path: root/source/l/mpfr/patches/patch08

diff -Naurd mpfr-4.2.0-a/PATCHES mpfr-4.2.0-b/PATCHES
--- mpfr-4.2.0-a/PATCHES	2023-05-17 17:17:28.512360351 +0000
+++ mpfr-4.2.0-b/PATCHES	2023-05-17 17:17:28.600360192 +0000
@@ -0,0 +1 @@
+compound
diff -Naurd mpfr-4.2.0-a/VERSION mpfr-4.2.0-b/VERSION
--- mpfr-4.2.0-a/VERSION	2023-05-12 15:08:39.325546612 +0000
+++ mpfr-4.2.0-b/VERSION	2023-05-17 17:17:28.600360192 +0000
@@ -1 +1 @@
-4.2.0-p7
+4.2.0-p8
diff -Naurd mpfr-4.2.0-a/src/compound.c mpfr-4.2.0-b/src/compound.c
--- mpfr-4.2.0-a/src/compound.c	2023-01-05 17:09:48.000000000 +0000
+++ mpfr-4.2.0-b/src/compound.c	2023-05-17 17:17:28.588360213 +0000
@@ -55,9 +55,9 @@
 mpfr_compound_si (mpfr_ptr y, mpfr_srcptr x, long n, mpfr_rnd_t rnd_mode)
 {
   int inexact, compared, k, nloop;
-  mpfr_t t;
-  mpfr_exp_t e;
-  mpfr_prec_t prec;
+  mpfr_t t, u;
+  mpfr_prec_t py, prec, extra;
+  mpfr_rnd_t rnd1;
   MPFR_ZIV_DECL (loop);
   MPFR_SAVE_EXPO_DECL (expo);
 
@@ -136,64 +136,185 @@
 
   MPFR_SAVE_EXPO_MARK (expo);
 
-  prec = MPFR_PREC(y);
-  prec += MPFR_INT_CEIL_LOG2 (prec) + 6;
+  py = MPFR_GET_PREC (y);
+  prec = py + MPFR_INT_CEIL_LOG2 (py) + 6;
 
   mpfr_init2 (t, prec);
+  mpfr_init2 (u, prec);
 
   k = MPFR_INT_CEIL_LOG2(SAFE_ABS (unsigned long, n));  /* thus |n| <= 2^k */
 
+  /* We compute u=log2p1(x) with prec+extra bits, since we lose some bits
+     in 2^u. */
+  extra = 0;
+  rnd1 = VSIGN (n) == MPFR_SIGN (x) ? MPFR_RNDD : MPFR_RNDU;
+
   MPFR_ZIV_INIT (loop, prec);
   for (nloop = 0; ; nloop++)
     {
-      /* we compute (1+x)^n as 2^(n*log2p1(x)) */
-      inexact = mpfr_log2p1 (t, x, MPFR_RNDN) != 0;
-      e = MPFR_GET_EXP(t);
-      /* |t - log2(1+x)| <= 1/2*ulp(t) = 2^(e-prec-1) */
-      inexact |= mpfr_mul_si (t, t, n, MPFR_RNDN) != 0;
-      /* |t - n*log2(1+x)| <= 2^(e2-prec-1) + |n|*2^(e-prec-1)
-                           <= 2^(e2-prec-1) + 2^(e+k-prec-1) <= 2^(e+k-prec)
-                          where |n| <= 2^k, and e2 is the new exponent of t. */
-      MPFR_ASSERTD(MPFR_GET_EXP(t) <= e + k);
-      e += k;
-      /* |t - n*log2(1+x)| <= 2^(e-prec) */
-      /* detect overflow */
-      if (nloop == 0 && mpfr_cmp_si (t, __gmpfr_emax) >= 0)
+      unsigned int inex;
+      mpfr_exp_t e, e2, ex;
+      mpfr_prec_t precu = MPFR_ADD_PREC (prec, extra);
+      mpfr_prec_t new_extra;
+      mpfr_rnd_t rnd2;
+
+      /* We compute (1+x)^n as 2^(n*log2p1(x)),
+         and we round toward 1, thus we round n*log2p1(x) toward 0,
+         thus for x*n > 0 we round log2p1(x) toward -Inf, and for x*n < 0
+         we round log2p1(x) toward +Inf. */
+      inex = mpfr_log2p1 (u, x, rnd1) != 0;
+      e = MPFR_GET_EXP (u);
+      /* |u - log2(1+x)| <= ulp(t) = 2^(e-precu) */
+      inex |= mpfr_mul_si (u, u, n, MPFR_RNDZ) != 0;
+      e2 = MPFR_GET_EXP (u);
+      /* |u - n*log2(1+x)| <= 2^(e2-precu) + |n|*2^(e-precu)
+                           <= 2^(e2-precu) + 2^(e+k-precu) <= 2^(e+k+1-precu)
+                          where |n| <= 2^k, and e2 is the new exponent of u. */
+      MPFR_ASSERTD (e2 <= e + k);
+      e += k + 1;
+      MPFR_ASSERTN (e2 <= MPFR_PREC_MAX);
+      new_extra = e2 > 0 ? e2 : 0;
+      /* |u - n*log2(1+x)| <= 2^(e-precu) */
+      /* detect overflow: since we rounded n*log2p1(x) toward 0,
+         if n*log2p1(x) >= __gmpfr_emax, we are sure there is overflow. */
+      if (mpfr_cmp_si (u, __gmpfr_emax) >= 0)
         {
           MPFR_ZIV_FREE (loop);
           mpfr_clear (t);
+          mpfr_clear (u);
           MPFR_SAVE_EXPO_FREE (expo);
           return mpfr_overflow (y, rnd_mode, 1);
         }
-      /* detect underflow */
-      if (nloop == 0 && mpfr_cmp_si (t, __gmpfr_emin - 1) <= 0)
+      /* detect underflow: similarly, since we rounded n*log2p1(x) toward 0,
+         if n*log2p1(x) < __gmpfr_emin-1, we are sure there is underflow. */
+      if (mpfr_cmp_si (u, __gmpfr_emin - 1) < 0)
         {
           MPFR_ZIV_FREE (loop);
           mpfr_clear (t);
+          mpfr_clear (u);
           MPFR_SAVE_EXPO_FREE (expo);
           return mpfr_underflow (y,
-                            (rnd_mode == MPFR_RNDN) ? MPFR_RNDZ : rnd_mode, 1);
+                            rnd_mode == MPFR_RNDN ? MPFR_RNDZ : rnd_mode, 1);
         }
       /* Detect cases where result is 1 or 1+ulp(1) or 1-1/2*ulp(1):
-         |2^t - 1| = |exp(t*log(2)) - 1| <= |t|*log(2) < |t| */
-      if (nloop == 0 && MPFR_GET_EXP(t) < - (mpfr_exp_t) MPFR_PREC(y))
+         |2^u - 1| = |exp(u*log(2)) - 1| <= |u|*log(2) < |u| */
+      if (nloop == 0 && MPFR_GET_EXP(u) < - py)
         {
-          /* since ulp(1) = 2^(1-PREC(y)), we have |t| < 1/4*ulp(1) */
+          /* since ulp(1) = 2^(1-py), we have |u| < 1/4*ulp(1) */
           /* mpfr_compound_near_one must be called in the extended
              exponent range, so that 1 is representable. */
-          inexact = mpfr_compound_near_one (y, MPFR_SIGN (t), rnd_mode);
+          inexact = mpfr_compound_near_one (y, MPFR_SIGN (u), rnd_mode);
           goto end;
         }
-      inexact |= mpfr_exp2 (t, t, MPFR_RNDA) != 0;
-      /* |t - (1+x)^n| <= ulp(t) + |t|*log(2)*2^(e-prec)
-                       < 2^(EXP(t)-prec) + 2^(EXP(t)+e-prec) */
-      e = (e >= 0) ? e + 1 : 1;
+      /* FIXME: mpfr_exp2 could underflow to the smallest positive number
+         since MPFR_RNDA is used, and this case will not be detected by
+         MPFR_CAN_ROUND (see BUGS). Either fix that, or do early underflow
+         detection (which may be necessary). */
+      /* round 2^u toward 1 */
+      rnd2 = MPFR_IS_POS (u) ? MPFR_RNDD : MPFR_RNDU;
+      inex |= mpfr_exp2 (t, u, rnd2) != 0;
+      /* we had |u - n*log2(1+x)| < 2^(e-precu)
+         thus u = n*log2(1+x) + delta with |delta| < 2^(e-precu)
+         then 2^u = (1+x)^n * 2^delta with |delta| < 2^(e-precu).
+         For |delta| < 0.5, |2^delta - 1| <= |delta| thus
+         |t - (1+x)^n| <= ulp(t) + |t|*2^(e-precu)
+                       < 2^(EXP(t)-prec) + 2^(EXP(t)+e-precu) */
+      e = (precu - prec >= e) ? 1 : e + 1 - (precu - prec);
       /* now |t - (1+x)^n| < 2^(EXP(t)+e-prec) */
 
-      if (MPFR_LIKELY (inexact == 0 ||
-                       MPFR_CAN_ROUND (t, prec - e, MPFR_PREC(y), rnd_mode)))
+      if (MPFR_LIKELY (!inex || MPFR_CAN_ROUND (t, prec - e, py, rnd_mode)))
         break;
 
+      /* If t fits in the target precision (or with 1 more bit), then we can
+         round, assuming the working precision is large enough, but the above
+         MPFR_CAN_ROUND() will fail because we cannot determine the ternary
+         value. However since we rounded t toward 1, we can determine it.
+         Since the error in the approximation t is at most 2^e ulp(t),
+         this error should be less than 1/2 ulp(y), thus we should have
+         prec - py >= e + 1. */
+      if (mpfr_min_prec (t) <= py + 1 && prec - py >= e + 1)
+        {
+          /* we add/subtract one ulp to get the correct rounding */
+          if (rnd2 == MPFR_RNDD) /* t was rounded downwards */
+            mpfr_nextabove (t);
+          else
+            mpfr_nextbelow (t);
+          break;
+        }
+
+      /* Detect particular cases where Ziv's strategy may take too much
+         memory and be too long, i.e. when x^n fits in the target precision
+         (+ 1 additional bit for rounding to nearest) and the exact result
+         (1+x)^n is very close to x^n.
+         Necessarily, x is a large even integer and n > 0 (thus n > 1).
+         Since this does not depend on the working precision, we only
+         check this at the first iteration (nloop == 0).
+         Hence the first "if" below and the kx < ex test of the second "if"
+         (x is an even integer iff its least bit 1 has exponent >= 1).
+         The second test of the second "if" corresponds to another simple
+         condition that implies that x^n fits in the target precision.
+         Here are the details:
+         Let k be the minimum length of the significand of x, and x' the odd
+         (integer) significand of x. This means  that 2^(k-1) <= x' < 2^k.
+         Thus 2^(n*(k-1)) <= (x')^n < 2^(k*n), and x^n has between n*(k-1)+1
+         and k*n bits. So x^n can fit into p bits only if p >= n*(k-1)+1,
+         i.e. n*(k-1) <= p-1.
+         Note that x >= 2^k, so that x^n >= 2^(k*n). Since raw overflow
+         has already been detected, k*n cannot overflow if computed with
+         the mpfr_exp_t type. Hence the second test of the second "if",
+         which cannot overflow. */
+      MPFR_ASSERTD (n < 0 || n > 1);
+      if (nloop == 0 && n > 1 && (ex = MPFR_GET_EXP (x)) >= 17)
+        {
+          mpfr_prec_t kx = mpfr_min_prec (x);
+          mpfr_prec_t p = py + (rnd_mode == MPFR_RNDN);
+
+          MPFR_LOG_MSG (("Check if x^n fits... n=%ld kx=%Pd p=%Pd\n",
+                         n, kx, p));
+          if (kx < ex && n * (mpfr_exp_t) (kx - 1) <= p - 1)
+            {
+              mpfr_t v;
+
+              /* Check whether x^n really fits into p bits. */
+              mpfr_init2 (v, p);
+              inexact = mpfr_pow_ui (v, x, n, MPFR_RNDZ);
+              if (inexact == 0)
+                {
+                  MPFR_LOG_MSG (("x^n fits into p bits\n", 0));
+                  /* (x+1)^n = x^n * (1 + 1/x)^n
+                     For directed rounding, we can round when (1 + 1/x)^n
+                     < 1 + 2^-p, and then the result is x^n,
+                     except for rounding up. Indeed, if (1 + 1/x)^n < 1 + 2^-p,
+                     1 <= (x+1)^n < x^n * (1 + 2^-p) = x^n + x^n/2^p
+                     < x^n + ulp(x^n).
+                     For rounding to nearest, we can round when (1 + 1/x)^n
+                     < 1 + 2^-p, and then the result is x^n when x^n fits
+                     into p-1 bits, and nextabove(x^n) otherwise. */
+                  mpfr_ui_div (t, 1, x, MPFR_RNDU);
+                  mpfr_add_ui (t, t, 1, MPFR_RNDU);
+                  mpfr_pow_ui (t, t, n, MPFR_RNDU);
+                  mpfr_sub_ui (t, t, 1, MPFR_RNDU);
+                  /* t cannot be zero */
+                  if (MPFR_GET_EXP(t) < - py)
+                    {
+                      mpfr_set (y, v, MPFR_RNDZ);
+                      if ((rnd_mode == MPFR_RNDN && mpfr_min_prec (v) == p)
+                          || rnd_mode == MPFR_RNDU || rnd_mode == MPFR_RNDA)
+                        {
+                          /* round up */
+                          mpfr_nextabove (y);
+                          inexact = 1;
+                        }
+                      else
+                        inexact = -1;
+                      mpfr_clear (v);
+                      goto end;
+                    }
+                }
+              mpfr_clear (v);
+            }
+        }
+
       /* Exact cases like compound(0.5,2) = 9/4 must be detected, since
          except for 1+x power of 2, the log2p1 above will be inexact,
          so that in the Ziv test, inexact != 0 and MPFR_CAN_ROUND will
@@ -211,6 +332,8 @@
 
       MPFR_ZIV_NEXT (loop, prec);
       mpfr_set_prec (t, prec);
+      extra = new_extra;
+      mpfr_set_prec (u, MPFR_ADD_PREC (prec, extra));
     }
 
   inexact = mpfr_set (y, t, rnd_mode);
@@ -218,6 +341,7 @@
  end:
   MPFR_ZIV_FREE (loop);
   mpfr_clear (t);
+  mpfr_clear (u);
 
   MPFR_SAVE_EXPO_FREE (expo);
   return mpfr_check_range (y, inexact, rnd_mode);
diff -Naurd mpfr-4.2.0-a/src/mpfr.h mpfr-4.2.0-b/src/mpfr.h
--- mpfr-4.2.0-a/src/mpfr.h	2023-05-12 15:08:39.321546616 +0000
+++ mpfr-4.2.0-b/src/mpfr.h	2023-05-17 17:17:28.596360199 +0000
@@ -27,7 +27,7 @@
 #define MPFR_VERSION_MAJOR 4
 #define MPFR_VERSION_MINOR 2
 #define MPFR_VERSION_PATCHLEVEL 0
-#define MPFR_VERSION_STRING "4.2.0-p7"
+#define MPFR_VERSION_STRING "4.2.0-p8"
 
 /* User macros:
    MPFR_USE_FILE:        Define it to make MPFR define functions dealing
diff -Naurd mpfr-4.2.0-a/src/version.c mpfr-4.2.0-b/src/version.c
--- mpfr-4.2.0-a/src/version.c	2023-05-12 15:08:39.325546612 +0000
+++ mpfr-4.2.0-b/src/version.c	2023-05-17 17:17:28.600360192 +0000
@@ -25,5 +25,5 @@
 const char *
 mpfr_get_version (void)
 {
-  return "4.2.0-p7";
+  return "4.2.0-p8";
 }
diff -Naurd mpfr-4.2.0-a/tests/tcompound.c mpfr-4.2.0-b/tests/tcompound.c
--- mpfr-4.2.0-a/tests/tcompound.c	2023-01-05 17:09:48.000000000 +0000
+++ mpfr-4.2.0-b/tests/tcompound.c	2023-05-17 17:17:28.588360213 +0000
@@ -238,18 +238,263 @@
   mpfr_clear (y);
 }
 
-static int
-mpfr_compound2 (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t rnd_mode)
+/* Failure with mpfr_compound_si from 2021-02-15 due to
+   incorrect underflow detection. */
+static void
+bug_20230206 (void)
 {
-  return mpfr_compound_si (y, x, 2, rnd_mode);
+  if (MPFR_PREC_MIN == 1)
+    {
+      mpfr_t x, y1, y2;
+      int inex1, inex2;
+      mpfr_flags_t flags1, flags2;
+#if MPFR_PREC_BITS >= 64
+      mpfr_exp_t emin;
+#endif
+
+      mpfr_inits2 (1, x, y1, y2, (mpfr_ptr) 0);
+      mpfr_set_ui_2exp (x, 1, -1, MPFR_RNDN);  /* x = 1/2 */
+
+      /* This first test is useful mainly for a 32-bit mpfr_exp_t type
+         (no failure with a 64-bit mpfr_exp_t type since the underflow
+         threshold in the extended exponent range is much lower). */
+
+      mpfr_set_ui_2exp (y1, 1, -1072124363, MPFR_RNDN);
+      inex1 = -1;
+      flags1 = MPFR_FLAGS_INEXACT;
+      mpfr_clear_flags ();
+      /* -1832808704 ~= -2^30 / log2(3/2) */
+      inex2 = mpfr_compound_si (y2, x, -1832808704, MPFR_RNDN);
+      flags2 = __gmpfr_flags;
+      if (!(mpfr_equal_p (y1, y2) &&
+            SAME_SIGN (inex1, inex2) &&
+            flags1 == flags2))
+        {
+          printf ("Error in bug_20230206 (1):\n");
+          printf ("Expected ");
+          mpfr_dump (y1);
+          printf ("  with inex = %d, flags =", inex1);
+          flags_out (flags1);
+          printf ("Got      ");
+          mpfr_dump (y2);
+          printf ("  with inex = %d, flags =", inex2);
+          flags_out (flags2);
+          exit (1);
+        }
+
+      /* This second test is for a 64-bit mpfr_exp_t type
+         (it is disabled with a 32-bit mpfr_exp_t type). */
+
+      /* The "#if" makes sure that 64-bit constants are supported, avoiding
+         a compilation failure. The "if" makes sure that the constant is
+         representable in a long (this would not be the case with 32-bit
+         unsigned long and 64-bit limb). It also ensures that mpfr_exp_t
+         has at least 64 bits. */
+#if MPFR_PREC_BITS >= 64
+      emin = mpfr_get_emin ();
+      set_emin (MPFR_EMIN_MIN);
+      mpfr_set_ui_2exp (y1, 1, -4611686018427366846, MPFR_RNDN);
+      inex1 = 1;
+      flags1 = MPFR_FLAGS_INEXACT;
+      mpfr_clear_flags ();
+      /* -7883729320669216768 ~= -2^62 / log2(3/2) */
+      inex2 = mpfr_compound_si (y2, x, -7883729320669216768, MPFR_RNDN);
+      flags2 = __gmpfr_flags;
+      if (!(mpfr_equal_p (y1, y2) &&
+            SAME_SIGN (inex1, inex2) &&
+            flags1 == flags2))
+        {
+          printf ("Error in bug_20230206 (2):\n");
+          printf ("Expected ");
+          mpfr_dump (y1);
+          printf ("  with inex = %d, flags =", inex1);
+          flags_out (flags1);
+          printf ("Got      ");
+          mpfr_dump (y2);
+          printf ("  with inex = %d, flags =", inex2);
+          flags_out (flags2);
+          exit (1);
+        }
+      set_emin (emin);
+#endif
+
+      mpfr_clears (x, y1, y2, (mpfr_ptr) 0);
+    }
 }
 
+/* Reported by Patrick Pelissier on 2023-02-11 for the master branch
+   (tgeneric_ui.c with GMP_CHECK_RANDOMIZE=1412991715).
+   On a 32-bit host, one gets Inf (overflow) instead of 0.1E1071805703.
+*/
+static void
+bug_20230211 (void)
+{
+  mpfr_t x, y1, y2;
+  int inex1, inex2;
+  mpfr_flags_t flags1, flags2;
+
+  mpfr_inits2 (1, x, y1, y2, (mpfr_ptr) 0);
+  mpfr_set_ui_2exp (x, 1, -1, MPFR_RNDN);  /* x = 1/2 */
+  mpfr_set_ui_2exp (y1, 1, 1071805702, MPFR_RNDN);
+  inex1 = 1;
+  flags1 = MPFR_FLAGS_INEXACT;
+  mpfr_clear_flags ();
+  inex2 = mpfr_compound_si (y2, x, 1832263949, MPFR_RNDN);
+  flags2 = __gmpfr_flags;
+  if (!(mpfr_equal_p (y1, y2) &&
+        SAME_SIGN (inex1, inex2) &&
+        flags1 == flags2))
+    {
+      printf ("Error in bug_20230211:\n");
+      printf ("Expected ");
+      mpfr_dump (y1);
+      printf ("  with inex = %d, flags =", inex1);
+      flags_out (flags1);
+      printf ("Got      ");
+      mpfr_dump (y2);
+      printf ("  with inex = %d, flags =", inex2);
+      flags_out (flags2);
+      exit (1);
+    }
+  mpfr_clears (x, y1, y2, (mpfr_ptr) 0);
+}
+
+/* Integer overflow with compound.c d04caeae04c6a83276916c4fbac1fe9b0cec3c8b
+   (2023-02-23) or 952fb0f5cc2df1fffde3eb54c462fdae5f123ea6 in the 4.2 branch
+   on "n * (kx - 1) + 1". Note: if the only effect is just a random value,
+   this probably doesn't affect the result (one might enter the "if" while
+   one shouldn't, but the real check is done inside the "if"). This test
+   fails if -fsanitize=undefined -fno-sanitize-recover is used or if the
+   processor emits a signal in case of integer overflow.
+   This test has been made obsolete by the "kx < ex" condition
+   in 2cb3123891dd46fe0258d4aec7f8655b8ec69aaf (master branch)
+   or f5cb40571bc3d1559f05b230cf4ffecaf0952852 (4.2 branch). */
+static void
+bug_20230517 (void)
+{
+  mpfr_exp_t old_emax;
+  mpfr_t x;
+
+  old_emax = mpfr_get_emax ();
+  set_emax (MPFR_EMAX_MAX);
+
+  mpfr_init2 (x, 123456);
+  mpfr_set_ui (x, 65536, MPFR_RNDN);
+  mpfr_nextabove (x);
+  mpfr_compound_si (x, x, LONG_MAX >> 16, MPFR_RNDN);
+  mpfr_clear (x);
+
+  set_emax (old_emax);
+}
+
+/* Inverse function on non-special cases...
+   One has x = (1+y)^n with y > -1 and x > 0. Thus y = x^(1/n) - 1.
+   The inverse function is useful
+     - to build and check hard-to-round cases (see bad_cases() in tests.c);
+     - to test the behavior close to the overflow and underflow thresholds.
+   The case x = 0 actually needs to be handled as it may occur with
+   bad_cases() due to rounding.
+*/
 static int
-mpfr_compound3 (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t rnd_mode)
+inv_compound (mpfr_ptr y, mpfr_srcptr x, long n, mpfr_rnd_t rnd_mode)
 {
-  return mpfr_compound_si (y, x, 3, rnd_mode);
+  mpfr_t t;
+  int inexact;
+  mpfr_prec_t precy, prect;
+  MPFR_ZIV_DECL (loop);
+  MPFR_SAVE_EXPO_DECL (expo);
+
+  MPFR_ASSERTN (n != 0);
+
+  if (MPFR_UNLIKELY (MPFR_IS_ZERO (x)))
+    {
+      if (n > 0)
+        return mpfr_set_si (y, -1, rnd_mode);
+      else
+        {
+          MPFR_SET_INF (y);
+          MPFR_SET_POS (y);
+          MPFR_RET (0);
+        }
+    }
+
+  MPFR_SAVE_EXPO_MARK (expo);
+
+  if (mpfr_equal_p (x, __gmpfr_one))
+    {
+      MPFR_SAVE_EXPO_FREE (expo);
+      mpfr_set_zero (y, 1);
+      MPFR_RET (0);
+    }
+
+  precy = MPFR_GET_PREC (y);
+  prect = precy + 20;
+  mpfr_init2 (t, prect);
+
+  MPFR_ZIV_INIT (loop, prect);
+  for (;;)
+    {
+      mpfr_exp_t expt1, expt2, err;
+      unsigned int inext;
+
+      if (mpfr_rootn_si (t, x, n, MPFR_RNDN) == 0)
+        {
+          /* With a huge t, this case would yield inext != 0 and a
+             MPFR_CAN_ROUND failure until a huge precision is reached
+             (as the result is very close to an exact point). Fortunately,
+             since t is exact, we can obtain the correctly rounded result
+             by doing the second operation to the target precision directly.
+          */
+          inexact = mpfr_sub_ui (y, t, 1, rnd_mode);
+          goto end;
+        }
+      expt1 = MPFR_GET_EXP (t);
+      /* |error| <= 2^(expt1-prect-1) */
+      inext = mpfr_sub_ui (t, t, 1, MPFR_RNDN);
+      if (MPFR_UNLIKELY (MPFR_IS_ZERO (t)))
+        goto cont;  /* cannot round yet */
+      expt2 = MPFR_GET_EXP (t);
+      err = 1;
+      if (expt2 < expt1)
+        err += expt1 - expt2;
+      /* |error(rootn)| <= 2^(err+expt2-prect-2)
+         and if mpfr_sub_ui is inexact:
+         |error| <= 2^(err+expt2-prect-2) + 2^(expt2-prect-1)
+                 <= (2^(err-1) + 1) * 2^(expt2-prect-1)
+                 <= 2^((err+1)+expt2-prect-2) */
+      if (inext)
+        err++;
+      /* |error| <= 2^(err+expt2-prect-2) */
+      if (MPFR_CAN_ROUND (t, prect + 2 - err, precy, rnd_mode))
+        break;
+
+    cont:
+      MPFR_ZIV_NEXT (loop, prect);
+      mpfr_set_prec (t, prect);
+    }
+
+  inexact = mpfr_set (y, t, rnd_mode);
+
+ end:
+  MPFR_ZIV_FREE (loop);
+  mpfr_clear (t);
+  MPFR_SAVE_EXPO_FREE (expo);
+  return mpfr_check_range (y, inexact, rnd_mode);
 }
 
+#define DEFN(N)                                                         \
+  static int mpfr_compound##N (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t r) \
+  { return mpfr_compound_si (y, x, N, r); }                             \
+  static int inv_compound##N (mpfr_ptr y, mpfr_srcptr x, mpfr_rnd_t r)  \
+  { return inv_compound (y, x, N, r); }
+
+DEFN(2)
+DEFN(3)
+DEFN(4)
+DEFN(5)
+DEFN(17)
+DEFN(120)
+
 #define TEST_FUNCTION mpfr_compound2
 #define test_generic test_generic_compound2
 #include "tgeneric.c"
@@ -258,17 +503,55 @@
 #define test_generic test_generic_compound3
 #include "tgeneric.c"
 
+#define TEST_FUNCTION mpfr_compound4
+#define test_generic test_generic_compound4
+#include "tgeneric.c"
+
+#define TEST_FUNCTION mpfr_compound5
+#define test_generic test_generic_compound5
+#include "tgeneric.c"
+
+#define TEST_FUNCTION mpfr_compound17
+#define test_generic test_generic_compound17
+#include "tgeneric.c"
+
+#define TEST_FUNCTION mpfr_compound120
+#define test_generic test_generic_compound120
+#include "tgeneric.c"
+
 int
 main (void)
 {
   tests_start_mpfr ();
 
   check_ieee754 ();
+  bug_20230206 ();
+  bug_20230211 ();
+  bug_20230517 ();
 
   test_generic_si (MPFR_PREC_MIN, 100, 100);
 
   test_generic_compound2 (MPFR_PREC_MIN, 100, 100);
   test_generic_compound3 (MPFR_PREC_MIN, 100, 100);
+  test_generic_compound4 (MPFR_PREC_MIN, 100, 100);
+  test_generic_compound5 (MPFR_PREC_MIN, 100, 100);
+  test_generic_compound17 (MPFR_PREC_MIN, 100, 100);
+  test_generic_compound120 (MPFR_PREC_MIN, 100, 100);
+
+  /* Note: For small n, we need a psup high enough to avoid too many
+     "f exact while f^(-1) inexact" occurrences in bad_cases(). */
+  bad_cases (mpfr_compound2, inv_compound2, "mpfr_compound2",
+             0, -256, 255, 4, 128, 240, 40);
+  bad_cases (mpfr_compound3, inv_compound3, "mpfr_compound3",
+             0, -256, 255, 4, 128, 120, 40);
+  bad_cases (mpfr_compound4, inv_compound4, "mpfr_compound4",
+             0, -256, 255, 4, 128, 80, 40);
+  bad_cases (mpfr_compound5, inv_compound5, "mpfr_compound5",
+             0, -256, 255, 4, 128, 80, 40);
+  bad_cases (mpfr_compound17, inv_compound17, "mpfr_compound17",
+             0, -256, 255, 4, 128, 80, 40);
+  bad_cases (mpfr_compound120, inv_compound120, "mpfr_compound120",
+             0, -256, 255, 4, 128, 80, 40);
 
   tests_end_mpfr ();
   return 0;
diff -Naurd mpfr-4.2.0-a/tests/tests.c mpfr-4.2.0-b/tests/tests.c
--- mpfr-4.2.0-a/tests/tests.c	2023-01-05 17:09:48.000000000 +0000
+++ mpfr-4.2.0-b/tests/tests.c	2023-05-17 17:17:28.588360213 +0000
@@ -1086,10 +1086,9 @@
                 }
               if (inex_inv)
                 {
-                  printf ("bad_cases: f exact while f^(-1) inexact,\n"
-                          "due to a poor choice of the parameters.\n");
-                  exit (1);
-                  /* alternatively, goto next_i */
+                  if (dbg)
+                    printf ("bad_cases: f exact while f^(-1) inexact\n");
+                  goto does_not_match;
                 }
               inex = 0;
               break;
@@ -1112,6 +1111,10 @@
           if (mpfr_nanflag_p () || mpfr_overflow_p () || mpfr_underflow_p ()
               || ! mpfr_equal_p (z, y))
             {
+              /* This may occur when psup is not large enough: evaluating
+                 x = (f^(-1))(y) then z = f(x) may not give back y if the
+                 precision of x is too small. */
+            does_not_match:
               if (dbg)
                 {
                   printf ("bad_cases: inverse doesn't match for %s\ny = ",