Bug#850329: [pkg-wine-party] Bug#850329: wine creates files with binary filenames

[Vincent Lefevre]

Control: reassign -1 autoconf 2.69-10
Control: retitle -1 autoconf generates code that can execute binary data as a 
shell script
Control: tags -1 -moreinfo -unreproducible
Control: severity -1 important

since this breaks Subversion in particular, requiring manual clean-up
of the directory (though the real culprit of the mess is dash).

On 2017-01-06 18:34:47 +0100, Jens Reyer wrote:
> are the tests that you run available somewhere, so that I can try to
> reproduce this?

I could reproduce it after removing the .wine directory, though it is
not necessarily related to wine at all (perhaps just a coincidence),
as I've now found the cause of the problem, and wine is not involved.
In a MPFR working copy (trunk), I just typed:

./configure --host=i586-mingw32msvc --disable-shared 
--with-gmp=/usr/local/gmp-mingw32 --enable-assert=full --enable-thread-safe

/usr/local/gmp-mingw32 is where a mingw32 version of GMP had been
installed last year, with:

$ ./configure --host=i586-mingw32msvc --disable-shared 
--prefix=/usr/local/gmp-mingw32 CC="i586-mingw32msvc-gcc 
-D__USE_MINGW_ANSI_STDIO"
$ make
$ make check LOG_COMPILER=wine
$ make install

Now, back to the problem, the following file was created by the
configure command in the MPFR directory:

-rw-r--r--  1 vlefevre vlefevre   0 2017-01-07 02:03:23 
pX\016\005\v\005\340\a\001\v\001\002\033\030*\004\340\024\0200\@\020\002\004\001\004\260\005\004}\030\006\003\@\001

In the config.log file, I can see:

configure:4348: checking whether we are cross compiling
configure:4356: i586-mingw32msvc-gcc -D__USE_MINGW_ANSI_STDIO -o conftest.exe 
-m32 -O2 -pedantic -fomit-frame-pointer -mtune=pentium -march=pentium  
-I/usr/local/gmp-mingw32/include  -L/usr/local/gmp-mingw32/lib conftest.c  >&5
configure:4360: $? = 0
configure:4367: ./conftest.exe
./conftest.exe: 1: ./conftest.exe: MZ
: not found
./conftest.exe: 2: ./conftest.exe: 
B/57X
@0B/70B
B/81
B/92
: not found
./conftest.exe: 1: ./conftest.exe: 
.text
: not found
./conftest.exe: 1: ./conftest.exe: Syntax error: word unexpected (expecting ")")
./conftest.exe: 1: ./conftest.exe: .data00
.rdata: not found
configure:4371: $? = 2
configure:4386: result: yes
./conftest.exe: 1: ./conftest.exe: 
.idata
: not found
./conftest.exe: 2: ./conftest.exe: 
[: not found

So, this is a bug in autoconf, which tries to execute a Windows
program, which is then interpreted by dash as a shell script:

  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=816313

but autoconf must have workarounds for known bugs, in particular
when the user has no control over the test like above.

Since --host is provided the check "whether we are cross compiling"
is useless, as one explicitly declares a cross-compilation:

  --host=HOST   cross-compile to build programs to run on HOST [BUILD]

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

Bug#850329: [pkg-wine-party] Bug#850329: wine creates files with binary filenames

[Jens Reyer]

control: tags -1 + moreinfo unreproducible
control: severity -1 minor

Hi Vincent,

are the tests that you run available somewhere, so that I can try to
reproduce this?

Since this seems to affect only your tests, I'm downgrading the
severity, like we do for other app-specific problems.


Greets
jre

Bug#850329: wine creates files with binary filenames

[Vincent Lefevre]

On 2017-01-05 23:48:37 +0100, Vincent Lefevre wrote:
> When running tests with wine, files with binary filenames were
> created in the current directory:
> 
> -rw-r--r--  1 vlefevre vlefevre   0 2017-01-05 13:38:36 
> nX\016\005\v\005\340\a\001\v\001\002\033\030*\004\340\024\0200\@\020\002\004\001\004\260\005\004G\247\005\003\@\001
> -rw-r--r--  1 vlefevre vlefevre   0 2017-01-05 13:39:18 
> nX\016\005\v\005\340\a\001\v\001\002\033\030*\004\340\024\0200\@\020\002\004\001\004\260\005\004q\247\005\003\@\001

I could reproduce this problem one time (this time, with only one
such file created), but now I can no longer reproduce it.

-- 
Vincent Lefèvre  - Web: 
100% accessible validated (X)HTML - Blog: 
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)

Bug#850329: wine creates files with binary filenames

[Vincent Lefevre]

Package: wine
Version: 1.8.6-1
Severity: important

When running tests with wine, files with binary filenames were
created in the current directory:

-rw-r--r--  1 vlefevre vlefevre   0 2017-01-05 13:38:36 
nX\016\005\v\005\340\a\001\v\001\002\033\030*\004\340\024\0200\@\020\002\004\001\004\260\005\004G\247\005\003\@\001
-rw-r--r--  1 vlefevre vlefevre   0 2017-01-05 13:39:18 
nX\016\005\v\005\340\a\001\v\001\002\033\030*\004\340\024\0200\@\020\002\004\001\004\260\005\004q\247\005\003\@\001

(which makes Subversion commands fail, then).

I don't know what has changed in wine, but there was no such problem
in a recent past.

I don't have much time at the moment to look at this more closely.
But if someone else noticed a similar problem...

-- Package-specific info:
/usr/bin/wine points to /usr/bin/wine-stable.

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500, 
'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.8.0-2-amd64 (SMP w/12 CPU cores)
Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages wine depends on:
ii  wine32  1.8.6-1
ii  wine64  1.8.6-1

wine recommends no packages.

Versions of packages wine suggests:
pn  dosbox   
pn  playonlinux  
pn  wine-binfmt  
pn  winetricks   

Versions of packages wine is related to:
ii  fonts-wine  1.8.6-1
ii  wine1.8.6-1
ii  wine32  1.8.6-1
ii  wine64  1.8.6-1

-- no debconf information