You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@subversion.apache.org by gs...@apache.org on 2010/06/08 02:47:22 UTC
svn commit: r952493 - in /subversion/trunk/subversion/libsvn_wc:
wc-queries.sql wc_db.c
Author: gstein
Date: Tue Jun 8 00:47:22 2010
New Revision: 952493
URL: http://svn.apache.org/viewvc?rev=952493&view=rev
Log:
The query that we used to fetch all children in BASE_NODE and WORKING_NODE
used a UNION between two SELECT statements. The idea was to have SQLite
remove all duplicates for us in a single query. Unfortunately, this caused
SQLite to create an ephemeral (temporary) table and place the results of
each query into that table. It created an index to remove dupliates. Then
it returned the values in that ephemeral table. For large numbers of
nodes, the construction of the table and its index becomes very costly.
This change rebuilds gather_children() in wc_db.c to do the duplicate
removal manually using a hash table. It does some simple scanning straight
into an array when it knows duplicates cannot exist (one of BASE or
WORKING is empty).
The performance problem of svn_wc__db_read_children() was first observed
in issue #3499. The actual performance improvement is untested so far, but
I'm assuming pburba can pick up this change and try in his scenario.
* subversion/libsvn_wc/wc_db.c:
(count_children): new helper to count the number of children of a given
PARENT_RELPATH within a specific table.
(add_children_to_hash): new helper to scan children, placing their names
into a hash table as keys (and the mapped values).
(union_children): new helper to scan both BASE_NODE and WORKING_NODE and
manually create a union of the resulting names using a hash table.
(single_table_children): new helper to return the children from a single
table.
(gather_children): rebuilt in terms of the above helpers
* subversion/libsvn_wc/wc-queries.sql:
(STMT_COUNT_BASE_NODE_CHILDREN, STMT_COUNT_WORKING_NODE_CHILDREN): new
statements to count the number of children for a given parent_relpath
(STMT_SELECT_WORKING_CHILDREN): removed
(STMT_SELECT_WORKING_NODE_CHILDREN): like STMT_SELECT_BASE_NODE_CHILDREN,
this scan all children in WORKING_NODE for a given parent_relpath
Modified:
subversion/trunk/subversion/libsvn_wc/wc-queries.sql
subversion/trunk/subversion/libsvn_wc/wc_db.c
Modified: subversion/trunk/subversion/libsvn_wc/wc-queries.sql
URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_wc/wc-queries.sql?rev=952493&r1=952492&r2=952493&view=diff
==============================================================================
--- subversion/trunk/subversion/libsvn_wc/wc-queries.sql (original)
+++ subversion/trunk/subversion/libsvn_wc/wc-queries.sql Tue Jun 8 00:47:22 2010
@@ -87,16 +87,21 @@ INSERT OR IGNORE INTO WORKING_NODE (
wc_id, local_relpath, parent_relpath, presence, kind)
VALUES (?1, ?2, ?3, 'incomplete', 'unknown');
+-- STMT_COUNT_BASE_NODE_CHILDREN
+SELECT COUNT(*) FROM BASE_NODE
+WHERE wc_id = ?1 AND parent_relpath = ?2;
+
+-- STMT_COUNT_WORKING_NODE_CHILDREN
+SELECT COUNT(*) FROM WORKING_NODE
+WHERE wc_id = ?1 AND parent_relpath = ?2;
+
-- STMT_SELECT_BASE_NODE_CHILDREN
select local_relpath from base_node
where wc_id = ?1 and parent_relpath = ?2;
--- STMT_SELECT_WORKING_CHILDREN
-select local_relpath from base_node
-where wc_id = ?1 and parent_relpath = ?2
-union
-select local_relpath from working_node
-where wc_id = ?1 and parent_relpath = ?2;
+-- STMT_SELECT_WORKING_NODE_CHILDREN
+SELECT local_relpath FROM WORKING_NODE
+WHERE wc_id = ?1 AND parent_relpath = ?2;
-- STMT_SELECT_WORKING_IS_FILE
select kind == 'file' from working_node
Modified: subversion/trunk/subversion/libsvn_wc/wc_db.c
URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_wc/wc_db.c?rev=952493&r1=952492&r2=952493&view=diff
==============================================================================
--- subversion/trunk/subversion/libsvn_wc/wc_db.c (original)
+++ subversion/trunk/subversion/libsvn_wc/wc_db.c Tue Jun 8 00:47:22 2010
@@ -813,40 +813,97 @@ insert_working_node(void *baton,
}
-/* */
static svn_error_t *
-gather_children(const apr_array_header_t **children,
- svn_boolean_t base_only,
- svn_wc__db_t *db,
- const char *local_abspath,
- apr_pool_t *result_pool,
- apr_pool_t *scratch_pool)
+count_children(int *count,
+ int stmt_idx,
+ svn_sqlite__db_t *sdb,
+ apr_int64_t wc_id,
+ const char *parent_relpath)
+{
+ svn_sqlite__stmt_t *stmt;
+
+ SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
+ SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
+ SVN_ERR(svn_sqlite__step_row(stmt));
+ *count = svn_sqlite__column_int(stmt, 0);
+ return svn_error_return(svn_sqlite__reset(stmt));
+}
+
+
+/* Each name is allocated in RESULT_POOL and stored into CHILDREN as a key
+ pointed to the same name. */
+static svn_error_t *
+add_children_to_hash(apr_hash_t *children,
+ int stmt_idx,
+ svn_sqlite__db_t *sdb,
+ apr_int64_t wc_id,
+ const char *parent_relpath,
+ apr_pool_t *result_pool)
{
- svn_wc__db_pdh_t *pdh;
- const char *local_relpath;
svn_sqlite__stmt_t *stmt;
- apr_array_header_t *child_names;
svn_boolean_t have_row;
- SVN_ERR_ASSERT(svn_dirent_is_absolute(local_abspath));
+ SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
+ SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
+ SVN_ERR(svn_sqlite__step(&have_row, stmt));
+ while (have_row)
+ {
+ const char *child_relpath = svn_sqlite__column_text(stmt, 0, NULL);
+ const char *name = svn_relpath_basename(child_relpath, result_pool);
- SVN_ERR(svn_wc__db_pdh_parse_local_abspath(&pdh, &local_relpath, db,
- local_abspath, svn_sqlite__mode_readonly,
- scratch_pool, scratch_pool));
- VERIFY_USABLE_PDH(pdh);
+ apr_hash_set(children, name, APR_HASH_KEY_STRING, name);
- SVN_ERR(svn_sqlite__get_statement(&stmt, pdh->wcroot->sdb,
- base_only
- ? STMT_SELECT_BASE_NODE_CHILDREN
- : STMT_SELECT_WORKING_CHILDREN));
- SVN_ERR(svn_sqlite__bindf(stmt, "is", pdh->wcroot->wc_id, local_relpath));
+ SVN_ERR(svn_sqlite__step(&have_row, stmt));
+ }
+
+ return svn_sqlite__reset(stmt);
+}
+
+
+static svn_error_t *
+union_children(const apr_array_header_t **children,
+ svn_sqlite__db_t *sdb,
+ apr_int64_t wc_id,
+ const char *parent_relpath,
+ apr_pool_t *result_pool,
+ apr_pool_t *scratch_pool)
+{
+ /* ### it would be nice to pre-size this hash table. */
+ apr_hash_t *names = apr_hash_make(scratch_pool);
+ apr_array_header_t *names_array;
+
+ /* All of the names get allocated in RESULT_POOL. */
+ SVN_ERR(add_children_to_hash(names, STMT_SELECT_BASE_NODE_CHILDREN,
+ sdb, wc_id, parent_relpath, result_pool));
+ SVN_ERR(add_children_to_hash(names, STMT_SELECT_WORKING_NODE_CHILDREN,
+ sdb, wc_id, parent_relpath, result_pool));
+
+ SVN_ERR(svn_hash_keys(&names_array, names, result_pool));
+ *children = names_array;
+
+ return SVN_NO_ERROR;
+}
+
+
+static svn_error_t *
+single_table_children(const apr_array_header_t **children,
+ int stmt_idx,
+ int start_size,
+ svn_sqlite__db_t *sdb,
+ apr_int64_t wc_id,
+ const char *parent_relpath,
+ apr_pool_t *result_pool)
+{
+ svn_sqlite__stmt_t *stmt;
+ apr_array_header_t *child_names;
+ svn_boolean_t have_row;
+
+ SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
+ SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
/* ### should test the node to ensure it is a directory */
- /* ### 10 is based on Subversion's average of 8.5 files per versioned
- ### directory in its repository. maybe use a different value? or
- ### count rows first? */
- child_names = apr_array_make(result_pool, 10, sizeof(const char *));
+ child_names = apr_array_make(result_pool, start_size, sizeof(const char *));
SVN_ERR(svn_sqlite__step(&have_row, stmt));
while (have_row)
@@ -866,6 +923,76 @@ gather_children(const apr_array_header_t
/* */
+static svn_error_t *
+gather_children(const apr_array_header_t **children,
+ svn_boolean_t base_only,
+ svn_wc__db_t *db,
+ const char *local_abspath,
+ apr_pool_t *result_pool,
+ apr_pool_t *scratch_pool)
+{
+ svn_wc__db_pdh_t *pdh;
+ const char *local_relpath;
+ int base_count;
+ int working_count;
+
+ SVN_ERR_ASSERT(svn_dirent_is_absolute(local_abspath));
+
+ SVN_ERR(svn_wc__db_pdh_parse_local_abspath(&pdh, &local_relpath, db,
+ local_abspath,
+ svn_sqlite__mode_readonly,
+ scratch_pool, scratch_pool));
+ VERIFY_USABLE_PDH(pdh);
+
+ if (base_only)
+ {
+ /* 10 is based on Subversion's average of 8.7 files per versioned
+ directory in its repository.
+
+ ### note "files". should redo count with subdirs included */
+ return svn_error_return(single_table_children(
+ children, STMT_SELECT_BASE_NODE_CHILDREN,
+ 10 /* start_size */,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id,
+ local_relpath, result_pool));
+ }
+
+ SVN_ERR(count_children(&base_count, STMT_COUNT_BASE_NODE_CHILDREN,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id, local_relpath));
+ SVN_ERR(count_children(&working_count, STMT_COUNT_WORKING_NODE_CHILDREN,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id, local_relpath));
+
+ if (base_count == 0)
+ {
+ if (working_count == 0)
+ {
+ *children = apr_array_make(result_pool, 0, sizeof(const char *));
+ return SVN_NO_ERROR;
+ }
+
+ return svn_error_return(single_table_children(
+ children, STMT_SELECT_WORKING_NODE_CHILDREN,
+ working_count,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id,
+ local_relpath, result_pool));
+ }
+ if (working_count == 0)
+ return svn_error_return(single_table_children(
+ children, STMT_SELECT_BASE_NODE_CHILDREN,
+ base_count,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id,
+ local_relpath, result_pool));
+
+ /* ### it would be nice to pass BASE_COUNT and WORKING_COUNT, but there is
+ ### nothing union_children() can do with those. */
+ return svn_error_return(union_children(children,
+ pdh->wcroot->sdb, pdh->wcroot->wc_id,
+ local_relpath,
+ result_pool, scratch_pool));
+}
+
+
+/* */
static void
flush_entries(const svn_wc__db_pdh_t *pdh)
{
Re: svn commit: r952493 - in /subversion/trunk/subversion/libsvn_wc:
wc-queries.sql wc_db.c
Posted by Paul Burba <pt...@gmail.com>.
On Mon, Jun 7, 2010 at 8:47 PM, <gs...@apache.org> wrote:
> Author: gstein
> Date: Tue Jun 8 00:47:22 2010
> New Revision: 952493
>
> URL: http://svn.apache.org/viewvc?rev=952493&view=rev
> Log:
> The query that we used to fetch all children in BASE_NODE and WORKING_NODE
> used a UNION between two SELECT statements. The idea was to have SQLite
> remove all duplicates for us in a single query. Unfortunately, this caused
> SQLite to create an ephemeral (temporary) table and place the results of
> each query into that table. It created an index to remove dupliates. Then
> it returned the values in that ephemeral table. For large numbers of
> nodes, the construction of the table and its index becomes very costly.
>
> This change rebuilds gather_children() in wc_db.c to do the duplicate
> removal manually using a hash table. It does some simple scanning straight
> into an array when it knows duplicates cannot exist (one of BASE or
> WORKING is empty).
>
> The performance problem of svn_wc__db_read_children() was first observed
> in issue #3499. The actual performance improvement is untested so far, but
> I'm assuming pburba can pick up this change and try in his scenario.
On Mon, Jun 7, 2010 at 8:53 PM, Greg Stein <gs...@gmail.com> wrote:
> Hey Paul,
>
> Can you try this change on your large-file-count working copies? I
> believe this should fix the performance problems you were seeing.
Greg,
Short Story: Hours to Seconds
Long Story: This does indeed solve the problems I was seeing:
My test repository was our test suite's Greek tree but with 17,000 1KB
files in a single directory:
Prior to r952493, update and status were taking *quite* some time:
svn st 01:23:33
svn up Gave up after an hour (i.e. lasted longer than my lunch).
With your fix in place, performance improves dramatically:
svn st 00:00:17
svn up 00:00:11
Paul
P.S. Thanks! I was nowhere near figuring this out :-\
Re: svn commit: r952493 - in /subversion/trunk/subversion/libsvn_wc:
wc-queries.sql wc_db.c
Posted by Paul Burba <pt...@gmail.com>.
On Mon, Jun 7, 2010 at 8:47 PM, <gs...@apache.org> wrote:
> Author: gstein
> Date: Tue Jun 8 00:47:22 2010
> New Revision: 952493
>
> URL: http://svn.apache.org/viewvc?rev=952493&view=rev
> Log:
> The query that we used to fetch all children in BASE_NODE and WORKING_NODE
> used a UNION between two SELECT statements. The idea was to have SQLite
> remove all duplicates for us in a single query. Unfortunately, this caused
> SQLite to create an ephemeral (temporary) table and place the results of
> each query into that table. It created an index to remove dupliates. Then
> it returned the values in that ephemeral table. For large numbers of
> nodes, the construction of the table and its index becomes very costly.
>
> This change rebuilds gather_children() in wc_db.c to do the duplicate
> removal manually using a hash table. It does some simple scanning straight
> into an array when it knows duplicates cannot exist (one of BASE or
> WORKING is empty).
>
> The performance problem of svn_wc__db_read_children() was first observed
> in issue #3499. The actual performance improvement is untested so far, but
> I'm assuming pburba can pick up this change and try in his scenario.
On Mon, Jun 7, 2010 at 8:53 PM, Greg Stein <gs...@gmail.com> wrote:
> Hey Paul,
>
> Can you try this change on your large-file-count working copies? I
> believe this should fix the performance problems you were seeing.
Greg,
Short Story: Hours to Seconds
Long Story: This does indeed solve the problems I was seeing:
My test repository was our test suite's Greek tree but with 17,000 1KB
files in a single directory:
Prior to r952493, update and status were taking *quite* some time:
svn st 01:23:33
svn up Gave up after an hour (i.e. lasted longer than my lunch).
With your fix in place, performance improves dramatically:
svn st 00:00:17
svn up 00:00:11
Paul
P.S. Thanks! I was nowhere near figuring this out :-\
Re: svn commit: r952493 - in
/subversion/trunk/subversion/libsvn_wc: wc-queries.sql wc_db.c
Posted by Julian Foad <ju...@btopenworld.com>.
On Tue, 2010-06-15, Julian Foad wrote:
> > Author: gstein
> > Date: Tue Jun 8 00:47:22 2010
> > New Revision: 952493
> >
> > URL: http://svn.apache.org/viewvc?rev=952493&view=rev
> > Log:
[...]
> > * subversion/libsvn_wc/wc_db.c:
> > (count_children): new helper to count the number of children of a given
> > PARENT_RELPATH within a specific table.
> > (add_children_to_hash): new helper to scan children, placing their names
> > into a hash table as keys (and the mapped values).
> > (union_children): new helper to scan both BASE_NODE and WORKING_NODE and
> > manually create a union of the resulting names using a hash table.
> > (single_table_children): new helper to return the children from a single
> > table.
>
> Hi Greg. Please could you copy these descriptions of the four new
> functions into doc strings in the source file.
Ping!
Thanks.
- Julian
Re: svn commit: r952493 - in
/subversion/trunk/subversion/libsvn_wc: wc-queries.sql wc_db.c
Posted by Julian Foad <ju...@wandisco.com>.
On Tue, 2010-06-08, c wrote:
> Author: gstein
> Date: Tue Jun 8 00:47:22 2010
> New Revision: 952493
>
> URL: http://svn.apache.org/viewvc?rev=952493&view=rev
> Log:
> The query that we used to fetch all children in BASE_NODE and WORKING_NODE
> used a UNION between two SELECT statements. The idea was to have SQLite
> remove all duplicates for us in a single query. Unfortunately, this caused
> SQLite to create an ephemeral (temporary) table and place the results of
> each query into that table. It created an index to remove dupliates. Then
> it returned the values in that ephemeral table. For large numbers of
> nodes, the construction of the table and its index becomes very costly.
>
> This change rebuilds gather_children() in wc_db.c to do the duplicate
> removal manually using a hash table. It does some simple scanning straight
> into an array when it knows duplicates cannot exist (one of BASE or
> WORKING is empty).
>
> The performance problem of svn_wc__db_read_children() was first observed
> in issue #3499. The actual performance improvement is untested so far, but
> I'm assuming pburba can pick up this change and try in his scenario.
>
> * subversion/libsvn_wc/wc_db.c:
> (count_children): new helper to count the number of children of a given
> PARENT_RELPATH within a specific table.
> (add_children_to_hash): new helper to scan children, placing their names
> into a hash table as keys (and the mapped values).
> (union_children): new helper to scan both BASE_NODE and WORKING_NODE and
> manually create a union of the resulting names using a hash table.
> (single_table_children): new helper to return the children from a single
> table.
Hi Greg. Please could you copy these descriptions of the four new
functions into doc strings in the source file.
Thanks.
- Julian
> (gather_children): rebuilt in terms of the above helpers
[...]
> Modified: subversion/trunk/subversion/libsvn_wc/wc_db.c
[...]
> static svn_error_t *
> +count_children(int *count,
> + int stmt_idx,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath)
> +{
> + svn_sqlite__stmt_t *stmt;
> +
> + SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
> + SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
> + SVN_ERR(svn_sqlite__step_row(stmt));
> + *count = svn_sqlite__column_int(stmt, 0);
> + return svn_error_return(svn_sqlite__reset(stmt));
> +}
> +
> +
> +/* Each name is allocated in RESULT_POOL and stored into CHILDREN as a key
> + pointed to the same name. */
> +static svn_error_t *
> +add_children_to_hash(apr_hash_t *children,
> + int stmt_idx,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool)
[...]
> +
> +static svn_error_t *
> +union_children(const apr_array_header_t **children,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool,
> + apr_pool_t *scratch_pool)
[...]
> +
> +static svn_error_t *
> +single_table_children(const apr_array_header_t **children,
> + int stmt_idx,
> + int start_size,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool)
[...]
> /* */
> +static svn_error_t *
> +gather_children(const apr_array_header_t **children,
> + svn_boolean_t base_only,
> + svn_wc__db_t *db,
> + const char *local_abspath,
> + apr_pool_t *result_pool,
> + apr_pool_t *scratch_pool)
Re: svn commit: r952493 - in /subversion/trunk/subversion/libsvn_wc:
wc-queries.sql wc_db.c
Posted by Greg Stein <gs...@gmail.com>.
Hey Paul,
Can you try this change on your large-file-count working copies? I
believe this should fix the performance problems you were seeing.
Cheers,
-g
On Mon, Jun 7, 2010 at 20:47, <gs...@apache.org> wrote:
> Author: gstein
> Date: Tue Jun 8 00:47:22 2010
> New Revision: 952493
>
> URL: http://svn.apache.org/viewvc?rev=952493&view=rev
> Log:
> The query that we used to fetch all children in BASE_NODE and WORKING_NODE
> used a UNION between two SELECT statements. The idea was to have SQLite
> remove all duplicates for us in a single query. Unfortunately, this caused
> SQLite to create an ephemeral (temporary) table and place the results of
> each query into that table. It created an index to remove dupliates. Then
> it returned the values in that ephemeral table. For large numbers of
> nodes, the construction of the table and its index becomes very costly.
>
> This change rebuilds gather_children() in wc_db.c to do the duplicate
> removal manually using a hash table. It does some simple scanning straight
> into an array when it knows duplicates cannot exist (one of BASE or
> WORKING is empty).
>
> The performance problem of svn_wc__db_read_children() was first observed
> in issue #3499. The actual performance improvement is untested so far, but
> I'm assuming pburba can pick up this change and try in his scenario.
>
> * subversion/libsvn_wc/wc_db.c:
> (count_children): new helper to count the number of children of a given
> PARENT_RELPATH within a specific table.
> (add_children_to_hash): new helper to scan children, placing their names
> into a hash table as keys (and the mapped values).
> (union_children): new helper to scan both BASE_NODE and WORKING_NODE and
> manually create a union of the resulting names using a hash table.
> (single_table_children): new helper to return the children from a single
> table.
> (gather_children): rebuilt in terms of the above helpers
>
> * subversion/libsvn_wc/wc-queries.sql:
> (STMT_COUNT_BASE_NODE_CHILDREN, STMT_COUNT_WORKING_NODE_CHILDREN): new
> statements to count the number of children for a given parent_relpath
> (STMT_SELECT_WORKING_CHILDREN): removed
> (STMT_SELECT_WORKING_NODE_CHILDREN): like STMT_SELECT_BASE_NODE_CHILDREN,
> this scan all children in WORKING_NODE for a given parent_relpath
>
> Modified:
> subversion/trunk/subversion/libsvn_wc/wc-queries.sql
> subversion/trunk/subversion/libsvn_wc/wc_db.c
>
> Modified: subversion/trunk/subversion/libsvn_wc/wc-queries.sql
> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_wc/wc-queries.sql?rev=952493&r1=952492&r2=952493&view=diff
> ==============================================================================
> --- subversion/trunk/subversion/libsvn_wc/wc-queries.sql (original)
> +++ subversion/trunk/subversion/libsvn_wc/wc-queries.sql Tue Jun 8 00:47:22 2010
> @@ -87,16 +87,21 @@ INSERT OR IGNORE INTO WORKING_NODE (
> wc_id, local_relpath, parent_relpath, presence, kind)
> VALUES (?1, ?2, ?3, 'incomplete', 'unknown');
>
> +-- STMT_COUNT_BASE_NODE_CHILDREN
> +SELECT COUNT(*) FROM BASE_NODE
> +WHERE wc_id = ?1 AND parent_relpath = ?2;
> +
> +-- STMT_COUNT_WORKING_NODE_CHILDREN
> +SELECT COUNT(*) FROM WORKING_NODE
> +WHERE wc_id = ?1 AND parent_relpath = ?2;
> +
> -- STMT_SELECT_BASE_NODE_CHILDREN
> select local_relpath from base_node
> where wc_id = ?1 and parent_relpath = ?2;
>
> --- STMT_SELECT_WORKING_CHILDREN
> -select local_relpath from base_node
> -where wc_id = ?1 and parent_relpath = ?2
> -union
> -select local_relpath from working_node
> -where wc_id = ?1 and parent_relpath = ?2;
> +-- STMT_SELECT_WORKING_NODE_CHILDREN
> +SELECT local_relpath FROM WORKING_NODE
> +WHERE wc_id = ?1 AND parent_relpath = ?2;
>
> -- STMT_SELECT_WORKING_IS_FILE
> select kind == 'file' from working_node
>
> Modified: subversion/trunk/subversion/libsvn_wc/wc_db.c
> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_wc/wc_db.c?rev=952493&r1=952492&r2=952493&view=diff
> ==============================================================================
> --- subversion/trunk/subversion/libsvn_wc/wc_db.c (original)
> +++ subversion/trunk/subversion/libsvn_wc/wc_db.c Tue Jun 8 00:47:22 2010
> @@ -813,40 +813,97 @@ insert_working_node(void *baton,
> }
>
>
> -/* */
> static svn_error_t *
> -gather_children(const apr_array_header_t **children,
> - svn_boolean_t base_only,
> - svn_wc__db_t *db,
> - const char *local_abspath,
> - apr_pool_t *result_pool,
> - apr_pool_t *scratch_pool)
> +count_children(int *count,
> + int stmt_idx,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath)
> +{
> + svn_sqlite__stmt_t *stmt;
> +
> + SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
> + SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
> + SVN_ERR(svn_sqlite__step_row(stmt));
> + *count = svn_sqlite__column_int(stmt, 0);
> + return svn_error_return(svn_sqlite__reset(stmt));
> +}
> +
> +
> +/* Each name is allocated in RESULT_POOL and stored into CHILDREN as a key
> + pointed to the same name. */
> +static svn_error_t *
> +add_children_to_hash(apr_hash_t *children,
> + int stmt_idx,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool)
> {
> - svn_wc__db_pdh_t *pdh;
> - const char *local_relpath;
> svn_sqlite__stmt_t *stmt;
> - apr_array_header_t *child_names;
> svn_boolean_t have_row;
>
> - SVN_ERR_ASSERT(svn_dirent_is_absolute(local_abspath));
> + SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
> + SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
> + SVN_ERR(svn_sqlite__step(&have_row, stmt));
> + while (have_row)
> + {
> + const char *child_relpath = svn_sqlite__column_text(stmt, 0, NULL);
> + const char *name = svn_relpath_basename(child_relpath, result_pool);
>
> - SVN_ERR(svn_wc__db_pdh_parse_local_abspath(&pdh, &local_relpath, db,
> - local_abspath, svn_sqlite__mode_readonly,
> - scratch_pool, scratch_pool));
> - VERIFY_USABLE_PDH(pdh);
> + apr_hash_set(children, name, APR_HASH_KEY_STRING, name);
>
> - SVN_ERR(svn_sqlite__get_statement(&stmt, pdh->wcroot->sdb,
> - base_only
> - ? STMT_SELECT_BASE_NODE_CHILDREN
> - : STMT_SELECT_WORKING_CHILDREN));
> - SVN_ERR(svn_sqlite__bindf(stmt, "is", pdh->wcroot->wc_id, local_relpath));
> + SVN_ERR(svn_sqlite__step(&have_row, stmt));
> + }
> +
> + return svn_sqlite__reset(stmt);
> +}
> +
> +
> +static svn_error_t *
> +union_children(const apr_array_header_t **children,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool,
> + apr_pool_t *scratch_pool)
> +{
> + /* ### it would be nice to pre-size this hash table. */
> + apr_hash_t *names = apr_hash_make(scratch_pool);
> + apr_array_header_t *names_array;
> +
> + /* All of the names get allocated in RESULT_POOL. */
> + SVN_ERR(add_children_to_hash(names, STMT_SELECT_BASE_NODE_CHILDREN,
> + sdb, wc_id, parent_relpath, result_pool));
> + SVN_ERR(add_children_to_hash(names, STMT_SELECT_WORKING_NODE_CHILDREN,
> + sdb, wc_id, parent_relpath, result_pool));
> +
> + SVN_ERR(svn_hash_keys(&names_array, names, result_pool));
> + *children = names_array;
> +
> + return SVN_NO_ERROR;
> +}
> +
> +
> +static svn_error_t *
> +single_table_children(const apr_array_header_t **children,
> + int stmt_idx,
> + int start_size,
> + svn_sqlite__db_t *sdb,
> + apr_int64_t wc_id,
> + const char *parent_relpath,
> + apr_pool_t *result_pool)
> +{
> + svn_sqlite__stmt_t *stmt;
> + apr_array_header_t *child_names;
> + svn_boolean_t have_row;
> +
> + SVN_ERR(svn_sqlite__get_statement(&stmt, sdb, stmt_idx));
> + SVN_ERR(svn_sqlite__bindf(stmt, "is", wc_id, parent_relpath));
>
> /* ### should test the node to ensure it is a directory */
>
> - /* ### 10 is based on Subversion's average of 8.5 files per versioned
> - ### directory in its repository. maybe use a different value? or
> - ### count rows first? */
> - child_names = apr_array_make(result_pool, 10, sizeof(const char *));
> + child_names = apr_array_make(result_pool, start_size, sizeof(const char *));
>
> SVN_ERR(svn_sqlite__step(&have_row, stmt));
> while (have_row)
> @@ -866,6 +923,76 @@ gather_children(const apr_array_header_t
>
>
> /* */
> +static svn_error_t *
> +gather_children(const apr_array_header_t **children,
> + svn_boolean_t base_only,
> + svn_wc__db_t *db,
> + const char *local_abspath,
> + apr_pool_t *result_pool,
> + apr_pool_t *scratch_pool)
> +{
> + svn_wc__db_pdh_t *pdh;
> + const char *local_relpath;
> + int base_count;
> + int working_count;
> +
> + SVN_ERR_ASSERT(svn_dirent_is_absolute(local_abspath));
> +
> + SVN_ERR(svn_wc__db_pdh_parse_local_abspath(&pdh, &local_relpath, db,
> + local_abspath,
> + svn_sqlite__mode_readonly,
> + scratch_pool, scratch_pool));
> + VERIFY_USABLE_PDH(pdh);
> +
> + if (base_only)
> + {
> + /* 10 is based on Subversion's average of 8.7 files per versioned
> + directory in its repository.
> +
> + ### note "files". should redo count with subdirs included */
> + return svn_error_return(single_table_children(
> + children, STMT_SELECT_BASE_NODE_CHILDREN,
> + 10 /* start_size */,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id,
> + local_relpath, result_pool));
> + }
> +
> + SVN_ERR(count_children(&base_count, STMT_COUNT_BASE_NODE_CHILDREN,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id, local_relpath));
> + SVN_ERR(count_children(&working_count, STMT_COUNT_WORKING_NODE_CHILDREN,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id, local_relpath));
> +
> + if (base_count == 0)
> + {
> + if (working_count == 0)
> + {
> + *children = apr_array_make(result_pool, 0, sizeof(const char *));
> + return SVN_NO_ERROR;
> + }
> +
> + return svn_error_return(single_table_children(
> + children, STMT_SELECT_WORKING_NODE_CHILDREN,
> + working_count,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id,
> + local_relpath, result_pool));
> + }
> + if (working_count == 0)
> + return svn_error_return(single_table_children(
> + children, STMT_SELECT_BASE_NODE_CHILDREN,
> + base_count,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id,
> + local_relpath, result_pool));
> +
> + /* ### it would be nice to pass BASE_COUNT and WORKING_COUNT, but there is
> + ### nothing union_children() can do with those. */
> + return svn_error_return(union_children(children,
> + pdh->wcroot->sdb, pdh->wcroot->wc_id,
> + local_relpath,
> + result_pool, scratch_pool));
> +}
> +
> +
> +/* */
> static void
> flush_entries(const svn_wc__db_pdh_t *pdh)
> {
>
>
>