> +typedef struct svn_diff_node_t svn_diff_node_t;
> +typedef struct svn_diff_tree_t svn_diff_tree_t;
> +typedef struct svn_diff_position_t svn_diff_position_t;
> +typedef struct svn_diff_lcs_t svn_diff_lcs_t;
Anything not publically exported in svn_diff.h will need the
double-underscore namespace protection. I won't mention it beyond
this, as it applies to a bunch of other stuff in libsvn_diff/diff.c
and you can find them all pretty easily.
> +typedef enum svn_diff_type_e
> +{
> + svn_diff_type_common,
> + svn_diff_type_diff_workingcopy,
> + svn_diff_type_diff_repository,
> + svn_diff_type_diff_common,
> + svn_diff_type_conflict
> +} svn_diff_type_e;
> +
> +struct svn_diff_t {
> + svn_diff_t *next;
> + svn_diff_type_e type;
> + apr_off_t baseline_start;
> + apr_off_t baseline_length;
> + apr_off_t workingcopy_start;
> + apr_off_t workingcopy_length;
> + apr_off_t repository_start;
> + apr_off_t repository_length;
> +};
Hmmm. These make me wonder about the general approach... Why would an
internal diff library know anything about working copies versus
repositories? IMHO, its callers should just hand it data_source_1 and
data_source_2 (plus possibly data_source_3 for diff3), and it will do
diffy things. It doesn't need to know that one of these sources is
from the wc, and another represents the repository.
In other words, I think it would be better to write this diff library
as though you plan to see it used by lots of other applications, not
just by this one version control system. That will result in the
cleanest interfaces and internal data flow.
I haven't reviewed the rest of the code, not with any rigor anyway, so
silence signifies only ignorance.
-K
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Apr 23 23:27:48 2002