> On 6 Feb 2025, at 02:51, Japin Li wrote:
>
> On Wed, 05 Feb 2025 at 18:17, Daniel Gustafsson wrote:
>>> On 5 Feb 2025, at 17:05, Tom Lane wrote:
>>
>>> While we're at it, could we make the adjacent "magic" string be
>>> "static const char *magic"? (Probably needs a couple more
>>> "const" m
On Wed, 05 Feb 2025 at 18:17, Daniel Gustafsson wrote:
>> On 5 Feb 2025, at 17:05, Tom Lane wrote:
>
>> While we're at it, could we make the adjacent "magic" string be
>> "static const char *magic"? (Probably needs a couple more
>> "const" modifiers at use sites, too.)
>
> Good point, from the l
> On 5 Feb 2025, at 17:05, Tom Lane wrote:
> While we're at it, could we make the adjacent "magic" string be
> "static const char *magic"? (Probably needs a couple more
> "const" modifiers at use sites, too.)
Good point, from the link referenced it's clear that FreeBSD has made that
change as w
Daniel Gustafsson writes:
> From a quick first inspection (and running the tests with the patch applied) I
> agree with this, these variables do not need to be static. I'll stare a bit
> more at this to make sure but seems like the right patch.
+1. All three of those variables are visibly assig
> On 5 Feb 2025, at 13:30, Japin Li wrote:
> When reviewing patch [1], I find that the static class specifier is
> unnecessary
> for the variables sp and ep in the function px_crypt_md5().
From a quick first inspection (and running the tests with the patch applied) I
agree with this, these vari
Hi,
When reviewing patch [1], I find that the static class specifier is unnecessary
for the variables sp and ep in the function px_crypt_md5().
diff --git a/contrib/pgcrypto/crypt-md5.c b/contrib/pgcrypto/crypt-md5.c
index d38721a1010..3d17b2340fe 100644
--- a/contrib/pgcrypto/crypt-md5.c
+++ b/
