Pádraig Brady writes:
> To give a little more context, this will avoid
> round trip issues like the following, by failing early:
>
> $ echo "HelloWorld==" | base64 -d | base64
> HelloWorlQ==
Thanks for background and patches! There are use-cases for bad inputs
(both for good and malicious p
On 27/10/2023 08:28, Simon Josefsson wrote:
Pádraig Brady writes:
To give a little more context, this will avoid
round trip issues like the following, by failing early:
$ echo "HelloWorld==" | base64 -d | base64
HelloWorlQ==
Thanks for background and patches! There are use-cases for
Pádraig Brady writes:
> However if there are good use-cases for bad inputs
> we may need to adjust this patch,
> rather than failing unconditionally.
>
> For example we could just flag non canonical input in the context,
> and leave it up to the caller how to deal with that.
That adds complexity
On 27/10/2023 10:23, Simon Josefsson wrote:
Pádraig Brady writes:
However if there are good use-cases for bad inputs
we may need to adjust this patch,
rather than failing unconditionally.
For example we could just flag non canonical input in the context,
and leave it up to the caller how to d
Paul Eggert wrote:
> @@ -66,39 +62,39 @@ main ()
> # endif
>{ /* Quiet NaN. */
> static memory_long_double x =
> - { LDBL80_WORDS (0x, 0xC333, 0x) };
> + { .word = LDBL80_WORDS (0x, 0xC333, 0x) };
> ASSERT (isnanl (x.value));
>}
But the
These patches add a module 'fenv-rounding', that implements the ISO C 99
functions
fegetround
fesetround
Most platforms have these functions nowadays, and the unit tests have not
uncovered any bugs. Therefore it's probably only relevant for portability
to older Android versions.
The most inte
