This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: Your INTERMEDIATE_ENCODING patch for Solaris
>>>>> "Pierre" == Pierre Muller <pierre.muller@ics-cnrs.unistra.fr> writes:
Pierre> Upon further investigation, I must confess that the OpenSolaris
Pierre> situation is a real mess:
Pierre> I get a working libiconv version for x86_64 solaris
Pierre> using CSW libiconv version 0x10D,
>From the version number, I guess this is GNU libiconv.
What happens if you don't build with libiconv, but rely instead on the
iconv in the Solaris libc?
Pierre> libiconv_open with parameter 'wchar_t' and 'UTF-8' works.
Pierre> but on my Sparc64 test machine, libiconv is
Pierre> on /usr/local/lib and its version is 0x10B.
Pierre> This later version fails on calls to libiconv_open
Pierre> with 'wchar_t' and 'ASCII'.
Ouch, but ok. We can check _LIBICONV_VERSION >= 0x10D.
This actually simplifies the patch a little; I didn't know about
_LIBICONV_VERSION before.
Could you try this? I also fixed the INTERMEDIATE_ENCODING problem you
pointed out.
Tom
2010-08-17 Tom Tromey <tromey@redhat.com>
* gdb_wchar.h: Check _LIBICONV_VERSION, __STDC_ISO_10646__.
Change how INTERMEDIATE_ENCODING is defined.
diff --git a/gdb/gdb_wchar.h b/gdb/gdb_wchar.h
index fca3fe4..78f59a1 100644
--- a/gdb/gdb_wchar.h
+++ b/gdb/gdb_wchar.h
@@ -23,7 +23,11 @@
Capable systems have the full suite: wchar_t support and iconv
(perhaps via GNU libiconv). On these machines, full functionality
- is available.
+ is available. Note that full functionality is dependent on us
+ being able to convert from an arbitrary encoding to wchar_t. In
+ practice this means we look for __STDC_ISO_10646__ (where we know
+ the name of the wchar_t encoding) or GNU libiconv, where we can use
+ "wchar_t".
DJGPP is known to have libiconv but not wchar_t support. On
systems like this, we use the narrow character functions. The full
@@ -35,8 +39,6 @@
wrappers for the wchar_t functionality we use. */
-#define INTERMEDIATE_ENCODING "wchar_t"
-
#if defined (HAVE_ICONV)
#include <iconv.h>
#else
@@ -45,9 +47,15 @@
#define PHONY_ICONV
#endif
-/* We use "btowc" as a sentinel to detect functioning wchar_t
- support. */
-#if defined (HAVE_ICONV) && defined (HAVE_WCHAR_H) && defined (HAVE_BTOWC)
+/* We use "btowc" as a sentinel to detect functioning wchar_t support.
+ We check for either __STDC_ISO_10646__ or a new-enough libiconv in
+ order to ensure we can convert to and from wchar_t. We choose
+ libiconv version 0x10D because it was reported that earlier
+ versions do not always accept "wchar_t" as an encoding argument on
+ Solaris. */
+#if defined (HAVE_ICONV) && defined (HAVE_WCHAR_H) && defined (HAVE_BTOWC) \
+ && (defined (__STDC_ISO_10646__) \
+ || defined (_LIBICONV_VERSION) && _LIBICONV_VERSION >= 0x10D)
#include <wchar.h>
#include <wctype.h>
@@ -63,6 +71,25 @@ typedef wint_t gdb_wint_t;
#define LCST(X) L ## X
+/* If __STDC_ISO_10646__ is defined, then the host wchar_t is UCS-4.
+ We exploit this fact in the hope that there are hosts that define
+ this but which do not support "wchar_t" as an encoding argument to
+ iconv_open. We put the endianness into the encoding name to avoid
+ hosts that emit a BOM when the unadorned name is used. */
+#if defined (__STDC_ISO_10646__)
+#if WORDS_BIGENDIAN
+#define INTERMEDIATE_ENCODING "UCS-4BE"
+#else
+#define INTERMEDIATE_ENCODING "UCS-4LE"
+#endif
+#elif defined (_LIBICONV_VERSION) && _LIBICONV_VERSION >= 0x10D
+#define INTERMEDIATE_ENCODING "wchar_t"
+#else
+/* This shouldn't happen, because the earlier #if should have filtered
+ out this case. */
+#error "Neither __STDC_ISO_10646__ nor _LIBICONV_VERSION defined"
+#endif
+
#else
typedef char gdb_wchar_t;
@@ -80,8 +107,9 @@ typedef int gdb_wint_t;
narrow encoding as our intermediate encoding. However, if we are
also providing a phony iconv, we might as well just stick with
"wchar_t". */
-#ifndef PHONY_ICONV
-#undef INTERMEDIATE_ENCODING
+#ifdef PHONY_ICONV
+#define INTERMEDIATE_ENCODING "wchar_t"
+#else
#define INTERMEDIATE_ENCODING host_charset ()
#endif