X-Git-Url: https://plomlompom.com/repos/?a=blobdiff_plain;f=src%2Freadwrite.c;h=e776ec2822636218908d237950ddb0939fa5bf2f;hb=823357b90906df451402358eda7b995143922496;hp=039dadea4a6855e1f2222241193897db4f987b4c;hpb=4d9b6ac142740aa337fc3e622902c9595c36f6d0;p=plomrogue diff --git a/src/readwrite.c b/src/readwrite.c index 039dade..e776ec2 100644 --- a/src/readwrite.c +++ b/src/readwrite.c @@ -6,7 +6,15 @@ -/* Read/write "x" from/to "file" as bigendian representation of "size" bits. */ +/* Read/write "x" from/to "file" as bigendian representation of "size" bits. + * Only multiples of 8 allowed for "size". On failure, return 1, else 0. + * + * As of of now, all other read/write functions build on top of these. TODO: + * Consider externing these so-far internal functions and dropping the + * interfaces to them, instead relying on their internal validity checks. + * (Usage of padded memory copies instead of directly manipulating * x as is + * done in read_uint* would need to be added, though.) + */ static uint8_t read_uintX_bigendian(FILE * file, uint32_t * x, uint8_t size); static uint8_t write_uintX_bigendian(FILE * file, uint32_t x, uint8_t size); @@ -14,10 +22,10 @@ static uint8_t write_uintX_bigendian(FILE * file, uint32_t x, uint8_t size); static uint8_t read_uintX_bigendian(FILE * file, uint32_t * x, uint8_t size) { - if (0 != size % 8) - { - return 1; - } + if (0 != size % 8) /* This bit number validity check is redundant */ + { /* as long as this function is only available */ + return 1; /* through extern interfaces pre-defining the */ + } /* bit number. TODO: Consider dropping it. */ int16_t bitshift = size - 8; * x = 0; @@ -38,10 +46,10 @@ static uint8_t read_uintX_bigendian(FILE * file, uint32_t * x, uint8_t size) static uint8_t write_uintX_bigendian(FILE * file, uint32_t x, uint8_t size) { - if (0 != size % 8) - { - return 1; - } + if (0 != size % 8) /* See comment */ + { /* on identical */ + return 1; /* code block in */ + } /* read_uintX_bigendian(). */ int16_t bitshift = size - 8; for (; bitshift >= 0; bitshift = bitshift - 8) @@ -58,6 +66,9 @@ static uint8_t write_uintX_bigendian(FILE * file, uint32_t x, uint8_t size) extern uint8_t read_uint8(FILE * file, uint8_t * x) { + /* Since read_uintX_bigendian() works on -- and zeroes -- four bytes, direct + * work on values of fewer bytes would corrupt immediate neighbor values. + */ uint32_t y = * x; uint8_t err = read_uintX_bigendian(file, &y, 8); * x = (uint8_t) y; @@ -68,6 +79,7 @@ extern uint8_t read_uint8(FILE * file, uint8_t * x) extern uint8_t read_uint16_bigendian(FILE * file, uint16_t * x) { + /* See read_uint8() introductory comment for rationale. */ uint32_t y = * x; uint8_t err = read_uintX_bigendian(file, &y, 16); * x = (uint16_t) y;