Merge pull request #536 from libtom/some-improvements

Some improvements

CMakeLists.txt

@@ -61,7 +61,7 @@ endif()
 # What compiler do we have and what are their...uhm... peculiarities
 if(CMAKE_C_COMPILER_ID MATCHES "(C|c?)lang")
     list(APPEND LTM_C_FLAGS -Wno-typedef-redefinition -Wno-tautological-compare -Wno-builtin-requires-header)
-    # Clang requires at least '-O1' for dead code eliminiation
+    # Clang requires at least '-O1' for dead code elimination
     set(CMAKE_C_FLAGS_DEBUG "-O1 ${CMAKE_C_FLAGS_DEBUG}")
 endif()
 if(CMAKE_C_COMPILER MATCHES "mingw")
@@ -103,11 +103,17 @@ target_link_options(${PROJECT_NAME} BEFORE PRIVATE
     ${LTM_LD_FLAGS}
 )
 
+set(PUBLIC_HEADERS tommath.h)
+set(C89 False CACHE BOOL "(Usually maintained automatically) Enable when the library is in c89 mode to package the correct header files on install")
+if(C89)
+    list(APPEND PUBLIC_HEADERS tommath_c89.h)
+endif()
+
 set_target_properties(${PROJECT_NAME} PROPERTIES
     OUTPUT_NAME tommath
     VERSION ${PROJECT_VERSION}
     SOVERSION ${PROJECT_VERSION_MAJOR}
-    PUBLIC_HEADER tommath.h
+    PUBLIC_HEADER "${PUBLIC_HEADERS}"
 )
 
 option(COMPILE_LTO "Build with LTO enabled")

astylerc

@@ -4,7 +4,7 @@
 # usage:
 #       astyle --options=astylerc *.[ch]
 
-# Do not create backup, annonying in the times of git
+# Do not create backup, annoying in the times of git
 suffix=none
 
 ## Bracket Style Options

changes.txt

@@ -19,7 +19,7 @@ v1.2.0
        -- Unified, safer and improved API's
        -- Less magic numbers - return values (where appropriate) and most flags are now enums,
           this was implemented in a backwards compatible way where return values were int.
-       -- API's with return values are now by default marked as "warn on unsused result", this
+       -- API's with return values are now by default marked as "warn on unused result", this
           can be disabled if required (which will most likely hide bugs), c.f. MP_WUR in tommath.h
        -- Provide a whole set of setters&getters for different primitive types (long, uint32_t, etc.)
        -- All those primitive setters are now optimized.
@@ -412,8 +412,8 @@ v0.13  -- tons of minor speed-ups in low level add, sub, mul_2 and div_2 which p
 Jan 17th, 2003
 v0.12  -- re-wrote the majority of the makefile so its more portable and will
           install via "make install" on most *nix platforms
-       -- Re-packaged all the source as seperate files.  Means the library a single
-          file packagage any more.  Instead of just adding "bn.c" you have to add
+       -- Re-packaged all the source as separate files.  Means the library a single
+          file package any more.  Instead of just adding "bn.c" you have to add
           libtommath.a
        -- Renamed "bn.h" to "tommath.h"
        -- Changes to the manual to reflect all of this
@@ -444,7 +444,7 @@ v0.08  -- Sped up the multipliers by moving the inner loop variables into a smal
        -- add etc/pprime.c program which makes numbers which are provably prime.
 
 Jan 1st, 2003
-v0.07  -- Removed alot of heap operations from core functions to speed them up
+v0.07  -- Removed a lot of heap operations from core functions to speed them up
        -- Added a root finding function [and mp_sqrt macro like from MPI]
        -- Added more to manual
 

demo/CMakeLists.txt

@@ -61,3 +61,6 @@ target_link_options(${LTM_TEST} BEFORE PUBLIC
 # CTest
 #-----------------------------------------------------------------------------
 add_test(NAME ${LTM_TEST} COMMAND ${LTM_TEST})
+
+find_program(MEMORYCHECK_COMMAND valgrind)
+set(MEMORYCHECK_COMMAND_OPTIONS "--trace-children=yes --leak-check=full")

demo/mtest_opponent.c

@@ -82,7 +82,7 @@ static int mtest_opponent(void)
 #endif
 
    for (;;) {
-      /* randomly clear and re-init one variable, this has the affect of triming the alloc space */
+      /* randomly clear and re-init one variable, this has the effect of trimming the alloc space */
       switch (abs(rand()) % 7) {
       case 0:
          mp_clear(&a);

demo/test.c

@@ -1548,7 +1548,7 @@ static int test_mp_decr(void)
    default(realprecision,1000);
    for(n=3,100,r = floor(a^(1/n));printf("\"" r "\", "))
 
-   All numbers as strings to simplifiy things, especially for the
+   All numbers as strings to simplify things, especially for the
    low-mp branch.
 */
 
@@ -2166,7 +2166,7 @@ static int test_mp_pack_unpack(void)
    uint8_t *buf = NULL;
 
    mp_order order = MP_LSB_FIRST;
-   mp_endian endianess = MP_NATIVE_ENDIAN;
+   mp_endian endianness = MP_NATIVE_ENDIAN;
 
    DOR(mp_init_multi(&a, &b, NULL));
    DO(mp_rand(&a, 15));
@@ -2180,9 +2180,9 @@ static int test_mp_pack_unpack(void)
    }
 
    DO(mp_pack((void *)buf, count, &written, order, 1uL,
-              endianess, 0uL, &a));
+              endianness, 0uL, &a));
    DO(mp_unpack(&b, count, order, 1uL,
-                endianess, 0uL, (const void *)buf));
+                endianness, 0uL, (const void *)buf));
 
    if (mp_cmp(&a, &b) != MP_EQ) {
       fprintf(stderr, "pack/unpack cycle failed\n");

doc/bn.tex

@@ -210,7 +210,7 @@ \subsubsection{OpenBSD}
 At this time two versions of \texttt{libtool} are installed and both are named \texttt{libtool},
 unfortunately but GNU \texttt{libtool} has been placed in \texttt{/usr/local/bin/} and the native
 version in \texttt{/usr/bin/}. The path might be different in other versions of OpenBSD but both
-programms differ in the output of \texttt{libtool --version}
+programs differ in the output of \texttt{libtool --version}
 \begin{alltt}
 $ /usr/local/bin/libtool --version
 libtool (GNU libtool) 2.4.2
@@ -519,7 +519,7 @@ \subsection{Single Initialization}
 \end{alltt}
 
 This function expects a pointer to an \texttt{mp\_int} structure and will initialize the members
-ofthe structure so the \texttt{mp\_int} represents the default integer which is zero.  If the
+of the structure so the \texttt{mp\_int} represents the default integer which is zero.  If the
 functions returns \texttt{MP\_OKAY} then the \texttt{mp\_int} is ready to be used by the other
 LibTomMath functions.
 
@@ -2213,7 +2213,7 @@ \subsection{Required Number of Tests}
 Determining the probability needed to pick the right column is a bit harder. Fips 186.4, for
 example has $2^{-80}$ for $512$ bit large numbers, $2^{-112}$ for $1024$ bits, and $2^{128}$ for
 $1536$ bits. It can be seen in table \ref{table:millerrabinrunsp1} that those combinations follow
-the diagonal from $(512,2^{-80})$ downwards and to the right to gain a lower probabilty of getting
+the diagonal from $(512,2^{-80})$ downwards and to the right to gain a lower probability of getting
 a composite declared a pseudoprime for the same amount of work or less.
 
 If this version of the library has the strong Lucas--Selfridge and/or the Frobenius--Underwood test

doc/makefile

@@ -40,7 +40,7 @@ manual:	mandvi
 
 # The file latexindent.pl is in several LaTeX distributions, if not:
 # https://ctan.org/pkg/latexindent
-# Its configuraion is well documented
+# Its configuration is well documented
 # http://mirrors.ctan.org/support/latexindent/documentation/latexindent.pdf
 pretty:
 	latexindent -s -w -m -l=.latexindent.yaml bn.tex

logs/before_after.dem

@@ -23,8 +23,8 @@ plot 'sqr-before.log' smooth bezier title "Squaring (without Karatsuba) (before)
 set output "expt-ba.png"
 plot 'expt-before.log' smooth bezier title "Exptmod (Montgomery) (before)", \
       'expt-after.log' smooth bezier title "Exptmod (Montgomery) (after)", \
-      'expt_dr-before.log' smooth bezier title "Exptmod (Dimminished Radix) (before)", \
-      'expt_dr-after.log' smooth bezier title "Exptmod (Dimminished Radix) (after)", \
+      'expt_dr-before.log' smooth bezier title "Exptmod (Diminished Radix) (before)", \
+      'expt_dr-after.log' smooth bezier title "Exptmod (Diminished Radix) (after)", \
       'expt_2k-before.log' smooth bezier title "Exptmod (2k Reduction) (before)", \
       'expt_2k-after.log' smooth bezier title "Exptmod (2k Reduction) (after)", \
       'expt_2kl-before.log' smooth bezier title "Exptmod (2k-l Reduction) (before)", \

logs/graphs.dem

@@ -9,7 +9,7 @@ set output "mult.png"
 plot 'sqr.log' smooth bezier title "Squaring (without Karatsuba)", 'sqr_kara.log' smooth bezier title "Squaring (Karatsuba)", 'mult.log' smooth bezier title "Multiplication (without Karatsuba)", 'mult_kara.log' smooth bezier title "Multiplication (Karatsuba)"
 
 set output "expt.png"
-plot 'expt.log' smooth bezier title "Exptmod (Montgomery)", 'expt_dr.log' smooth bezier title "Exptmod (Dimminished Radix)", 'expt_2k.log' smooth bezier title "Exptmod (2k Reduction)"
+plot 'expt.log' smooth bezier title "Exptmod (Montgomery)", 'expt_dr.log' smooth bezier title "Exptmod (Diminished Radix)", 'expt_2k.log' smooth bezier title "Exptmod (2k Reduction)"
 
 set output "invmod.png"
 plot 'invmod.log' smooth bezier title "Modular Inverse"

makefile

@@ -130,7 +130,7 @@ zipup: clean astyle new_file docs
 	@#   As the pdf creation modifies the tex files, git sometimes detects the
 	@#   modified files, but misses that it's put back to its original version.
 	@git update-index --refresh
-	@git diff-index --quiet HEAD -- || ( echo "FAILURE: uncommited changes or not a git" && exit 1 )
+	@git diff-index --quiet HEAD -- || ( echo "FAILURE: uncommitted changes or not a git" && exit 1 )
 	rm -rf libtommath-$(VERSION) ltm-$(VERSION).*
 	@# files/dirs excluded from "git archive" are defined in .gitattributes
 	git archive --format=tar --prefix=libtommath-$(VERSION)/ HEAD | tar x
@@ -173,7 +173,8 @@ c89:
 	-e 's/uint\([0-9][0-9]*\)_t/mp_u\1/g' \
 	-e 's/int\([0-9][0-9]*\)_t/mp_i\1/g' \
 	-e 's/__func__/MP_FUNCTION_NAME/g' \
-	*.c tommath.h tommath_private.h demo/*.c demo/*.h etc/*.c
+	-e 's/set(C89 False/set(C89 True/g' \
+	*.c tommath.h tommath_private.h demo/*.c demo/*.h etc/*.c CMakeLists.txt
 
 c99:
 	@echo "Applying substitutions for c99 compatibility..."
@@ -196,7 +197,8 @@ c99:
 	-e 's/mp_u\([0-9][0-9]*\)/uint\1_t/g' \
 	-e 's/mp_i\([0-9][0-9]*\)/int\1_t/g' \
 	-e 's/MP_FUNCTION_NAME/__func__/g' \
-	*.c tommath.h tommath_private.h demo/*.c demo/*.h etc/*.c
+	-e 's/set(C89 True/set(C89 False/g' \
+	*.c tommath.h tommath_private.h demo/*.c demo/*.h etc/*.c CMakeLists.txt
 
 astyle:
 	@echo "   * run astyle on all sources"

makefile_include.mk

@@ -95,6 +95,10 @@ endif # COMPILE_SIZE
 
 ifneq ($(findstring clang,$(CC)),)
 LTM_CFLAGS += -Wno-typedef-redefinition -Wno-tautological-compare -Wno-builtin-requires-header
+ifdef IGNORE_SPEED
+#for dead code eliminiation
+LTM_CFLAGS += -O1
+endif
 endif
 ifneq ($(findstring mingw,$(CC)),)
 LTM_CFLAGS += -Wno-shadow

mp_2expt.c

@@ -15,7 +15,7 @@ mp_err mp_2expt(mp_int *a, int b)
    /* zero a as per default */
    mp_zero(a);
 
-   /* grow a to accomodate the single bit */
+   /* grow a to accommodate the single bit */
    if ((err = mp_grow(a, (b / MP_DIGIT_BIT) + 1)) != MP_OKAY) {
       return err;
    }

mp_clamp.c

@@ -6,7 +6,7 @@
 /* trim unused digits
  *
  * This is used to ensure that leading zero digits are
- * trimed and the leading "used" digit will be non-zero
+ * trimmed and the leading "used" digit will be non-zero
  * Typically very fast.  Also fixes the sign if there
  * are no more leading digits
  */

mp_cmp_mag.c

@@ -3,7 +3,7 @@
 /* LibTomMath, multiple-precision integer library -- Tom St Denis */
 /* SPDX-License-Identifier: Unlicense */
 
-/* compare maginitude of two ints (unsigned) */
+/* compare magnitude of two ints (unsigned) */
 mp_ord mp_cmp_mag(const mp_int *a, const mp_int *b)
 {
    int n;

mp_exptmod.c

@@ -5,7 +5,7 @@
 
 /* this is a shell function that calls either the normal or Montgomery
  * exptmod functions.  Originally the call to the montgomery code was
- * embedded in the normal function but that wasted alot of stack space
+ * embedded in the normal function but that wasted a lot of stack space
  * for nothing (since 99% of the time the Montgomery code would be called)
  */
 mp_err mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)

mp_lshd.c

@@ -27,7 +27,7 @@ mp_err mp_lshd(mp_int *a, int b)
    a->used += b;
 
    /* much like mp_rshd this is implemented using a sliding window
-    * except the window goes the otherway around.  Copying from
+    * except the window goes the other way around.  Copying from
     * the bottom to the top.  see mp_rshd.c for more info.
     */
    for (x = a->used; x --> b;) {

mp_montgomery_calc_normalization.c

@@ -7,7 +7,7 @@
  * shifts with subtractions when the result is greater than b.
  *
  * The method is slightly modified to shift B unconditionally upto just under
- * the leading bit of b.  This saves alot of multiple precision shifting.
+ * the leading bit of b.  This saves a lot of multiple precision shifting.
  */
 mp_err mp_montgomery_calc_normalization(mp_int *a, const mp_int *b)
 {

mp_mul_2.c

@@ -10,7 +10,7 @@ mp_err mp_mul_2(const mp_int *a, mp_int *b)
    int x, oldused;
    mp_digit r;
 
-   /* grow to accomodate result */
+   /* grow to accommodate result */
    if ((err = mp_grow(b, a->used + 1)) != MP_OKAY) {
       return err;
    }

mp_prime_is_prime.c

@@ -181,7 +181,7 @@ mp_err mp_prime_is_prime(const mp_int *a, int t, bool *result)
        * The mp_digit's have a defined bit-size but the size of the
        * array a.dp is a simple 'int' and this library can not assume full
        * compliance to the current C-standard (ISO/IEC 9899:2011) because
-       * it gets used for small embeded processors, too. Some of those MCUs
+       * it gets used for small embedded processors, too. Some of those MCUs
        * have compilers that one cannot call standard compliant by any means.
        * Hence the ugly type-fiddling in the following code.
        */
@@ -213,12 +213,12 @@ mp_err mp_prime_is_prime(const mp_int *a, int t, bool *result)
         The function mp_rand() goes to some length to use a cryptographically
         good PRNG. That also means that the chance to always get the same base
         in the loop is non-zero, although very low.
-        If the BPSW test and/or the addtional Frobenious test have been
+        If the BPSW test and/or the additional Frobenious test have been
         performed instead of just the Miller-Rabin test with the bases 2 and 3,
         a single extra test should suffice, so such a very unlikely event
         will not do much harm.
 
-        To preemptivly answer the dangling question: no, a witness does not
+        To preemptively answer the dangling question: no, a witness does not
         need to be prime.
       */
       for (ix = 0; ix < t; ix++) {

mp_reduce_setup.c

@@ -4,7 +4,7 @@
 /* SPDX-License-Identifier: Unlicense */
 
 /* pre-calculate the value required for Barrett reduction
- * For a given modulus "b" it calulates the value required in "a"
+ * For a given modulus "b" it calculates the value required in "a"
  */
 mp_err mp_reduce_setup(mp_int *a, const mp_int *b)
 {

mtest/mpi.c

@@ -418,7 +418,7 @@ void mp_exch(mp_int *mp1, mp_int *mp2)
 
   Release the storage used by an mp_int, and void its fields so that
   if someone calls mp_clear() again for the same int later, we won't
-  get tollchocked.
+  get tolchocked.
  */
 
 void   mp_clear(mp_int *mp)

s_mp_montgomery_reduce_comba.c

@@ -61,7 +61,7 @@ mp_err s_mp_montgomery_reduce_comba(mp_int *x, const mp_int *n, mp_digit rho)
       /* a = a + mu * m * b**i
        *
        * This is computed in place and on the fly.  The multiplication
-       * by b**i is handled by offseting which columns the results
+       * by b**i is handled by offsetting which columns the results
        * are added to.
        *
        * Note the comba method normally doesn't handle carries in the

s_mp_mul_comba.c

@@ -43,7 +43,7 @@ mp_err s_mp_mul_comba(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       ty = MP_MIN(b->used-1, ix);
       tx = ix - ty;
 
-      /* this is the number of times the loop will iterrate, essentially
+      /* this is the number of times the loop will iterate, essentially
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
       iy = MP_MIN(a->used-tx, ty+1);

s_mp_mul_high_comba.c

@@ -35,7 +35,7 @@ mp_err s_mp_mul_high_comba(const mp_int *a, const mp_int *b, mp_int *c, int digs
       ty = MP_MIN(b->used-1, ix);
       tx = ix - ty;
 
-      /* this is the number of times the loop will iterrate, essentially its
+      /* this is the number of times the loop will iterate, essentially its
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
       iy = MP_MIN(a->used-tx, ty+1);

s_mp_mul_karatsuba.c

@@ -27,7 +27,7 @@
  * are saved.  Note also that the call to mp_mul can end up back
  * in this function if the a0, a1, b0, or b1 are above the threshold.
  * This is known as divide-and-conquer and leads to the famous
- * O(N**lg(3)) or O(N**1.584) work which is asymptopically lower than
+ * O(N**lg(3)) or O(N**1.584) work which is asymptotically lower than
  * the standard O(N**2) that the baseline/comba methods use.
  * Generally though the overhead of this method doesn't pay off
  * until a certain size (N ~ 80) is reached.

s_mp_sqr_comba.c

@@ -39,7 +39,7 @@ mp_err s_mp_sqr_comba(const mp_int *a, mp_int *b)
       ty = MP_MIN(a->used-1, ix);
       tx = ix - ty;
 
-      /* this is the number of times the loop will iterrate, essentially
+      /* this is the number of times the loop will iterate, essentially
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
       iy = MP_MIN(a->used-tx, ty+1);