On Tue, Aug 11, 2009 at 21:10, Hyrum K. Wright<hyrum_at_hyrumwright.org> wrote:
>...
> +++ trunk/subversion/libsvn_client/ra.c Tue Aug 11 12:10:11 2009 (r38683)
> @@ -374,80 +374,25 @@ svn_client_uuid_from_url(const char **uu
>
>
> svn_error_t *
> -svn_client_uuid_from_path(const char **uuid,
> - const char *path,
> - svn_wc_adm_access_t *adm_access,
> - svn_client_ctx_t *ctx,
> - apr_pool_t *pool)
> +svn_client_uuid_from_path2(const char **uuid,
> + const char *local_abspath,
> + svn_client_ctx_t *ctx,
> + apr_pool_t *result_pool,
> + apr_pool_t *scratch_pool)
> {
> const svn_wc_entry_t *entry;
> - svn_boolean_t is_root;
>
> - SVN_ERR(svn_wc__entry_versioned(&entry, path, adm_access,
> - TRUE, /* show deleted */ pool));
> + SVN_ERR(svn_wc__get_entry_versioned(&entry, ctx->wc_ctx, local_abspath,
> + svn_node_unknown, TRUE, /* show deleted */
> + FALSE, scratch_pool, scratch_pool));
>
> if (entry->uuid)
Why test this?
> - {
> - *uuid = entry->uuid;
> - return SVN_NO_ERROR;
> - }
> -
> - /* ## Probably never reached after the 1.6/1.7 WC rewrite */
> -
> - SVN_ERR(svn_wc_is_wc_root(&is_root, path, adm_access, pool));
> -
> - if (!is_root)
> - {
> - /* Working copies have a single uuid, as all contents is from a single
> - repository */
> -
> - svn_error_t *err;
> - svn_wc_adm_access_t *parent_access;
> - const char *parent = svn_dirent_dirname(path, pool);
> -
> - /* Open the parents administrative area to fetch the uuid.
> - Subversion 1.0 and later have the uuid in every checkout root */
> -
> - SVN_ERR(svn_wc__adm_open_in_context(&parent_access, ctx->wc_ctx, parent,
> - FALSE, 0, ctx->cancel_func,
> - ctx->cancel_baton, pool));
> -
> - err = svn_client_uuid_from_path(uuid, svn_dirent_dirname(path, pool),
> - parent_access, ctx, pool);
> -
> - svn_error_clear(svn_wc_adm_close2(parent_access, pool));
> -
> - return svn_error_return(err);
> - }
> -
> - /* We may have a working copy without uuid */
> - if (entry->url)
> - {
> - /* You can enter this case by copying a new subdirectory with 1.0-1.5
> - * # svn mkdir newdir
> - * # cp newdir /tmp/new-wc
> - * and then check /tmp/new-wc
> - *
> - * See also:
> - * http://subversion.tigris.org/servlets/ReadMsg?list=dev&msgNo=101831
> - * Message-ID: <877jgjtkus.fsf_at_debian2.lan> */
> -
> - /* fallback to using the network. */
> - SVN_ERR(svn_client_uuid_from_url(uuid, entry->url, ctx, pool));
> - }
> - else
> - {
> - /* Excluded path will fall into this code branch, since the missed
> - fields in the entry for excluded path is not filled. But it is just
> - ok. */
> - return svn_error_createf(SVN_ERR_ENTRY_MISSING_URL, NULL,
> - _("'%s' has no URL"),
> - svn_dirent_local_style(path, pool));
> - }
> + *uuid = apr_pstrdup(result_pool, entry->uuid);
If the assumption is that it is always present, then why bother to
test? And even better: raise an assertion. Otherwise, all this deleted
code needs to be restored somehow.
Or maybe the upgrade to f12 needs to track down the UUID. And if it
can't find one (tho the above comments implies it *is* somewhere in
the WC), then it should bail? Or say how to get it?
If the value is *not* there, then we need to have a better solution somewhere.
In short, the "if entry->uuid" is weasel-y. Go all-in, or not.
>...
Cheers,
-g
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2382711
Received on 2009-08-12 00:32:33 CEST