Yes!
this solution is really more upstream/friendly, so if you forward please drop a note on the patch with the link,
and the last problem I found:
dget -u http://mentors.debian.net/debian/pool/main/n/neko/neko_2.0.0-4.dsccd neko-2.0.0/
dpkg-buildpackage -S -sa
[...]
export DEB_CFLAGS_MAINT_APPEND = -fno-strict-aliasing
/bin/sh: 1: export: : bad variable name
I guess changing it to
export DEB_CFLAGS_MAINT_APPEND = -fno-strict-aliasing
(with no initial tab) fixes the issue.
BTW the hardening stuff gained some gcc warnings that might be addressed upstream:
cc -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -Wl,-z,relro -pthread -o vm/others.o -c vm/others.c
vm/others.c: In function 'neko_val_throw':
vm/others.c:484:2: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
if( *(char**)vm->start == jit_handle_trap )
^
vm/others.c: In function 'neko_val_rethrow':
vm/others.c:493:2: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
if( *(char**)vm->start == jit_handle_trap )
^
cc -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -Wl,-z,relro -pthread -o libs/std/misc.o -c libs/std/misc.c
libs/std/misc.c: In function 'float_of_bytes':
libs/std/misc.c:81:2: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
f = *(float*)val_string(s);
^
libs/std/misc.c: In function 'double_of_bytes':
libs/std/misc.c:101:2: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
f = *(double*)val_string(s);
^
gcc -O3 -DEAPI -fPIC -pthread $(pkg-config --cflags apr-1) -pthread -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -c -I../../vm -I/usr/include sqlite.c
In file included from sqlite.c:22:0:
sqlite.c: In function 'result_next':
../../vm/neko.h:332:23: warning: 'f' may be used uninitialized in this function [-Wmaybe-uninitialized]
#define alloc_field neko_alloc_field
^
sqlite.c:225:10: note: 'f' was declared here
value f;
^
protocol.c: In function 'protocol_read_answer':
protocol.c:401:21: warning: pointer targets in passing argument 2 of 'proto_read' differ in signedness [-Wpointer-sign]
if( !proto_read(p,header,4) )
^
protocol.c:126:13: note: expected 'char *' but argument is of type 'unsigned char *'
static bool proto_read( proto *p, char *str, int len ) {
^
gcc -O3 -DEAPI -fPIC -pthread $(pkg-config --cflags apr-1) -pthread -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -D_LARGEFILE64_SOURCE -I../common -c -I../../vm -I/usr/include/apache2 -I/usr/include/apr-1.0 mod_tora.c
mod_tora.c: In function 'mod_tora_do_init':
mod_tora.c:289:6: warning: unused variable 'tmp' [-Wunused-variable]
int tmp = 0;
^
gcc -L/build/neko-2.0.0/src -L/usr/lib -L/usr/lib/mysql -o ../../bin/mod_tora2.ndll socket.o protocol.o mod_tora.o -shared -L../../bin -pthread -lneko -Wl,-z,relro
gcc -O3 -DEAPI -fPIC -pthread $(pkg-config --cflags apr-1) -pthread -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -D_LARGEFILE64_SOURCE -c -I../../vm -I/usr/include/apache2 -I/usr/include/apr-1.0 mod_neko.c
mod_neko.c: In function 'mod_neko_do_init':
mod_neko.c:322:6: warning: unused variable 'tmp' [-Wunused-variable]
int tmp = 0;
^
mod_neko.c: At top level:
mod_neko.c:117:13: warning: 'null_print' defined but not used [-Wunused-function]
static void null_print( const char *data, int size, void *_c ) {
^
gcc -O3 -DEAPI -fPIC -pthread $(pkg-config --cflags apr-1) -pthread -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wall -fPIC -I vm -DCOMPACT_TABLE -D_GNU_SOURCE -I libs/common -D_LARGEFILE64_SOURCE -c -I../../vm -I/usr/include/apache2 -I/usr/include/apr-1.0 cgi.c
ui.c: In function 'ui_main':
ui.c:151:2: warning: 'g_thread_init' is deprecated [-Wdeprecated-declarations]
g_thread_init(NULL);
the last one can be fixed with a
// g_thread_init has been deprecated since version 2.32 and should not be used in newly-written code. This function is no longer necessary. The GLib threading system is automatically initialized at the start of your program.
#if !(GLIB_CHECK_VERSION(2,32,0))
g_thread_init(NULL);
#endif
block I guess
(I'm not asking you to fix the above, but I would appreciate if you can forward them upstream :)
cheers,
G.
Il Giovedì 24 Settembre 2015 12:51, Andy Li <andy@onthewings.net> ha scritto:
Hi Gianfranco,
Thanks for the review!
Good idea about using the *FLAGS env vars. I've updated the patch to use those and will also propose the change to upstream.
I've also added back -Wall as suggested.
The package has just been reuploaded to mentors.
Best regards,
Andy
On Thu, Sep 24, 2015 at 5:54 PM, Gianfranco Costamagna <costamagnagianfranco@yahoo.it> wrote:
Hi Andy,
>
>
>
>the packaging looks good.
>
>I did two builds with DH_VERBOSE=1 in a sid chroot
>(the unstable and mentors versions)
>
>I did a meld between them, and I see some "issues"
>
>-"Wall" flag has disappeared
>
>(I guess not a real problem here :) )
>
>
>- many lines are like this one
>gcc -O3 -DEAPI -fPIC -pthread $(pkg-config --cflags apr-1) -pthread $(dpkg-buildflags --get CFLAGS) -c -I../../vm -I/usr/include/mysql mysql.c
>
>evaluating $(dpkg-buildflags --get CFLAGS) at each file build (and link, the same is for LDFLAGS) is somewhat an overkill.
>
>I see they come from debian/patches/harden.patch,
>just as TODO you might ask upstream to use *FLAGS from the environment, and drop that patch.
>
>
>but we are nitpicking here :)
>
>So if you can readd the -Wall flag I guess we are fine there :)
>
>cheers,
>
>G.
>