Daniel Shahaf wrote:
stef...@apache.org wrote on Sat, Aug 14, 2010 at 13:20:22 -0000:
Author: stefan2
Date: Sat Aug 14 13:20:22 2010
New Revision: 985487
URL: http://svn.apache.org/viewvc?rev=985487&view=rev
Log:
APR file I/O is a high frequency operation as the respective streams directly
map their requests
onto the file layer. Thus, introduce svn_io_file_putc() as symmetrical
counterpart to *_getc()
and use both for single char read requests because their APR implementation
tends to involve
far less overhead.
Also, introduce and use svn_io_file_read_full2 that optionally doesn't create
an error object
upon EOF that may be discarded by the caller anyways. This actually translates
into significant
runtime savings because constructing the svn_error_t for file errors involves
expensive locale
dependent functions.
* subversion/include/svn_io.h
(svn_io_file_putc, svn_io_file_read_full2): declare new API functions
* subversion/libsvn_subr/svn_io.c
(svn_io_file_putc, svn_io_file_read_full2): implement new API functions
* subversion/libsvn_subr/stream.c
(read_handler_apr, write_handler_apr): use the I/O function best suitable for
the request
Modified:
subversion/branches/performance/subversion/include/svn_io.h
subversion/branches/performance/subversion/libsvn_subr/io.c
subversion/branches/performance/subversion/libsvn_subr/stream.c
Modified: subversion/branches/performance/subversion/include/svn_io.h
URL:
http://svn.apache.org/viewvc/subversion/branches/performance/subversion/include/svn_io.h?rev=985487&r1=985486&r2=985487&view=diff
==============================================================================
--- subversion/branches/performance/subversion/include/svn_io.h (original)
+++ subversion/branches/performance/subversion/include/svn_io.h Sat Aug 14
13:20:22 2010
@@ -1745,6 +1745,15 @@ svn_io_file_getc(char *ch,
apr_pool_t *pool);
+/** Wrapper for apr_file_putc().
+ * @since New in 1.7
+ */
+svn_error_t *
+svn_io_file_putc(char ch,
+ apr_file_t *file,
+ apr_pool_t *pool);
+
+
/** Wrapper for apr_file_info_get(). */
svn_error_t *
svn_io_file_info_get(apr_finfo_t *finfo,
@@ -1770,6 +1779,20 @@ svn_io_file_read_full(apr_file_t *file,
apr_pool_t *pool);
+/** Wrapper for apr_file_read_full().
+ * If eof_is_ok is set, no svn_error_t error object
+ * will be created upon EOF.
+ * @since New in 1.7
+ */
+svn_error_t *
+svn_io_file_read_full2(apr_file_t *file,
+ void *buf,
+ apr_size_t nbytes,
+ apr_size_t *bytes_read,
+ svn_boolean_t eof_is_ok,
+ apr_pool_t *pool);
+
+
Why didn't you deprecate svn_io_file_read_full() in the same commit?
By the time I wrote that code almost 3 months ago, I tried to minimize
the impact (code churn) on the code base.
Usually we do it as follows:
+ the declaration of svn_io_file_read_full2() is *above* that
of svn_io_file_read_full (not critical)
+ mark the old function SVN_DEPRECATED (preprocessor symbol) and
doxygen @deprecated
+ change the doc string of the old function to be a diff against the new
function's docs
+ in the appropriate *.c file, upgrade the definition in-place
+ re-define the old function in lib_$foo/deprecated.c as a wrapper around the
new function
Took me a number of commits but it should be o.k. by r986492.
Why you didn't make svn_io_file_read_full() a deprecated wrapper around
svn_io_file_read_full2()?
See above ;)
-- Stefan^2.
