| commit | 2571400d7e151568ce5fbe896c5a1be1755de327 | |
| author | Jake Bailey (Hacklang) <jakebailey@fb.com> | |
| Fri, 18 Jan 2019 18:36:27 +0000 (18 10:36 -0800) | ||
| committer | Hhvm Bot <hhvm-bot@users.noreply.github.com> | |
| Fri, 18 Jan 2019 18:45:31 +0000 (18 10:45 -0800) | ||
| tree | 12b6f1381af11cce4e132d1465f04c0f52f8263f | treesnapshot (tar.gz zip) |
| parent | 593cc29d0217983e2f2a781eaa705ee6db3fbb62 | commitdiff |
Rename write_through to write_around in SharedMem.WithCache
Summary: This `write_through` function was added in {D3023663}, but I think it is a bit of a misnomer. The strategy of writing directly to a backing store without writing to a cache layer in front of it is called write-around, not write-through. The `add` function implements the write-through strategy, since it synchronously writes to both the cache and the backing store before allowing the caller to continue execution (making the existence of an alternative named `write_through` all the more confusing).
Reviewed By: pittsw
Differential Revision: D13659733
fbshipit-source-id: 1254fb941ab89e3b7785d1ec46954fe8c427337f
Summary: This `write_through` function was added in {D3023663}, but I think it is a bit of a misnomer. The strategy of writing directly to a backing store without writing to a cache layer in front of it is called write-around, not write-through. The `add` function implements the write-through strategy, since it synchronously writes to both the cache and the backing store before allowing the caller to continue execution (making the existence of an alternative named `write_through` all the more confusing).
Reviewed By: pittsw
Differential Revision: D13659733
fbshipit-source-id: 1254fb941ab89e3b7785d1ec46954fe8c427337f