[Bug gcov-profile/61889] New: [4.10 Regression] gcov-tool.c uses nftw, ftw.h

[Bug gcov-profile/61889] New: [4.10 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

            Bug ID: 61889
           Summary: [4.10 Regression] gcov-tool.c uses nftw, ftw.h
           Product: gcc
           Version: 4.10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: gcov-profile
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rainer@emrich-ebersheim.de

Bootstrap fails:

g++ -c   -g  -DIN_GCC    -fno-exceptions -fno-rtti -fasynchronous-unwind-tables
-W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wno-format
-Wmissing-format-attribute -Woverloaded-virtual -pedantic -Wno-long-long
-Wno-variadic-macr  os -Wno-overlength-strings   -DHAVE_CONFIG_H -I. -I.
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/.
-I../../../../../../../opt/devel/gnu/src 
/gcc-mingw-w64/gcc-4.10.0/gcc/../include -I./../intl
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/../libcpp/include
-I/opt/devel/SCRATCH/tmp.s61dWEwM1g/install/include
-I/opt/devel/SCRATCH/tmp.s61dWEwM1g/instal  l/include
-I/opt/devel/SCRATCH/tmp.s61dWEwM1g/install/include 
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/../libdecnumber
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/../libdecnumber/
 bid -I../libdecnumber
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/../libbacktrace
-DCLOOG_INT_GMP -I/opt/devel/SCRATCH/tmp.s61dWEwM1g/install/include
-I/opt/devel/SCRATCH/tmp.s61dWEwM1g/install/include  -o gco  v-tool.o -MT
gcov-tool.o -MMD -MP -MF ./.deps/gcov-tool.TPo
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/gcov-tool.c
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-4.10.0/gcc/gcov-tool.c:38:17:
fatal error: ftw.h: No such file or directory
 #include <ftw.h>
                 ^
compilation terminated.
Makefile:1063: recipe for target 'gcov-tool.o' failed
make: *** [gcov-tool.o] Error 1

This is not very portable and breaks bootstrap for mingw targets sine nearly
two weeks.

[Bug gcov-profile/61889] [4.10 Regression] gcov-tool.c uses nftw, ftw.h

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |build
           Priority|P3                          |P1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-07-23
                 CC|                            |xinliangli at gmail dot com
   Target Milestone|---                         |4.10.0
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  Please fix (for example by providing nftw from libiberty).

[Bug gcov-profile/61889] [4.10 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ktietz at gcc dot gnu.org

--- Comment #2 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Yes, issue is here that ftw API is used for cleanup.

...
static int
unlink_profile_dir (const char *path)
{
    return nftw(path, unlink_gcda_file, 64, FTW_DEPTH | FTW_PHYS);
}
...

This method isn't present on none glibc-systems.  So there should be at least a
header-check (HAVE_FTW_H) be added to gcov-tool.c and to configure.

As Richard mentioned, libiberty would be a good place to add implementation.

[Bug gcov-profile/61889] [4.10 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #3 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 23.07.2014 17:17, schrieb ktietz at gcc dot gnu.org:
> --- Comment #2 from Kai Tietz <ktietz at gcc dot gnu.org> --- Yes, issue is
> here that ftw API is used for cleanup.
> 
> ... static int unlink_profile_dir (const char *path) { return nftw(path,
> unlink_gcda_file, 64, FTW_DEPTH | FTW_PHYS); } ...
> 
> This method isn't present on none glibc-systems.  So there should be at
> least a header-check (HAVE_FTW_H) be added to gcov-tool.c and to
> configure.
> 
> As Richard mentioned, libiberty would be a good place to add
> implementation.
> 
It's still breaking bootstrap on mingw targets, more than 3 weeks now!

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJT3qPrAAoJEB3HOsWs+KJbzMgH/0kciCmfQu+Bqc7drVDPmabH
xgjF2RKEMXg8XhueSpq5MADkh8cUvGv0skExtknwiEgr84TxK/SwinmDvCADUVWY
AXQuODgDnrRS8VNugyckfVZ11wm/uK2ywfp8zlRyNreR54OezOXtXWAtJ0Z8wDc6
0GUvJqK+SP6RbKJLNrRDiowN8bwq+cGK/TAdgIRq7bztY69nzkO604MY8zDQsbaS
ykPv+uwF9uCrmDuKA/L0uYBgd+E2t8WkMi4AbYJebqaiegLrYj16oEPYaJem1j39
uIqIpP/p6ylkQPmclSePOd7TbKTVi++YP3N3czNTXjvogYc20fyDoUl5/9cKCSs=
=IM/a
-----END PGP SIGNATURE-----

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING
      Known to fail|4.10.0                      |5.0

--- Comment #4 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Issue got fixed on mingw-w64's side by providing on trunk ftw/nftw API.
Nevertheless it would be good that for ftw include and use at least a
header-check is done.  That gcov data isn't deleted isn't necessarily that bad.
 IMO better then disabling complete tool.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #6 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Yes, I remember.  I didn't comment on it.
The following checks aren't ok.
'#if !defined(_WIN32)'

you should disable those parts *only* if ftw API isn't present. This you should
check by a HAVE_FTW_H configure-check-macro.  To disable it just because _WIN32
is defined is wrong.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[xur at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #7 from xur at google dot com ---
OK. I'll fix this and submit another patch.

On Wed, Aug 20, 2014 at 11:26 AM, ktietz at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org> wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
>
> --- Comment #6 from Kai Tietz <ktietz at gcc dot gnu.org> ---
> Yes, I remember.  I didn't comment on it.
> The following checks aren't ok.
> '#if !defined(_WIN32)'
>
> you should disable those parts *only* if ftw API isn't present. This you should
> check by a HAVE_FTW_H configure-check-macro.  To disable it just because _WIN32
> is defined is wrong.
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #8 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
(In reply to xur from comment #7)
> OK. I'll fix this and submit another patch.
What is the status for that?

> 
> On Wed, Aug 20, 2014 at 11:26 AM, ktietz at gcc dot gnu.org
> <gcc-bugzilla@gcc.gnu.org> wrote:
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
> >
> > --- Comment #6 from Kai Tietz <ktietz at gcc dot gnu.org> ---
> > Yes, I remember.  I didn't comment on it.
> > The following checks aren't ok.
> > '#if !defined(_WIN32)'
> >
> > you should disable those parts *only* if ftw API isn't present. This you should
> > check by a HAVE_FTW_H configure-check-macro.  To disable it just because _WIN32
> > is defined is wrong.
> >
Quoting Kai Tietz:
Issue got fixed on mingw-w64's side by providing on trunk ftw/nftw API.

Nevertheless there is still an issue with the use of "mkdir":
g++ -c   -g -DIN_GCC    -fno-exceptions -fno-rtti -fasynchronous-unwind-tables
-W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wno-format
-Wmissing-format-attribute -Woverloaded-virtual -pedantic -Wno-long-long
-Wno-variadic-macros -Wno-overlength-strings   -DHAVE_CONFIG_H -I. -I.
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/.
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/../include
-I./../intl
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/../libcpp/include
-I/opt/devel/SCRATCH/tmp.Yed3qsDdVp/install/include
-I/opt/devel/SCRATCH/tmp.Yed3qsDdVp/install/include
-I/opt/devel/SCRATCH/tmp.Yed3qsDdVp/install/include 
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/../libdecnumber
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/../libdecnumber/bid
-I../libdecnumber
-I../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/../libbacktrace
-DCLOOG_INT_GMP -I/opt/devel/SCRATCH/tmp.Yed3qsDdVp/install/include
-DCLOOG_INT_GMP -I/opt/devel/SCRATCH/tmp.Yed3qsDdVp/install/include  -o
gcov-tool.o -MT gcov-tool.o -MMD -MP -MF ./.deps/gcov-tool.TPo
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/gcov-tool.c
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/gcov-tool.c:95:21:
error: macro "mkdir" requires 2 arguments, but only 1 given
       if (mkdir (out) == -1 && errno != EEXIST)
                     ^
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/gcov-tool.c:
In function 'void gcov_output_files(const char*, gcov_info*)':
../../../../../../../opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/gcov-tool.c:95:27:
error: ISO C++ forbids comparison between pointer and integer [-fpermissive]
       if (mkdir (out) == -1 && errno != EEXIST)
                           ^

This clashes with the macro definition for mkdir in system.h:

/* Some systems have mkdir that takes a single argument.  */
#ifdef MKDIR_TAKES_ONE_ARG
# define mkdir(a,b) mkdir (a)
#endif 

Removing the wrong #if !defined(_WIN32) solves the issue for gcov-tool.c:

Index: gcc/gcov-tool.c
===================================================================
--- gcc/gcov-tool.c    (Revision 215554)
+++ gcc/gcov-tool.c    (Arbeitskopie)
@@ -89,11 +89,7 @@ gcov_output_files (const char *out, stru
   /* Try to make directory if it doesn't already exist.  */
   if (access (out, F_OK) == -1)
     {
-#if !defined(_WIN32)
       if (mkdir (out, S_IRWXU | S_IRWXG | S_IRWXO) == -1 && errno != EEXIST)
-#else
-      if (mkdir (out) == -1 && errno != EEXIST)
-#endif
         fatal_error ("Cannot make directory %s", out);
     } else
       unlink_profile_dir (out); 

At a second point the issue is more complex. In libgcc/libgcov-driver-system.c
there is the following code:

#ifdef TARGET_POSIX_IO
            && mkdir (filename, 0755) == -1
#else
            && mkdir (filename) == -1
#endif

libgcov-driver-system.c is used in two places, in the gcc directory and in the
the libgcc directory for libgcov. In the first case the macro for mkdir is
defined in the second it isn't.

An ugly hack solves this issue and lets me at least build gcc-5.0.0 on
x86_64-w64-mingw32.

Index: libgcc/libgcov-driver-system.c
===================================================================
--- libgcc/libgcov-driver-system.c    (Revision 215554)
+++ libgcc/libgcov-driver-system.c    (Arbeitskopie)
@@ -66,6 +66,9 @@ create_file_directory (char *filename)
 #ifdef TARGET_POSIX_IO
             && mkdir (filename, 0755) == -1
 #else
+#ifdef mkdir
+#undef mkdir
+#endif
             && mkdir (filename) == -1
 #endif
             /* The directory might have been made by another process.  */ 


Someone with more insight should have a look on this, please!

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #9 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Created attachment 33548
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33548&action=edit
Proposed patch to fix the mingw case.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fxcoudert at gcc dot gnu.org

--- Comment #10 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
Why is this 2.5 month old bug WAITING? Looks like a clear issue to me.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #11 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Dear friends this issue seems to become a never ending story.
In my understanding the person causing the issue is responsible for a fix.
There are several hints in this thread how to solve the issue. So please get on
to it.
Otherwise we have a massive issue here.
And last but not least I don't see any reason why this bug has status WAITING.
If there is a solid reason please post it here.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[StaffLeavers at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #12 from StaffLeavers at arm dot com ---
tony.wang no longer works for ARM.

Your email will be forwarded to their line manager.


Please do not reply to this email.
If you need more information, please email real-postmaster@arm.com

Thank you.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[StaffLeavers at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #13 from StaffLeavers at arm dot com ---
tony.wang no longer works for ARM.

Your email will be forwarded to their line manager.


Please do not reply to this email.
If you need more information, please email real-postmaster@arm.com

Thank you.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
                 CC|tony.wang at arm dot com           |

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[StaffLeavers at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #14 from StaffLeavers at arm dot com ---
tony.wang no longer works for ARM.

Your email will be forwarded to their line manager.


Please do not reply to this email.
If you need more information, please email real-postmaster@arm.com

Thank you.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[StaffLeavers at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #15 from StaffLeavers at arm dot com ---
tony.wang no longer works for ARM.

Your email will be forwarded to their line manager.


Please do not reply to this email.
If you need more information, please email real-postmaster@arm.com

Thank you.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[xur at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #16 from xur at google dot com ---
I sent a patch to fix this, a few weeks ago, but I have got the review
or approval.

https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00186.html

Honza, could you take a quick look?


-Rong

On Mon, Oct 13, 2014 at 12:29 PM, rainer@emrich-ebersheim.de
<gcc-bugzilla@gcc.gnu.org> wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
>
> --- Comment #11 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
> Dear friends this issue seems to become a never ending story.
> In my understanding the person causing the issue is responsible for a fix.
> There are several hints in this thread how to solve the issue. So please get on
> to it.
> Otherwise we have a massive issue here.
> And last but not least I don't see any reason why this bug has status WAITING.
> If there is a solid reason please post it here.
>
> --
> You are receiving this mail because:
> You are the assignee for the bug.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #17 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
(In reply to xur from comment #16)
> I sent a patch to fix this, a few weeks ago, but I have got the review
> or approval.
> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00186.html

Kai Tietz, mingw maintainer, commented on it (comment #6 above) and it is not
OK as is (support should be checked, not hard-coded).

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[xur at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #18 from xur at google dot com ---
This patch is after Kai Tietz's comment. and it does check the nfw headers.

On Mon, Oct 13, 2014 at 2:40 PM, fxcoudert at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org> wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
>
> --- Comment #17 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
> (In reply to xur from comment #16)
>> I sent a patch to fix this, a few weeks ago, but I have got the review
>> or approval.
>> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00186.html
>
> Kai Tietz, mingw maintainer, commented on it (comment #6 above) and it is not
> OK as is (support should be checked, not hard-coded).
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #19 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Hi Xur,

I asked you in my intial support to check for existance of FTW-API, and not to
implement it for Win32.

So first, send patch checking in a valid way if API can be used.

The ftw/nftw emulation you wrote seems to me more suitable for libiberty.  And
again in most places the check for _WIN32 isn't right.  You should check
instead for mingw-target, means for __MINGW32__ instead, as for cygwin this
macro might be defined in some circumstances.  And make sure that you disable
code-paths only for mingw-targets iff we don't have the FTW-API.

I would suggest to make out this patch 2 separate patches.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[xur at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #20 from xur at google dot com ---
Thanks for the comments. I'll work on this to get it fixed this time.

Let me understand your idea correctly:
We will have two patches: The first one will check FTW-API and make
the gcov-tool build configurable.
if -disable-gcov-tool is specified, we will not build gcov-tool.
if -enable-gcov-tool is specified, we will build gcov-tool
if neither specified, we will check the FTW-API and build gcovtool if
FTW-API is available.

The second patch is to emulate FTW in libiberty for MINGW32?
I'm a little confused here. libiberty is built after the configure. Do
we need to a special handling of MINGW32 in config?

Thanks,

-Rong

On Mon, Oct 13, 2014 at 3:03 PM, ktietz at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org> wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
>
> --- Comment #19 from Kai Tietz <ktietz at gcc dot gnu.org> ---
> Hi Xur,
>
> I asked you in my intial support to check for existance of FTW-API, and not to
> implement it for Win32.
>
> So first, send patch checking in a valid way if API can be used.
>
> The ftw/nftw emulation you wrote seems to me more suitable for libiberty.  And
> again in most places the check for _WIN32 isn't right.  You should check
> instead for mingw-target, means for __MINGW32__ instead, as for cygwin this
> macro might be defined in some circumstances.  And make sure that you disable
> code-paths only for mingw-targets iff we don't have the FTW-API.
>
> I would suggest to make out this patch 2 separate patches.
>
> --
> You are receiving this mail because:
> You are the assignee for the bug.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #21 from Kai Tietz <ktietz at gcc dot gnu.org> ---
(In reply to xur from comment #20)
> Thanks for the comments. I'll work on this to get it fixed this time.
> 
> Let me understand your idea correctly:
> We will have two patches: The first one will check FTW-API and make
> the gcov-tool build configurable.
> if -disable-gcov-tool is specified, we will not build gcov-tool.
> if -enable-gcov-tool is specified, we will build gcov-tool
> if neither specified, we will check the FTW-API and build gcovtool if
> FTW-API is available.

No, we always check if FTW-API is present.  The presence of FTW-API should have
no impact on case that gcov-tool gets built, or not.
Instead gcov-tool should take care that for cases where FTW-API isn't used the
specific part of functionality (it is used for file-unlinking) is simply
disabled.
Otherwise the standard code-path should be used.
Btw, as I already mentioned has mingw-w64 flavor the ftw/nftw API on its master
version.  And it works OOTB with current code.  So no need to break that by
assuming wild things in configure.

> The second patch is to emulate FTW in libiberty for MINGW32?
> I'm a little confused here. libiberty is built after the configure. Do
> we need to a special handling of MINGW32 in config?
Well, if we put that into libiberty, or put it in gcov-tool directly - I would
assume that you alway have dependency to libiberty, otherwise you should add it
for this tool, as it prevents you from reinventing the wheel again.
Anyway this part is something different and should be handled via the ML and
not via the bug-tracker.


> Thanks,
> 
> -Rong

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #22 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, can you just check for nftw with the arguments you want to use in a compile
test in configure and if that fail, perhaps for now don't do that at all?
The next step would be to provide in libiberty a nftw implementation for some
or all targets that don't have it.  mingw surely isn't the only problematic
target.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[tbsaunde at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #23 from tbsaunde at gcc dot gnu.org ---
Author: tbsaunde
Date: Tue Feb 10 03:40:20 2015
New Revision: 220566

URL: https://gcc.gnu.org/viewcvs?rev=220566&root=gcc&view=rev
Log:
Support gcov-tool without ftw.h

    gcc/

    PR gcov-profile/61889
    * config.in: regenerate.
    * configure.in: Likewise.
    * configure.ac: Check for ftw.h.
    * gcov-tool.c: Check for ftw.h before using nftw.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config.in
    trunk/gcc/configure
    trunk/gcc/configure.ac
    trunk/gcc/gcov-tool.c

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #26 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Instead of the #undef mkdir you'd IMHO better just use (mkdir) (filename)
in the second case.
Anyway, if you've posted your patch to gcc-patches, you should be pinging it
until it is reviewed.  Kai, can you please have a look?

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #27 from Kai Tietz <ktietz at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #26)
> Instead of the #undef mkdir you'd IMHO better just use (mkdir) (filename)
> in the second case.
> Anyway, if you've posted your patch to gcc-patches, you should be pinging it
> until it is reviewed.  Kai, can you please have a look?

Sure, I will have a look.  Was this patch sent to ML?  If so I am sorry for not
noticing it.  Could you please ping it?

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #28 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
http://gcc.gnu.org/ml/gcc-patches/2014-09/msg02137.html
?

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #29 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Am 10.02.2015 12:12, schrieb jakub at gcc dot gnu.org:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889
> 
> --- Comment #28 from Jakub Jelinek <jakub at gcc dot gnu.org> --- 
> http://gcc.gnu.org/ml/gcc-patches/2014-09/msg02137.html ?
> 
Yes, that's it.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #30 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Yes, this patch slipped under my radar.  It would be good if you - Rainer -
would have pinged on it.  As far as I recalled I awaited at that time a full
patch by Rong on this subject. (See comment #16 and after that).

Nevertheless patch looks ok.  I will give it some testing and will apply it
after that.  The change to undef mkdir in libgcc's part of the patch looks to
me being the real patch.  The patch in gcc's part is more cosmetic part getting
rid of this (also mentioned to Rong too) wrong _WIN32 check.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #31 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
(In reply to Kai Tietz from comment #30)
> Yes, this patch slipped under my radar.  It would be good if you - Rainer -
> would have pinged on it.  As far as I recalled I awaited at that time a full
> patch by Rong on this subject. (See comment #16 and after that).
Sorry, I had not much time the last few month.

> Nevertheless patch looks ok.  I will give it some testing and will apply it
> after that.  The change to undef mkdir in libgcc's part of the patch looks
> to me being the real patch.  The patch in gcc's part is more cosmetic part
> getting rid of this (also mentioned to Rong too) wrong _WIN32 check.
This part is necessary because otherwise the definition of the mkdir macro in
system.h expects 2 arguments what isn't satisfied.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #32 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Author: ktietz
Date: Tue Feb 10 14:13:13 2015
New Revision: 220582

URL: https://gcc.gnu.org/viewcvs?rev=220582&root=gcc&view=rev
Log:
2015-02-10  Rainer Emrich  <rainer@emrich-ebersheim.de>

    PR gcov-profile/61889
    * libgcc/libgcov-driver-system.c: undefine clashing macro for mkdir.


Modified:
    trunk/libgcc/ChangeLog
    trunk/libgcc/libgcov-driver-system.c

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #33 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Author: ktietz
Date: Tue Feb 10 14:14:58 2015
New Revision: 220584

URL: https://gcc.gnu.org/viewcvs?rev=220584&root=gcc&view=rev
Log:
2015-02-10  Rainer Emrich  <rainer@emrich-ebersheim.de>

    PR gcov-profile/61889
    * gcov-tool.c: Remove wrong #if !defined(_WIN32)


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/gcov-tool.c

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #34 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Fixed.

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[terry.guo at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

Terry Guo <terry.guo at arm dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |terry.guo at arm dot com

--- Comment #35 from Terry Guo <terry.guo at arm dot com> ---
I am building trunk for mingw and still having issue:

/home/build/work/GCC-5-0-build/src/gcc/gcc/../libgcc/libgcov-util.c:55:17:
fatal error: ftw.h: No such file or directory
compilation terminated.
make[1]: *** [libgcov-util.o] Error 1

Is this something we just missed?

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #37 from Kai Tietz <ktietz at gcc dot gnu.org> ---
I confirm that in libgcc we still have an issue ...
Could you please make a new report for libgcc's libgcov-util.c for it.

Thanks in advance

[Bug gcov-profile/61889] [5 Regression] gcov-tool.c uses nftw, ftw.h

[terry.guo at arm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61889

--- Comment #38 from Terry Guo <terry.guo at arm dot com> ---
(In reply to Kai Tietz from comment #37)
> I confirm that in libgcc we still have an issue ...
> Could you please make a new report for libgcc's libgcov-util.c for it.
> 
> Thanks in advance

Reported it at https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65038. Not sure we
can mark them as duplication.