Re: [PATCH v2 03/10] reftable/basics: provide new `reftable_buf` interface

From: Taylor Blau <me@ttaylorr.com>
Date: Oct 15, 2024, 19:27 UTC
Message-ID: <Zw7CIRoXrLr5pd5A@nand.local>
In-Reply-To: <CAPig+cSXJaETg4Sq3Zw8=37M15CdLEUkdMb2U2X4NPSOAFwmyw@mail.gmail.com>

On Tue, Oct 15, 2024 at 01:10:59AM -0400, Eric Sunshine wrote:

Show 18 quoted lines

> On Tue, Oct 15, 2024 at 12:38 AM Patrick Steinhardt <ps@pks.im> wrote:
> > On Mon, Oct 14, 2024 at 06:34:55PM -0400, Taylor Blau wrote:
> > > On Mon, Oct 14, 2024 at 03:02:24PM +0200, Patrick Steinhardt wrote:
> > > > +/*
> > > > + * Add the given bytes to the buffer. Returns 0 on success,
> > > > + * REFTABLE_OUT_OF_MEMORY_ERROR on allocation failure.
> > > > + */
> > > > +int reftable_buf_add(struct reftable_buf *buf, const void *data, size_t len);
> > >
> > > Is there a reason that data is a void-pointer here and not a const char
> > > *?
> >
> > Only that it emulates `strbuf_add()`, which also uses a void pointer.
>
> The reason for that is because strbuf is a generic byte-array which
> may contain embedded NULs, and the `const void *` plus `len`
> emphasizes this property, whereas `const char *` would imply a
> C-string with no embedded NULs.

Thanks, that was the explanation I was missing. Perhaps it is worth re-stating in the commit message here to avoid confusing readers like I was when I first read Patrick's patch ;-).

Thanks, Taylor