Changes:
* A new numeric escape sequence for string literals has been
introduced. A backslash followed by an integer literal and a
semicolon is interpreted as character with the specified ordinal
number. E.g.: "abc\128;def". With the old numeric escape sequence
this would be "abc\128\def". Terminating the escape sequence
with ; instead of \ improves the readability, especially when
escape sequences follow each other. Instead of "\8\\12\\7\8\9\\\"
it is now possible to write "\8;\12;\7;8\9;\\" The old escape
sequence is still supported, but it is deprecated and should not
be used. Future versions will warn about the deprecated escape
sequence.
* The conversion functions 'bytes' and 'bytes2Int' in bytedata.s7c
have been improved to support all combinations of signedness and
endianess.
* The functions bStriBe and bStriLe have been removed from
bstring.s7i. The function 'bytes' in bytedata.s7c can be used as
replacement.
* Checks for lowestSetBit and UNSIGNED conversions with the
functions 'bytes' and 'bytes2Int' have been added to chkint.sd7.
* Checks for UNSIGNED conversions with the functions 'bytes' and
'bytes2BigInt' have been added to chkbig.sd7.
* Checks for the function rand have been added to chkset.sd7.
* The parser has been improved to write error messages when the
source file contains unexpected UTF-8 continuation bytes, solitary
UTF-8 start bytes, extended control codes from the C1 set or an
UTF-16 byte order mark.
* In tar.s7i the functions getoct, putspc and putoct have been
improved. The computation of the size of padding data has also
been improved.
* The compiler has been improved to optimize string initialisations
with the function strZero. StrZero creates a string of arbitrary
length with '\0;' (zero) bytes.
* The compiler has been improved to optimize integer exponentiation
(action INT_POW). Now the multiplication pattern is determined at
compile time and the number of multiplications is reduced.
* The performance of the function setCard in set_rtl.c has been
improved by a factor of 2.5.
* The performance of the function setRand in set_rtl.c has been
improved by a factor of 2.3.
* Loop unrolling inspired by Duff's device has been introduced in
striutl.h. The unrolling is used for tight loops which copy byte
data to Seed7 strings in fil_rtl.c, soc_rtl.c and bst_rtl.c. The
performance improvement has been verified with gcc and valgrind.
* The example program err.sd7, which checks compile time error
messages, has been improved.
* The old way to store the data of constant bigIntegers has been
removed from compiler and runtime library.
* The buffers used in intRadix and intRadixPow2 (file int_rtl.c) have
been enlarged to fit for the most negative number with a base of 2.
* In itflib.c the functions itf_cpy and itf_cpy2 have been fixed to
work correctly, when the destination is converted from a struct to
an interface.
* In traceutl.c the tracing of characters has been improved in the
functions prot_char, prot_os_stri, prot_stri_unquoted and
prot_bstri.
* In ut8_rtl.c the function ut8Getc has been improved to use smarter
checks for ranges of characters.
* The deprecated variant of the function str, which has a second
parameter to specify a numeric base, has been removed. The
operators radix and RADIX can be used instead.
* Interpreter and compiler have been improved to support the new
actions INT_BYTES_BE, INT_BYTES_LE, INT_BYTES_BE_2_INT and
INT_BYTES_LE_2_INT. The functions intBytesBe, int_bytesBe,
intBytesLe, int_bytesLe, intBytesBe2Int, int_bytesBe2Int,
intBytesLe2Int and int_bytesLe2Int have been added.
* Support for the actions BIG_CLIT, INT_TO_BSTRI_BE, INT_TO_BSTRI_LE,
INT_BYTES_BE_2_UINT and INT_BYTES_LE_2_UINT has been removed from
interpreter and compiler. The functions bigImport, bigCLit,
big_clit, intToBStriBe, int_toBStriBe, intToBStriLe, int_toBStriLe,
intBytesBe2UInt, int_bytes_be_2_uint, intBytesLe2UInt and
int_bytes_le_2_uint have been removed.
140a1d1269