Обсуждение: ConnInfo assumes signed char

ConnInfo, defined in connection.h, assumes that char = signed char.

This is not true of Linux on arm, powerpc and s390, and possibly other
platforms.  As a result:

gcc -DHAVE_CONFIG_H -I. -I. -I. -g -O2 -c dlg_specific.c -MT dlg_specific.lo -MD -MP -MF .deps/dlg_specific.TPlo  -fPIC
-DPIC-o dlg_specific.lo 
dlg_specific.c: In function `getDSNdefaults':
dlg_specific.c:398: warning: comparison is always false due to limited range of data type
dlg_specific.c:400: warning: comparison is always false due to limited range of data type
dlg_specific.c:402: warning: comparison is always false due to limited range of data type
dlg_specific.c:404: warning: comparison is always false due to limited range of data type
dlg_specific.c:406: warning: comparison is always false due to limited range of data type
dlg_specific.c: In function `getDSNinfo':
dlg_specific.c:484: warning: comparison is always false due to limited range of data type
dlg_specific.c:491: warning: comparison is always false due to limited range of data type
dlg_specific.c:498: warning: comparison is always false due to limited range of data type
dlg_specific.c:505: warning: comparison is always false due to limited range of data type
dlg_specific.c:512: warning: comparison is always false due to limited range of data type

The problem still exists in current CVS.

Where char may contain a negative value, it should explicitly be
declared as signed char.

--
Oliver Elphick                                Oliver.Elphick@lfix.co.uk
Isle of Wight, UK                             http://www.lfix.co.uk/oliver
GPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839  932A 614D 4C34 3E1D 0C1C
                 ========================================
     "But the LORD is in his holy temple; let all the earth
      keep silence before him."           Habakkuk 2:20

Oliver Elphick <olly@lfix.co.uk> writes:
> ConnInfo, defined in connection.h, assumes that char = signed char.

I agree, this is a genuine bug.  It appears to have been introduced
post-7.2, as the ODBC sources in the REL7_2 branch don't seem to
try to assign negative values to these fields.

            regards, tom lane

Oliver Elphick wrote:
>
> ConnInfo, defined in connection.h, assumes that char = signed char.
>
> This is not true of Linux on arm, powerpc and s390, and possibly other
> platforms.  As a result:
>
> gcc -DHAVE_CONFIG_H -I. -I. -I. -g -O2 -c dlg_specific.c -MT dlg_specific.lo -MD -MP -MF .deps/dlg_specific.TPlo
-fPIC-DPIC -o dlg_specific.lo 
> dlg_specific.c: In function `getDSNdefaults':
> dlg_specific.c:398: warning: comparison is always false due to limited range of data type
> dlg_specific.c:400: warning: comparison is always false due to limited range of data type
> dlg_specific.c:402: warning: comparison is always false due to limited range of data type
> dlg_specific.c:404: warning: comparison is always false due to limited range of data type
> dlg_specific.c:406: warning: comparison is always false due to limited range of data type
> dlg_specific.c: In function `getDSNinfo':
> dlg_specific.c:484: warning: comparison is always false due to limited range of data type
> dlg_specific.c:491: warning: comparison is always false due to limited range of data type
> dlg_specific.c:498: warning: comparison is always false due to limited range of data type
> dlg_specific.c:505: warning: comparison is always false due to limited range of data type
> dlg_specific.c:512: warning: comparison is always false due to limited range of data type
>
> The problem still exists in current CVS.
>
> Where char may contain a negative value, it should explicitly be
> declared as signed char.

OK I changed them to *signed char*s (maybe).
Thanks.

regards,
Hiroshi Inoue
    http://www.geocities.jp/inocchichichi/psqlodbc/