sse: remove unbalanced HEDLEY_DIAGNOSTIC_PUSH

Also cleanup some illegal double underscore variable names Closes: #910
simd-everywhere · May 4, 2023 · 02c7a67 · 02c7a67
1 parent b0b370a
commit 02c7a67
Show file tree

Hide file tree

Showing 9 changed files with 57 additions and 72 deletions.
diff --git a/simde-common.h b/simde-common.h
@@ -1092,10 +1092,9 @@ SIMDE_DIAGNOSTIC_DISABLE_CPP98_COMPAT_PEDANTIC_
     HEDLEY_GCC_VERSION_CHECK(4,3,0)
 #  define SIMDE_BUG_IGNORE_SIGN_CONVERSION(expr) (__extension__ ({ \
        HEDLEY_DIAGNOSTIC_PUSH  \
-       HEDLEY_DIAGNOSTIC_POP  \
        _Pragma("GCC diagnostic ignored \"-Wsign-conversion\"") \
        __typeof__(expr) simde_bug_ignore_sign_conversion_v_= (expr); \
-       HEDLEY_DIAGNOSTIC_PUSH  \
+       HEDLEY_DIAGNOSTIC_POP  \
        simde_bug_ignore_sign_conversion_v_; \
      }))
 #else

diff --git a/x86/avx.h b/x86/avx.h
@@ -2127,7 +2127,7 @@ simde_mm256_round_ps (simde__m256 a, const int rounding) {
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_round_ps(a, rounding) SIMDE_STATEMENT_EXPR_(({ \
     simde__m256_private \
-      simde_mm256_round_ps_r_, \
+      simde_mm256_round_ps_r_ = simde__m256_to_private(simde_mm256_setzero_ps()), \
       simde_mm256_round_ps_a_ = simde__m256_to_private(a); \
     \
     for (size_t simde_mm256_round_ps_i = 0 ; simde_mm256_round_ps_i < (sizeof(simde_mm256_round_ps_r_.m128) / sizeof(simde_mm256_round_ps_r_.m128[0])) ; simde_mm256_round_ps_i++) { \
@@ -2201,7 +2201,7 @@ simde_mm256_round_pd (simde__m256d a, const int rounding) {
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_round_pd(a, rounding) SIMDE_STATEMENT_EXPR_(({ \
     simde__m256d_private \
-      simde_mm256_round_pd_r_, \
+      simde_mm256_round_pd_r_ = simde__m256d_to_private(simde_mm256_setzero_pd()), \
       simde_mm256_round_pd_a_ = simde__m256d_to_private(a); \
     \
     for (size_t simde_mm256_round_pd_i = 0 ; simde_mm256_round_pd_i < (sizeof(simde_mm256_round_pd_r_.m128d) / sizeof(simde_mm256_round_pd_r_.m128d[0])) ; simde_mm256_round_pd_i++) { \
@@ -3093,7 +3093,7 @@ simde_mm256_cmp_ps
 #elif defined(SIMDE_STATEMENT_EXPR_) && SIMDE_NATURAL_VECTOR_SIZE_LE(128)
   #define simde_mm256_cmp_ps(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m256_private \
-      simde_mm256_cmp_ps_r_, \
+      simde_mm256_cmp_ps_r_ = simde__m256_to_private(simde_mm256_setzero_ps()), \
       simde_mm256_cmp_ps_a_ = simde__m256_to_private((a)), \
       simde_mm256_cmp_ps_b_ = simde__m256_to_private((b)); \
     \

diff --git a/x86/avx2.h b/x86/avx2.h
@@ -3989,11 +3989,11 @@ simde_mm256_shuffle_epi32 (simde__m256i a, const int imm8)
        simde_mm_shuffle_epi32(simde_mm256_extracti128_si256(a, 0), (imm8)))
 #elif defined(SIMDE_SHUFFLE_VECTOR_)
 #  define simde_mm256_shuffle_epi32(a, imm8) (__extension__ ({ \
-      const simde__m256i_private simde__tmp_a_ = simde__m256i_to_private(a); \
+      const simde__m256i_private simde_tmp_a_ = simde__m256i_to_private(a); \
       simde__m256i_from_private((simde__m256i_private) { .i32 = \
           SIMDE_SHUFFLE_VECTOR_(32, 32, \
-                                (simde__tmp_a_).i32, \
-                                (simde__tmp_a_).i32, \
+                                (simde_tmp_a_).i32, \
+                                (simde_tmp_a_).i32, \
                                 ((imm8)     ) & 3, \
                                 ((imm8) >> 2) & 3, \
                                 ((imm8) >> 4) & 3, \
@@ -4017,11 +4017,11 @@ simde_mm256_shuffle_epi32 (simde__m256i a, const int imm8)
        simde_mm_shufflehi_epi16(simde_mm256_extracti128_si256(a, 0), (imm8)))
 #elif defined(SIMDE_SHUFFLE_VECTOR_)
 #  define simde_mm256_shufflehi_epi16(a, imm8) (__extension__ ({ \
-      const simde__m256i_private simde__tmp_a_ = simde__m256i_to_private(a); \
+      const simde__m256i_private simde_tmp_a_ = simde__m256i_to_private(a); \
       simde__m256i_from_private((simde__m256i_private) { .i16 = \
         SIMDE_SHUFFLE_VECTOR_(16, 32, \
-          (simde__tmp_a_).i16, \
-          (simde__tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
           0, 1, 2, 3, \
           (((imm8)     ) & 3) + 4, \
           (((imm8) >> 2) & 3) + 4, \
@@ -4053,11 +4053,11 @@ simde_mm256_shuffle_epi32 (simde__m256i a, const int imm8)
        simde_mm_shufflelo_epi16(simde_mm256_extracti128_si256(a, 0), (imm8)))
 #elif defined(SIMDE_SHUFFLE_VECTOR_)
 #  define simde_mm256_shufflelo_epi16(a, imm8) (__extension__ ({ \
-      const simde__m256i_private simde__tmp_a_ = simde__m256i_to_private(a); \
+      const simde__m256i_private simde_tmp_a_ = simde__m256i_to_private(a); \
       simde__m256i_from_private((simde__m256i_private) { .i16 = \
         SIMDE_SHUFFLE_VECTOR_(16, 32, \
-          (simde__tmp_a_).i16, \
-          (simde__tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
           (((imm8)     ) & 3), \
           (((imm8) >> 2) & 3), \
           (((imm8) >> 4) & 3), \

diff --git a/x86/avx512/cmp.h b/x86/avx512/cmp.h
@@ -237,7 +237,7 @@ simde_mm512_cmp_ps_mask (simde__m512 a, simde__m512 b, const int imm8)
 #elif defined(SIMDE_STATEMENT_EXPR_) && SIMDE_NATURAL_VECTOR_SIZE_LE(128)
   #define simde_mm512_cmp_ps_mask(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512_private \
-      simde_mm512_cmp_ps_mask_r_, \
+      simde_mm512_cmp_ps_mask_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
       simde_mm512_cmp_ps_mask_a_ = simde__m512_to_private((a)), \
       simde_mm512_cmp_ps_mask_b_ = simde__m512_to_private((b)); \
     \
@@ -250,7 +250,7 @@ simde_mm512_cmp_ps_mask (simde__m512 a, simde__m512 b, const int imm8)
 #elif defined(SIMDE_STATEMENT_EXPR_) && SIMDE_NATURAL_VECTOR_SIZE_LE(256)
   #define simde_mm512_cmp_ps_mask(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512_private \
-      simde_mm512_cmp_ps_mask_r_, \
+      simde_mm512_cmp_ps_mask_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
       simde_mm512_cmp_ps_mask_a_ = simde__m512_to_private((a)), \
       simde_mm512_cmp_ps_mask_b_ = simde__m512_to_private((b)); \
     \
@@ -485,7 +485,7 @@ simde_mm512_cmp_pd_mask (simde__m512d a, simde__m512d b, const int imm8)
 #elif defined(SIMDE_STATEMENT_EXPR_) && SIMDE_NATURAL_VECTOR_SIZE_LE(128)
   #define simde_mm512_cmp_pd_mask(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512d_private \
-      simde_mm512_cmp_pd_mask_r_, \
+      simde_mm512_cmp_pd_mask_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
       simde_mm512_cmp_pd_mask_a_ = simde__m512d_to_private((a)), \
       simde_mm512_cmp_pd_mask_b_ = simde__m512d_to_private((b)); \
     \
@@ -498,7 +498,7 @@ simde_mm512_cmp_pd_mask (simde__m512d a, simde__m512d b, const int imm8)
 #elif defined(SIMDE_STATEMENT_EXPR_) && SIMDE_NATURAL_VECTOR_SIZE_LE(256)
   #define simde_mm512_cmp_pd_mask(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512d_private \
-      simde_mm512_cmp_pd_mask_r_, \
+      simde_mm512_cmp_pd_mask_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
       simde_mm512_cmp_pd_mask_a_ = simde__m512d_to_private((a)), \
       simde_mm512_cmp_pd_mask_b_ = simde__m512d_to_private((b)); \
     \

diff --git a/x86/avx512/range.h b/x86/avx512/range.h
@@ -128,7 +128,7 @@ simde_mm256_range_ps (simde__m256 a, simde__m256 b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_range_ps(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m256_private \
-        simde_mm256_range_ps_r_, \
+        simde_mm256_range_ps_r_ = simde__m256_to_private(simde_mm256_setzero_ps()), \
         simde_mm256_range_ps_a_ = simde__m256_to_private(a), \
         simde_mm256_range_ps_b_ = simde__m256_to_private(b); \
       \
@@ -208,7 +208,7 @@ simde_mm512_range_ps (simde__m512 a, simde__m512 b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_range_ps(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m512_private \
-        simde_mm512_range_ps_r_, \
+        simde_mm512_range_ps_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
         simde_mm512_range_ps_a_ = simde__m512_to_private(a), \
         simde_mm512_range_ps_b_ = simde__m512_to_private(b); \
       \
@@ -221,7 +221,7 @@ simde_mm512_range_ps (simde__m512 a, simde__m512 b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_range_ps(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m512_private \
-        simde_mm512_range_ps_r_, \
+        simde_mm512_range_ps_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
         simde_mm512_range_ps_a_ = simde__m512_to_private(a), \
         simde_mm512_range_ps_b_ = simde__m512_to_private(b); \
       \
@@ -368,7 +368,7 @@ simde_mm256_range_pd (simde__m256d a, simde__m256d b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_range_pd(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m256d_private \
-        simde_mm256_range_pd_r_, \
+        simde_mm256_range_pd_r_ = simde__m256d_to_private(simde_mm256_setzero_pd()), \
         simde_mm256_range_pd_a_ = simde__m256d_to_private(a), \
         simde_mm256_range_pd_b_ = simde__m256d_to_private(b); \
       \
@@ -448,7 +448,7 @@ simde_mm512_range_pd (simde__m512d a, simde__m512d b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_range_pd(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m512d_private \
-        simde_mm512_range_pd_r_, \
+        simde_mm512_range_pd_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
         simde_mm512_range_pd_a_ = simde__m512d_to_private(a), \
         simde_mm512_range_pd_b_ = simde__m512d_to_private(b); \
       \
@@ -461,7 +461,7 @@ simde_mm512_range_pd (simde__m512d a, simde__m512d b, int imm8)
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_range_pd(a, b, imm8) SIMDE_STATEMENT_EXPR_(({ \
       simde__m512d_private \
-        simde_mm512_range_pd_r_, \
+        simde_mm512_range_pd_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
         simde_mm512_range_pd_a_ = simde__m512d_to_private(a), \
         simde_mm512_range_pd_b_ = simde__m512d_to_private(b); \
       \

diff --git a/x86/avx512/round.h b/x86/avx512/round.h
@@ -10,7 +10,7 @@ SIMDE_BEGIN_DECLS_
 #if SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_x_mm512_round_ps(a, rounding) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512_private \
-      simde_x_mm512_round_ps_r_, \
+      simde_x_mm512_round_ps_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
       simde_x_mm512_round_ps_a_ = simde__m512_to_private(a); \
     \
     for (size_t simde_x_mm512_round_ps_i = 0 ; simde_x_mm512_round_ps_i < (sizeof(simde_x_mm512_round_ps_r_.m256) / sizeof(simde_x_mm512_round_ps_r_.m256[0])) ; simde_x_mm512_round_ps_i++) { \
@@ -148,7 +148,7 @@ SIMDE_BEGIN_DECLS_
 #if SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_x_mm512_round_pd(a, rounding) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512d_private \
-      simde_x_mm512_round_pd_r_, \
+      simde_x_mm512_round_pd_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
       simde_x_mm512_round_pd_a_ = simde__m512d_to_private(a); \
     \
     for (size_t simde_x_mm512_round_pd_i = 0 ; simde_x_mm512_round_pd_i < (sizeof(simde_x_mm512_round_pd_r_.m256d) / sizeof(simde_x_mm512_round_pd_r_.m256d[0])) ; simde_x_mm512_round_pd_i++) { \

diff --git a/x86/avx512/roundscale.h b/x86/avx512/roundscale.h
@@ -73,7 +73,7 @@ SIMDE_BEGIN_DECLS_
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_roundscale_ps(a, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m256_private \
-      simde_mm256_roundscale_ps_r_, \
+      simde_mm256_roundscale_ps_r_ = simde__m256_to_private(simde_mm256_setzero_ps()), \
       simde_mm256_roundscale_ps_a_ = simde__m256_to_private(a); \
     \
     for (size_t simde_mm256_roundscale_ps_i = 0 ; simde_mm256_roundscale_ps_i < (sizeof(simde_mm256_roundscale_ps_r_.m128) / sizeof(simde_mm256_roundscale_ps_r_.m128[0])) ; simde_mm256_roundscale_ps_i++) { \
@@ -141,7 +141,7 @@ SIMDE_BEGIN_DECLS_
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_roundscale_ps(a, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512_private \
-      simde_mm512_roundscale_ps_r_, \
+      simde_mm512_roundscale_ps_r_ = simde__m512_to_private(simde_mm512_setzero_ps()), \
       simde_mm512_roundscale_ps_a_ = simde__m512_to_private(a); \
     \
     for (size_t simde_mm512_roundscale_ps_i = 0 ; simde_mm512_roundscale_ps_i < (sizeof(simde_mm512_roundscale_ps_r_.m256) / sizeof(simde_mm512_roundscale_ps_r_.m256[0])) ; simde_mm512_roundscale_ps_i++) { \
@@ -265,7 +265,7 @@ SIMDE_BEGIN_DECLS_
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(128) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm256_roundscale_pd(a, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m256d_private \
-      simde_mm256_roundscale_pd_r_, \
+      simde_mm256_roundscale_pd_r_ = simde__m256d_to_private(simde_mm256_setzero_pd()), \
       simde_mm256_roundscale_pd_a_ = simde__m256d_to_private(a); \
     \
     for (size_t simde_mm256_roundscale_pd_i = 0 ; simde_mm256_roundscale_pd_i < (sizeof(simde_mm256_roundscale_pd_r_.m128d) / sizeof(simde_mm256_roundscale_pd_r_.m128d[0])) ; simde_mm256_roundscale_pd_i++) { \
@@ -333,7 +333,7 @@ SIMDE_BEGIN_DECLS_
 #elif SIMDE_NATURAL_VECTOR_SIZE_LE(256) && defined(SIMDE_STATEMENT_EXPR_)
   #define simde_mm512_roundscale_pd(a, imm8) SIMDE_STATEMENT_EXPR_(({ \
     simde__m512d_private \
-      simde_mm512_roundscale_pd_r_, \
+      simde_mm512_roundscale_pd_r_ = simde__m512d_to_private(simde_mm512_setzero_pd()), \
       simde_mm512_roundscale_pd_a_ = simde__m512d_to_private(a); \
     \
     for (size_t simde_mm512_roundscale_pd_i = 0 ; simde_mm512_roundscale_pd_i < (sizeof(simde_mm512_roundscale_pd_r_.m256d) / sizeof(simde_mm512_roundscale_pd_r_.m256d[0])) ; simde_mm512_roundscale_pd_i++) { \

diff --git a/x86/sse.h b/x86/sse.h
@@ -3982,11 +3982,11 @@ simde_mm_sfence (void) {
 #  define simde_mm_shuffle_pi16(a, imm8) _mm_shuffle_pi16(a, imm8)
 #elif defined(SIMDE_SHUFFLE_VECTOR_)
 #  define simde_mm_shuffle_pi16(a, imm8) (__extension__ ({ \
-      const simde__m64_private simde__tmp_a_ = simde__m64_to_private(a); \
+      const simde__m64_private simde_tmp_a_ = simde__m64_to_private(a); \
       simde__m64_from_private((simde__m64_private) { .i16 = \
         SIMDE_SHUFFLE_VECTOR_(16, 8, \
-          (simde__tmp_a_).i16, \
-          (simde__tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
+          (simde_tmp_a_).i16, \
           (((imm8)     ) & 3), \
           (((imm8) >> 2) & 3), \
           (((imm8) >> 4) & 3), \
@@ -4592,11 +4592,6 @@ simde_mm_ucomineq_ss (simde__m128 a, simde__m128 b) {
 #  endif
 #endif
 
-#if defined(SIMDE_DIAGNOSTIC_DISABLE_UNINITIALIZED_)
-  HEDLEY_DIAGNOSTIC_PUSH
-  SIMDE_DIAGNOSTIC_DISABLE_UNINITIALIZED_
-#endif
-
 SIMDE_FUNCTION_ATTRIBUTES
 simde__m128
 simde_mm_unpackhi_ps (simde__m128 a, simde__m128 b) {