summaryrefslogtreecommitdiff
path: root/printf
diff options
context:
space:
mode:
authorKevin Ryde <user42@zip.com.au>2002-12-10 23:29:59 +0100
committerKevin Ryde <user42@zip.com.au>2002-12-10 23:29:59 +0100
commit0cd846ea92e1dee733aeaa2ad57182d8ebc4d558 (patch)
treec5f221c98718856c6fa3a05b3375e7ea6a08c63b /printf
parentce8749e0ab1c0aaeb9dc353e3e2d61d51f7201a0 (diff)
downloadgmp-0cd846ea92e1dee733aeaa2ad57182d8ebc4d558.tar.gz
* printf/doprnt.c: Comments on "ll" versus "L".
Diffstat (limited to 'printf')
-rw-r--r--printf/doprnt.c7
1 files changed, 7 insertions, 0 deletions
diff --git a/printf/doprnt.c b/printf/doprnt.c
index c2de3a6d7..8f7c85ddf 100644
--- a/printf/doprnt.c
+++ b/printf/doprnt.c
@@ -94,6 +94,13 @@ MA 02111-1307, USA. */
"Z" was a type marker for size_t in old glibc, but there seems no need to
provide access to that now "z" is standard.
+ In GMP 4.1.1 we documented "ll" and "L" as being equivalent, but in C99
+ in fact "ll" is just for long long and "L" just for long double.
+ Apparentely GLIBC allows "L" for long long though. This doesn't affect
+ us as such, since both are passed through to the C library. To be
+ consistent with what we said before, the two are treated equivalently
+ here, and it's left to the C library to do what it thinks with them.
+
Possibilities:
"b" might be nice for binary output, and could even be supported for the