doc: fix wrong claim in Bitfield Allocation Rules.

* txr.1: Rewriting the last two sentences to fix the wrong
claim of its last two sentences, which contradicts the correct
statement in the the previous paragraph. The statements are
only correct about big-endian, not about both types of
machine.

1 files changed, 6 insertions, 9 deletions

diff --git a/txr.1 b/txr.1
index 4b9a3e80..34463d4a 100644
--- a/txr.1
+++ b/txr.1
@@ -67529,15 +67529,12 @@ position (bit 0). If another one-bit-wide is allocated, it is placed into
 bit 6 on big-endian, and bit 1 on little-endian.
 
 More generally, whenever a bitfield is allocated for a big-endian machine, and
-the storage unit is determined into which that bitfield shall be placed, that
-field will be placed at that unit's most significant bit position that is still
-available for allocation. On a little-endian machine, that field will be placed
-at the least significant available bit position of the storage unit. On either
-type of machine, that field will be placed at the lowest-addressed byte or
-bytes that are available, and within that byte or those bytes, the field
-will occupy the most significant bit positions that are available. If the
-field partially occupies a byte, the bits that remain free are the least
-significant ones.
+the storage unit is determined into which that bitfield shall be placed, the
+most significant bits of that storage unit are filled first on a big-endian
+machine, whereas the least significant bits are filled first on a little-endian
+machine.  From this it follows that on either type of machine, that field shall
+be placed at the lowest-addressed byte or bytes in which unallocated bits
+remain.
 
 .SS* FFI Call Descriptors