Re: [PATCH 12/12] git-zlib: widen `git_deflate_bound()` to `size_t`
- From
Johannes Schindelin <johannes.schindelin@gmx.de>
- Date
- Aug 5, 2026, 13:58 UTC
- Message-ID
- <7fe7ca44-f3c4-7e99-b7fe-81e8467e294e@gmx.de>
- In-Reply-To
- <anMBBW_arzuri4Qo@pks.im>
Hi Patrick,
On Wed, 5 Aug 2026, Patrick Steinhardt wrote:
Show 11 quoted lines
> On Thu, Jul 09, 2026 at 04:49:39PM +0000, Johannes Schindelin via GitGitGadget wrote: > > From: Johannes Schindelin <johannes.schindelin@gmx.de> > > > > All four `unsigned long`/`int`/`ssize_t` receivers across archive-zip, > > diff, http-push and t/helper/test-pack-deltas were widened to `size_t` > > in the prior commits, and remote-curl and fast-import were already > > there. With every caller prepared, both the parameter and the return > > type can now move without introducing any silent narrowing. > > Nit, feel free to ignore: I feel like all of these patches could've been > squashed into a single one, as they're trivial enough.
I like them trivial and small ;-)
Show 8 quoted lines
> > For inputs above zlib's `uLong` range (i.e. >4 GiB on platforms where > > `uLong` is 32-bit, notably 64-bit Windows), defer to zlib's stored-block > > formula (the same fallback it would itself use for an unknown stream > > state) plus the worst-case wrapper overhead. The existing path through > > `deflateBound()` is unchanged for inputs that fit. > > A link or something like that to the formula would've helped here, as > I'm not familiar with this mechanism.
Right. I added two references to the commit message.
Show 28 quoted lines
>
> > diff --git a/git-zlib.c b/git-zlib.c
> > index d21adb3bf5..ebbbcc6d1a 100644
> > --- a/git-zlib.c
> > +++ b/git-zlib.c
> > @@ -167,9 +167,21 @@ int git_inflate(git_zstream *strm, int flush)
> > return status;
> > }
> >
> > -unsigned long git_deflate_bound(git_zstream *strm, unsigned long size)
> > +size_t git_deflate_bound(git_zstream *strm, size_t size)
> > {
> > - return deflateBound(&strm->z, size);
> > +#if SIZE_MAX > ULONG_MAX
> > + if (size > maximum_unsigned_value_of_type(uLong))
> > + /*
> > + * deflateBound() takes uLong, which is 32-bit on
> > + * Windows. For inputs above that range, return zlib's
> > + * stored-block formula (the conservative path it would
> > + * itself use for an unknown stream state) plus the
> > + * worst-case wrapper overhead.
> > + */
> > + return size + (size >> 5) + (size >> 7) + (size >> 11)
> > + + 7 + 18;
> > +#endif
>
> So is the idea here that we estimate the highest number of bytes that
> the deflated size could end up with?Precisely. And the formula in zlib is a bit complex, it calculates a "fixedlen" and a "storelen" for two different ways to represent the worst case size. But for large values, only `storelen` matters, therefore we can get away with a much simpler logic here.
FWIW zlib v1.3.2 added `deflateBound_z()`, which accepts `size_t` (or more precisely: `z_size_t`). However, v1.3.2 is only 7 months old, so I'll be retired by the time Debian stable gets it :-P
Ciao, Johannes
> > Patrick >