[PATCH BlueZ] obexd: Fix compile error due to lack of header

[PATCH BlueZ] obexd: Fix compile error due to lack of header

[ERAMOTO Masaya]

When compiling with gcc 7.2.0, gcc outputs the following messages.
This patch adds the header stdlib.h to each obex-related header including
functions with the type ssize_t, so the same problem does not also happen
in case using there headers individually.

  In file included from obexd/plugins/mas.c:41:0:
  ./obexd/src/obex.h:37:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
   ssize_t obex_get_size(struct obex_session *os);
   ^~~~~~~
   size_t
  ./obexd/src/obex.h:45:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
   ssize_t obex_get_apparam(struct obex_session *os, const uint8_t **buffer);
   ^~~~~~~
   size_t
  ./obexd/src/obex.h:46:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
   ssize_t obex_get_non_header_data(struct obex_session *os,
   ^~~~~~~
   size_t
  In file included from obexd/plugins/mas.c:43:0:
  ./obexd/src/mimetype.h:36:2: error: expected specifier-qualifier-list before ‘ssize_t’
    ssize_t (*get_next_header)(void *object, void *buf, size_t mtu,
    ^~~~~~~
  In file included from obexd/plugins/mas.c:46:0:
  obexd/plugins/filesystem.h:24:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
   ssize_t string_read(void *object, void *buf, size_t count);
   ^~~~~~~
   size_t
---
 obexd/plugins/filesystem.h | 2 ++
 obexd/src/mimetype.h       | 2 ++
 obexd/src/obex.h           | 2 ++
 3 files changed, 6 insertions(+)

diff --git a/obexd/plugins/filesystem.h b/obexd/plugins/filesystem.h
index f95773beb..441b377a2 100644
--- a/obexd/plugins/filesystem.h
+++ b/obexd/plugins/filesystem.h
@@ -21,6 +21,8 @@
  *
  */
 
+#include <stdlib.h>
+
 ssize_t string_read(void *object, void *buf, size_t count);
 gboolean is_filename(const char *name);
 int verify_path(const char *path);
diff --git a/obexd/src/mimetype.h b/obexd/src/mimetype.h
index 79529b890..52b544073 100644
--- a/obexd/src/mimetype.h
+++ b/obexd/src/mimetype.h
@@ -21,6 +21,8 @@
  *
  */
 
+#include <stdlib.h>
+
 typedef gboolean (*obex_object_io_func) (void *object, int flags, int err,
 							void *user_data);
 
diff --git a/obexd/src/obex.h b/obexd/src/obex.h
index fc1674755..769bbcd8a 100644
--- a/obexd/src/obex.h
+++ b/obexd/src/obex.h
@@ -22,6 +22,8 @@
  *
  */
 
+#include <stdlib.h>
+
 #define OBJECT_SIZE_UNKNOWN -1
 #define OBJECT_SIZE_DELETE -2
 
-- 
2.14.1

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[Luiz Augusto von Dentz]

Hi Eramoto,

On Thu, Oct 26, 2017 at 9:05 AM, ERAMOTO Masaya
<eramoto.masaya@jp.fujitsu.com> wrote:
> When compiling with gcc 7.2.0, gcc outputs the following messages.
> This patch adds the header stdlib.h to each obex-related header including
> functions with the type ssize_t, so the same problem does not also happen
> in case using there headers individually.

Weird this is just happening now, so what did change in 7.2? Usually
we have stdlib.h included in the c file not in the header.

>   In file included from obexd/plugins/mas.c:41:0:
>   ./obexd/src/obex.h:37:1: error: unknown type name =E2=80=98ssize_t=E2=
=80=99; did you mean =E2=80=98size_t=E2=80=99?
>    ssize_t obex_get_size(struct obex_session *os);
>    ^~~~~~~
>    size_t
>   ./obexd/src/obex.h:45:1: error: unknown type name =E2=80=98ssize_t=E2=
=80=99; did you mean =E2=80=98size_t=E2=80=99?
>    ssize_t obex_get_apparam(struct obex_session *os, const uint8_t **buff=
er);
>    ^~~~~~~
>    size_t
>   ./obexd/src/obex.h:46:1: error: unknown type name =E2=80=98ssize_t=E2=
=80=99; did you mean =E2=80=98size_t=E2=80=99?
>    ssize_t obex_get_non_header_data(struct obex_session *os,
>    ^~~~~~~
>    size_t
>   In file included from obexd/plugins/mas.c:43:0:
>   ./obexd/src/mimetype.h:36:2: error: expected specifier-qualifier-list b=
efore =E2=80=98ssize_t=E2=80=99
>     ssize_t (*get_next_header)(void *object, void *buf, size_t mtu,
>     ^~~~~~~
>   In file included from obexd/plugins/mas.c:46:0:
>   obexd/plugins/filesystem.h:24:1: error: unknown type name =E2=80=98ssiz=
e_t=E2=80=99; did you mean =E2=80=98size_t=E2=80=99?
>    ssize_t string_read(void *object, void *buf, size_t count);
>    ^~~~~~~
>    size_t
> ---
>  obexd/plugins/filesystem.h | 2 ++
>  obexd/src/mimetype.h       | 2 ++
>  obexd/src/obex.h           | 2 ++
>  3 files changed, 6 insertions(+)
>
> diff --git a/obexd/plugins/filesystem.h b/obexd/plugins/filesystem.h
> index f95773beb..441b377a2 100644
> --- a/obexd/plugins/filesystem.h
> +++ b/obexd/plugins/filesystem.h
> @@ -21,6 +21,8 @@
>   *
>   */
>
> +#include <stdlib.h>
> +
>  ssize_t string_read(void *object, void *buf, size_t count);
>  gboolean is_filename(const char *name);
>  int verify_path(const char *path);
> diff --git a/obexd/src/mimetype.h b/obexd/src/mimetype.h
> index 79529b890..52b544073 100644
> --- a/obexd/src/mimetype.h
> +++ b/obexd/src/mimetype.h
> @@ -21,6 +21,8 @@
>   *
>   */
>
> +#include <stdlib.h>
> +
>  typedef gboolean (*obex_object_io_func) (void *object, int flags, int er=
r,
>                                                         void *user_data);
>
> diff --git a/obexd/src/obex.h b/obexd/src/obex.h
> index fc1674755..769bbcd8a 100644
> --- a/obexd/src/obex.h
> +++ b/obexd/src/obex.h
> @@ -22,6 +22,8 @@
>   *
>   */
>
> +#include <stdlib.h>
> +
>  #define OBJECT_SIZE_UNKNOWN -1
>  #define OBJECT_SIZE_DELETE -2
>
> --
> 2.14.1
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-bluetooth=
" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



--=20
Luiz Augusto von Dentz

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[ERAMOTO Masaya]

Hi Luiz,

On 10/26/2017 09:05 PM, Luiz Augusto von Dentz wrote:
> Hi Eramoto,
> 
> On Thu, Oct 26, 2017 at 9:05 AM, ERAMOTO Masaya
> <eramoto.masaya@jp.fujitsu.com> wrote:
>> When compiling with gcc 7.2.0, gcc outputs the following messages.
>> This patch adds the header stdlib.h to each obex-related header including
>> functions with the type ssize_t, so the same problem does not also happen
>> in case using there headers individually.
> 
> Weird this is just happening now, so what did change in 7.2? Usually
> we have stdlib.h included in the c file not in the header.


I do not know what change in gcc 7.2 triggers this issue since I have only
upgraded OS. obexd/plugins/mas.c does not always include stdlib.h, so I
guess gcc 7.2 has enhanced type checking in header file.


Regards,
Eramoto

> 
>>   In file included from obexd/plugins/mas.c:41:0:
>>   ./obexd/src/obex.h:37:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
>>    ssize_t obex_get_size(struct obex_session *os);
>>    ^~~~~~~
>>    size_t
>>   ./obexd/src/obex.h:45:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
>>    ssize_t obex_get_apparam(struct obex_session *os, const uint8_t **buffer);
>>    ^~~~~~~
>>    size_t
>>   ./obexd/src/obex.h:46:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
>>    ssize_t obex_get_non_header_data(struct obex_session *os,
>>    ^~~~~~~
>>    size_t
>>   In file included from obexd/plugins/mas.c:43:0:
>>   ./obexd/src/mimetype.h:36:2: error: expected specifier-qualifier-list before ‘ssize_t’
>>     ssize_t (*get_next_header)(void *object, void *buf, size_t mtu,
>>     ^~~~~~~
>>   In file included from obexd/plugins/mas.c:46:0:
>>   obexd/plugins/filesystem.h:24:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
>>    ssize_t string_read(void *object, void *buf, size_t count);
>>    ^~~~~~~
>>    size_t
>> ---
>>  obexd/plugins/filesystem.h | 2 ++
>>  obexd/src/mimetype.h       | 2 ++
>>  obexd/src/obex.h           | 2 ++
>>  3 files changed, 6 insertions(+)
>>
>> diff --git a/obexd/plugins/filesystem.h b/obexd/plugins/filesystem.h
>> index f95773beb..441b377a2 100644
>> --- a/obexd/plugins/filesystem.h
>> +++ b/obexd/plugins/filesystem.h
>> @@ -21,6 +21,8 @@
>>   *
>>   */
>>
>> +#include <stdlib.h>
>> +
>>  ssize_t string_read(void *object, void *buf, size_t count);
>>  gboolean is_filename(const char *name);
>>  int verify_path(const char *path);
>> diff --git a/obexd/src/mimetype.h b/obexd/src/mimetype.h
>> index 79529b890..52b544073 100644
>> --- a/obexd/src/mimetype.h
>> +++ b/obexd/src/mimetype.h
>> @@ -21,6 +21,8 @@
>>   *
>>   */
>>
>> +#include <stdlib.h>
>> +
>>  typedef gboolean (*obex_object_io_func) (void *object, int flags, int err,
>>                                                         void *user_data);
>>
>> diff --git a/obexd/src/obex.h b/obexd/src/obex.h
>> index fc1674755..769bbcd8a 100644
>> --- a/obexd/src/obex.h
>> +++ b/obexd/src/obex.h
>> @@ -22,6 +22,8 @@
>>   *
>>   */
>>
>> +#include <stdlib.h>
>> +
>>  #define OBJECT_SIZE_UNKNOWN -1
>>  #define OBJECT_SIZE_DELETE -2
>>
>> --
>> 2.14.1
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 
> 

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[Bastien Nocera]

On Thu, 2017-10-26 at 15:05 +0300, Luiz Augusto von Dentz wrote:
> Hi Eramoto,
> 
> On Thu, Oct 26, 2017 at 9:05 AM, ERAMOTO Masaya
> <eramoto.masaya@jp.fujitsu.com> wrote:
> > When compiling with gcc 7.2.0, gcc outputs the following messages.
> > This patch adds the header stdlib.h to each obex-related header
> > including
> > functions with the type ssize_t, so the same problem does not also
> > happen
> > in case using there headers individually.
> 
> Weird this is just happening now, so what did change in 7.2? Usually
> we have stdlib.h included in the c file not in the header.

It's not a gcc change, but a glibc change. See the patch labelled "Fix
compilation error on newer glibc" I just sent.

Cheers

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[Bastien Nocera]

On Thu, 2017-10-26 at 15:05 +0900, ERAMOTO Masaya wrote:
> When compiling with gcc 7.2.0, gcc outputs the following messages.
> This patch adds the header stdlib.h to each obex-related header
> including
> functions with the type ssize_t, so the same problem does not also
> happen
> in case using there headers individually.

I already sent a patch for this about 2 months ago, look for "obexd:
Fix compilation error on F27"

Cheers

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[ERAMOTO Masaya]

Hi Bastien,

On 10/26/2017 09:39 PM, Bastien Nocera wrote:
> On Thu, 2017-10-26 at 15:05 +0900, ERAMOTO Masaya wrote:
>> When compiling with gcc 7.2.0, gcc outputs the following messages.
>> This patch adds the header stdlib.h to each obex-related header
>> including
>> functions with the type ssize_t, so the same problem does not also
>> happen
>> in case using there headers individually.
> 
> I already sent a patch for this about 2 months ago, look for "obexd:
> Fix compilation error on F27"

Sorry for I had missed this patch.

Now, it seems that you are blocking the patch set including this patch
due to that Florian commented about one of the patch set. Is there any
upgrades for the patch set?


Regards,
Eramoto

Re: [PATCH BlueZ] obexd: Fix compile error due to lack of header

[Bastien Nocera]

On Fri, 2017-10-27 at 16:55 +0900, ERAMOTO Masaya wrote:
> Hi Bastien,
> 
> On 10/26/2017 09:39 PM, Bastien Nocera wrote:
> > On Thu, 2017-10-26 at 15:05 +0900, ERAMOTO Masaya wrote:
> > > When compiling with gcc 7.2.0, gcc outputs the following
> > > messages.
> > > This patch adds the header stdlib.h to each obex-related header
> > > including
> > > functions with the type ssize_t, so the same problem does not
> > > also
> > > happen
> > > in case using there headers individually.
> > 
> > I already sent a patch for this about 2 months ago, look for
> > "obexd:
> > Fix compilation error on F27"
> 
> Sorry for I had missed this patch.
> 
> Now, it seems that you are blocking the patch set including this
> patch
> due to that Florian commented about one of the patch set.

I'm not blocking anything, it didn't get reviewed at all.

>  Is there any
> upgrades for the patch set?

The patch set is just a bunch of unrelated fixes, the patches can be
committed individually.

I think one of us needs to figure out what the change is to put the
justification in the commit message before that's going to be accepted
though.