss
> FAIL: gdb.python/py-prettyprint.exp: print ref
> FAIL: gdb.python/py-prettyprint.exp: print derived
> FAIL: gdb.python/py-prettyprint.exp: print ns
> FAIL: gdb.python/py-prettyprint.exp: print ns
> FAIL: gdb.python/py-prettyprint.exp: print ns
> FAIL: gdb.python/py-prettyprint.exp: print ns2
> FAIL: gdb.python/py-prettyprint.exp: print x
> FAIL: gdb.python/py-prettyprint.exp: print cstring
> FAIL: gdb.python/py-prettyprint.exp: print estring
> FAIL: gdb.python/py-prettyprint.exp: print estring2
> FAIL: gdb.python/py-prettyprint.exp: print hint_error
> FAIL: gdb.python/py-prettyprint.exp: print c
> FAIL: gdb.python/py-prettyprint.exp: print nstype
> FAIL: gdb.python/py-prettyprint.exp: print nstype on one line
> ERROR: remote_download to host of ./gdb.python/py-prettyprint.py to 
> /home/rurban/Software/gdb-7.6.90/gdb/testsuite/gdb.python/py-prettyprint.py: 
> cp: cannot stat './gdb.python/py-prettyprint.py': No such file or directory
> FAIL: gdb.python/py-prettyprint.exp: info locals
> FAIL: gdb.python/py-prettyprint.exp: print ss enabled #1
> FAIL: gdb.python/py-prettyprint.exp: python disable_lookup_function ()
> FAIL: gdb.python/py-prettyprint.exp: python enable_lookup_function ()
> FAIL: gdb.python/py-prettyprint.exp: print ss enabled #2

I can reproduce most of these fails, quoted above, (on amd64 ubuntu
machine) in an in-tree build and test.  Looks we need to tweak
testsuite/lib/*.exp files to support in-tree build well.

-- 
Yao (齐尧)

Re: GDB 7.6.90 available for testing

[Joel Brobecker]

> Joel, I did notice this -- not a regression, but a new test that fails:
> 
> gdb.ada/mi_interface.exp: Create ggg1 varobj: FAIL
> gdb.ada/mi_interface.exp: list ggg1's children: FAIL
> 
> 
> >From gdb.log:
> 
> -var-create ggg1 * ggg1
> ^error,msg="-var-create: unable to create variable object"
> (gdb) 
> 
> I didn't dig any deeper though.

My guess is that it's related to the compiler producing some
invalid debugging info. I will try to reproduce on my VM as
soon as I have a chance.

-- 
Joel

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> From: Joel Brobecker <brobecker@adacore.com>
> Date: Wed,  8 Jan 2014 14:14:28 +0400 (RET)
> 
> I have just finished creating the gdb-7.6.90 pre-release.
> It is available for download at the following location:
> 
>     ftp://sourceware.org/pub/gdb/snapshots/branch/gdb-7.6.90.tar.bz2
> 
> A gzip'ed version is also available: gdb-7.6.90.tar.gz.
> 
> Please give it a test if you can and report any problems you might find.

The gdbserver part fails to build on MinGW:

  gcc -O2 -gdwarf-2 -g3 -D__USE_MINGW_ACCESS     -I. -I. -I./../common  -I./../regformats -I./../ -I./../../include  -I./../gnulib/import -Ibuild-gnulib-gdbserver/import -Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral-Wno-char-subscripts -Werror -DGDBSERVER -c -o agent.o -MT agent.o -MMD -MP  -MF .deps/agent.Tpo ../common/agent.c
  In file included from ./server.h:98,
		   from ../common/agent.c:21:
  ./target.h:24:27: target/resume.h: No such file or directory
  ./target.h:25:25: target/wait.h: No such file or directory
  ./target.h:26:31: target/waitstatus.h: No such file or directory
  In file included from ./server.h:98,
		   from ../common/agent.c:21:
  ./target.h:43: error: field `kind' has incomplete type
  ./target.h:120: warning: "struct target_waitstatus" declared inside parameter list
  ./target.h:120: warning: its scope is only this definition or declaration, which is probably not what you want
  ./target.h:512: warning: "struct target_waitstatus" declared inside parameter list
  In file included from ./server.h:100,
		   from ../common/agent.c:21:
  ./gdbthread.h:34: error: field `last_resume_kind' has incomplete type
  ./gdbthread.h:37: error: field `last_status' has incomplete type
  ../common/agent.c: In function `agent_run_command':
  ../common/agent.c:240: error: `resume_continue' undeclared (first use in this function)
  ../common/agent.c:240: error: (Each undeclared identifier is reported only once
  ../common/agent.c:240: error: for each function it appears in.)
  ../common/agent.c:278: error: storage size of 'status' isn't known
  ../common/agent.c:287: error: `resume_stop' undeclared (first use in this function)
  ../common/agent.c:278: warning: unused variable `status'
  Makefile:516: recipe for target `agent.o' failed
  make[4]: *** [agent.o] Error 1
  make[4]: Leaving directory `/d/gnu/gdb-7.6.90/gdb/gdbserver'
  Makefile:1323: recipe for target `subdir_do' failed
  make[3]: *** [subdir_do] Error 1

This is because of the "-I./../" part on the GCC command line.  My
version of GCC doesn't like the trailing slash.

That slash comes from this snippet in gdbserver/Makefile.in:

  INCLUDE_CFLAGS = -I. -I${srcdir} -I$(srcdir)/../common \
	  -I$(srcdir)/../regformats -I$(srcdir)/../ -I$(INCLUDE_DIR) \
	  $(INCGNU)

If I remove the trailing slash there, the build runs to completion.

OK to push the following (with a suitable log entry)?

--- gdb/gdbserver/Makefile.in~0	2014-01-08 11:23:36.000000000 +0200
+++ gdb/gdbserver/Makefile.in	2014-01-11 10:49:53.774500000 +0200
@@ -106,7 +106,7 @@
 # e.g.: "target/wait.h".
 #
 INCLUDE_CFLAGS = -I. -I${srcdir} -I$(srcdir)/../common \
-	-I$(srcdir)/../regformats -I$(srcdir)/../ -I$(INCLUDE_DIR) \
+	-I$(srcdir)/../regformats -I$(srcdir)/.. -I$(INCLUDE_DIR) \
 	$(INCGNU)
 
 # M{H,T}_CFLAGS, if defined, has host- and target-dependent CFLAGS

Re: GDB 7.6.90 available for testing

[Ricard Wanderlof]

On Wed, 8 Jan 2014, Joel Brobecker wrote:

> Hello,
>
> I have just finished creating the gdb-7.6.90 pre-release.
> It is available for download at the following location:
>
>    ftp://sourceware.org/pub/gdb/snapshots/branch/gdb-7.6.90.tar.bz2
>
> A gzip'ed version is also available: gdb-7.6.90.tar.gz.

Is there a corresponding git tag one could use for checking out?

/Ricard
- 
Ricard Wolf Wanderlöf                           ricardw(at)axis.com
Axis Communications AB, Lund, Sweden            www.axis.com
Phone +46 46 272 2016                           Fax +46 46 13 61 30

Re: GDB 7.6.90 available for testing

[Joel Brobecker]

> >A gzip'ed version is also available: gdb-7.6.90.tar.gz.
> 
> Is there a corresponding git tag one could use for checking out?

Not for pre-releases. Only releases get a tag.
The SHA1 you're looking for is 05baa5c341872225b02c0b370724ac0d043f3e0e.

But if you're going to do some testing from git, I think it'd be
more useful if you used the tip of the gdb-7.7 branch instead.
That way, you'll be testing the latest fixes as well.

-- 
Joel

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

Ping!

> Date: Sat, 11 Jan 2014 10:55:45 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: gdb-patches@sourceware.org
> 
> > From: Joel Brobecker <brobecker@adacore.com>
> > Date: Wed,  8 Jan 2014 14:14:28 +0400 (RET)
> > 
> > I have just finished creating the gdb-7.6.90 pre-release.
> > It is available for download at the following location:
> > 
> >     ftp://sourceware.org/pub/gdb/snapshots/branch/gdb-7.6.90.tar.bz2
> > 
> > A gzip'ed version is also available: gdb-7.6.90.tar.gz.
> > 
> > Please give it a test if you can and report any problems you might find.
> 
> The gdbserver part fails to build on MinGW:
> 
>   gcc -O2 -gdwarf-2 -g3 -D__USE_MINGW_ACCESS     -I. -I. -I./../common  -I./../regformats -I./../ -I./../../include  -I./../gnulib/import -Ibuild-gnulib-gdbserver/import -Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral-Wno-char-subscripts -Werror -DGDBSERVER -c -o agent.o -MT agent.o -MMD -MP  -MF .deps/agent.Tpo ../common/agent.c
>   In file included from ./server.h:98,
> 		   from ../common/agent.c:21:
>   ./target.h:24:27: target/resume.h: No such file or directory
>   ./target.h:25:25: target/wait.h: No such file or directory
>   ./target.h:26:31: target/waitstatus.h: No such file or directory
>   In file included from ./server.h:98,
> 		   from ../common/agent.c:21:
>   ./target.h:43: error: field `kind' has incomplete type
>   ./target.h:120: warning: "struct target_waitstatus" declared inside parameter list
>   ./target.h:120: warning: its scope is only this definition or declaration, which is probably not what you want
>   ./target.h:512: warning: "struct target_waitstatus" declared inside parameter list
>   In file included from ./server.h:100,
> 		   from ../common/agent.c:21:
>   ./gdbthread.h:34: error: field `last_resume_kind' has incomplete type
>   ./gdbthread.h:37: error: field `last_status' has incomplete type
>   ../common/agent.c: In function `agent_run_command':
>   ../common/agent.c:240: error: `resume_continue' undeclared (first use in this function)
>   ../common/agent.c:240: error: (Each undeclared identifier is reported only once
>   ../common/agent.c:240: error: for each function it appears in.)
>   ../common/agent.c:278: error: storage size of 'status' isn't known
>   ../common/agent.c:287: error: `resume_stop' undeclared (first use in this function)
>   ../common/agent.c:278: warning: unused variable `status'
>   Makefile:516: recipe for target `agent.o' failed
>   make[4]: *** [agent.o] Error 1
>   make[4]: Leaving directory `/d/gnu/gdb-7.6.90/gdb/gdbserver'
>   Makefile:1323: recipe for target `subdir_do' failed
>   make[3]: *** [subdir_do] Error 1
> 
> This is because of the "-I./../" part on the GCC command line.  My
> version of GCC doesn't like the trailing slash.
> 
> That slash comes from this snippet in gdbserver/Makefile.in:
> 
>   INCLUDE_CFLAGS = -I. -I${srcdir} -I$(srcdir)/../common \
> 	  -I$(srcdir)/../regformats -I$(srcdir)/../ -I$(INCLUDE_DIR) \
> 	  $(INCGNU)
> 
> If I remove the trailing slash there, the build runs to completion.
> 
> OK to push the following (with a suitable log entry)?
> 
> --- gdb/gdbserver/Makefile.in~0	2014-01-08 11:23:36.000000000 +0200
> +++ gdb/gdbserver/Makefile.in	2014-01-11 10:49:53.774500000 +0200
> @@ -106,7 +106,7 @@
>  # e.g.: "target/wait.h".
>  #
>  INCLUDE_CFLAGS = -I. -I${srcdir} -I$(srcdir)/../common \
> -	-I$(srcdir)/../regformats -I$(srcdir)/../ -I$(INCLUDE_DIR) \
> +	-I$(srcdir)/../regformats -I$(srcdir)/.. -I$(INCLUDE_DIR) \
>  	$(INCGNU)
>  
>  # M{H,T}_CFLAGS, if defined, has host- and target-dependent CFLAGS
> 

Re: GDB 7.6.90 available for testing

[Pedro Alves]

On 01/11/2014 08:55 AM, Eli Zaretskii wrote:

> 
> This is because of the "-I./../" part on the GCC command line.  My
> version of GCC doesn't like the trailing slash.

Guessing that's an old gcc.  This seems to have been fixed in 2007:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22133

> That slash comes from this snippet in gdbserver/Makefile.in:
> 
>   INCLUDE_CFLAGS = -I. -I${srcdir} -I$(srcdir)/../common \
> 	  -I$(srcdir)/../regformats -I$(srcdir)/../ -I$(INCLUDE_DIR) \
> 	  $(INCGNU)
> 
> If I remove the trailing slash there, the build runs to completion.
> 
> OK to push the following (with a suitable log entry)?

Sure.

Thanks,
-- 
Pedro Alves

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> Date: Mon, 13 Jan 2014 17:51:32 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: Joel Brobecker <brobecker@adacore.com>, gdb-patches@sourceware.org
> 
> On 01/11/2014 08:55 AM, Eli Zaretskii wrote:
> 
> > 
> > This is because of the "-I./../" part on the GCC command line.  My
> > version of GCC doesn't like the trailing slash.
> 
> Guessing that's an old gcc.

Your guess is correct.  (I also tried a newer GCC, and the problem
never happened there.)  But I think that slash is the odd one out
anyway: the other -I arguments don't have it.

> > OK to push the following (with a suitable log entry)?
> 
> Sure.

Thanks.

As this will be my first push when we are branched, what are the
procedures for that with git?  I'm guessing

  git checkout gdb-7.7-branch
  (hack, hack)
  git commit
  git push

Is that correct?  If so, what about doing the same on master -- should
I merge it myself and then push, or are these merges handled in some
other way by someone else?

TIA

Re: GDB 7.6.90 available for testing

[Pedro Alves]

On 01/13/2014 06:27 PM, Eli Zaretskii wrote:
>> Date: Mon, 13 Jan 2014 17:51:32 +0000
>> From: Pedro Alves <palves@redhat.com>
>> CC: Joel Brobecker <brobecker@adacore.com>, gdb-patches@sourceware.org
>>
>> On 01/11/2014 08:55 AM, Eli Zaretskii wrote:
>>
>>>
>>> This is because of the "-I./../" part on the GCC command line.  My
>>> version of GCC doesn't like the trailing slash.
>>
>> Guessing that's an old gcc.
> 
> Your guess is correct.  (I also tried a newer GCC, and the problem
> never happened there.)  But I think that slash is the odd one out
> anyway: the other -I arguments don't have it.
> 
>>> OK to push the following (with a suitable log entry)?
>>
>> Sure.
> 
> Thanks.
> 
> As this will be my first push when we are branched, what are the
> procedures for that with git?  I'm guessing
> 
>   git checkout gdb-7.7-branch
>   (hack, hack)
>   git commit
>   git push

Yeah.

As a matter of principle, I prefer fixing master first though, and
then backport to any stable branches.  So that there's never a chance
of the branch getting a fix that's not in mainline (and so the
same bug could reappear in mainline later in the next release).
(You never know when you might be interrupted, and then forget to
merge a change.)

> Is that correct?  If so, what about doing the same on master -- should
> I merge it myself and then push, or are these merges handled in some
> other way by someone else?

You have to do it yourself.  I use stgit, so I just rebase my series
on top of master or whatever stable branch, and then push, but with
raw git, I'd use git cherry-pick.

Thanks,
-- 
Pedro Alves

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> Date: Mon, 13 Jan 2014 18:44:24 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: brobecker@adacore.com, gdb-patches@sourceware.org
> 
> As a matter of principle, I prefer fixing master first though, and
> then backport to any stable branches.

OK, will do.

> You have to do it yourself.  I use stgit, so I just rebase my series
> on top of master or whatever stable branch, and then push, but with
> raw git, I'd use git cherry-pick.

Is there any difference between merge and cherry-pick in this case?

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> Date: Mon, 13 Jan 2014 20:50:28 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: brobecker@adacore.com, gdb-patches@sourceware.org
> 
> Is there any difference between merge and cherry-pick in this case?

It seems msysgit's cherry-pick somehow converted gdbserver/ChangeLog
to DOS CRLF EOL format.  So I needed one additional commit on the
branch to fix that, before pushing.  Is that OK, or should I have done
something different?

Re: GDB 7.6.90 available for testing

[Pedro Alves]

On 01/13/2014 07:15 PM, Eli Zaretskii wrote:
>> Date: Mon, 13 Jan 2014 20:50:28 +0200
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: brobecker@adacore.com, gdb-patches@sourceware.org
>>
>> Is there any difference between merge and cherry-pick in this case?

merge would attempt to merge the whole of master into the branch
starting from the branching point.

> It seems msysgit's cherry-pick somehow converted gdbserver/ChangeLog
> to DOS CRLF EOL format.  So I needed one additional commit on the
> branch to fix that, before pushing.  Is that OK, or should I have done
> something different?

Ouch, sounds like a git bug...

Yeah, it would have been better if that broken EOL change
hadn't made it into the repository.  E.g., now "git blame"
on gdbserver's ChangeLog will show that commit as author
of all of the file's lines.  Luckily, as it was on the branch,
not master, it probably won't ever affect anyone.  You could,
for example, have squashed those two commits into one (with git
rebase) before pushing, ending up with a single commit that
looked just like the one in master.  Oh well, what's done is
done now.  ;-)

-- 
Pedro Alves

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> Date: Mon, 13 Jan 2014 19:53:29 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: brobecker@adacore.com, gdb-patches@sourceware.org
> 
> On 01/13/2014 07:15 PM, Eli Zaretskii wrote:
> >> Date: Mon, 13 Jan 2014 20:50:28 +0200
> >> From: Eli Zaretskii <eliz@gnu.org>
> >> Cc: brobecker@adacore.com, gdb-patches@sourceware.org
> >>
> >> Is there any difference between merge and cherry-pick in this case?
> 
> merge would attempt to merge the whole of master into the branch
> starting from the branching point.

I meant the other way around: from the branch to master.

Re: GDB 7.6.90 available for testing

[Eli Zaretskii]

> Date: Mon, 13 Jan 2014 19:53:29 +0000
> From: Pedro Alves <palves@redhat.com>
> CC: brobecker@adacore.com, gdb-patches@sourceware.org
> 
> > It seems msysgit's cherry-pick somehow converted gdbserver/ChangeLog
> > to DOS CRLF EOL format.  So I needed one additional commit on the
> > branch to fix that, before pushing.  Is that OK, or should I have done
> > something different?
> 
> Ouch, sounds like a git bug...

For the record: it's not a git bug.  It's a bug in git-merge-changelog
when it is built on Windows: it uses text-mode I/O.  It didn't come
with git, I built it myself.

I guess cherry-pick of a ChangeLog counts as a merge operation, so git
invoked git-merge-changelog, which is installed here as a merge driver
for ChangeLog files.

So now I need to fix the Windows build of git-merge-changelog to avoid
this problem in the future.