Re: [PATCH] Define PRISTINE.checksum as always being SHA-1

Julian Foad wrote:
> Bert and I were just discussing the checksums in PRISTINE. We think its
> primary key should be SHA-1 always. A secondary index can be built on
> the MD5 column if required.
>
> Can we patch the docco like this?

+1

There was agreement on always using SHA1 hashes in the 'pristine store
design' thread. See where Greg says "nono" in
http://mail-archives.apache.org/mod_mbox/subversion-dev/201002.mbox/%3C6cca3db31002171652u15ae7e87ha5b09e4cb170af2a@mail.gmail.com%3E

The possibility of using MD5 to index was also named in that thread, but did
not resonate well.

~Neels

>
> [[[
> * subversion/libsvn_wc/wc-metadata.sql
> (PRISTINE): Define "checksum" as always being SHA-1.
> Document the uniqueness properties of both checksum fields.
> Document when "md5_checksum" can be null.
> ]]]
>
> [[[
> Index: subversion/libsvn_wc/wc-metadata.sql
> ===================================================================
> --- subversion/libsvn_wc/wc-metadata.sql (revision 911752)
> +++ subversion/libsvn_wc/wc-metadata.sql (working copy)
> @@ -172,7 +172,9 @@
> and ACTUAL_NODE tables.
> */
> CREATE TABLE PRISTINE (
> - /* ### the hash algorithm (MD5 or SHA-1) is encoded in this value */
> + /* The SHA-1 checksum of the pristine text. This is a unique key. The
> + SHA-1 checksum of a pristine text is assumed to be unique among all
> + pristine texts referenced from this database. */
> checksum TEXT NOT NULL PRIMARY KEY,
>
> /* ### enumerated values specifying type of compression. NULL implies
> @@ -189,7 +191,8 @@
> refcount INTEGER NOT NULL,
>
> /* Alternative MD5 checksum used for communicating with older
> - repositories. */
> + repositories. Not guaranteed to be unique among table rows.
> + NULL if not (yet) calculated. */
> md5_checksum TEXT
> );
> ]]]
>
> - Julian
>
>