Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Subversion Wiki" for 
change notification.

The "Svn19ApiReview" page has been changed by StefanFuhrmann:
https://wiki.apache.org/subversion/Svn19ApiReview?action=diff&rev1=11&rev2=12

Comment:
First round of API review has been completed.

  ----
  == Needs Review ==
  
+ All public API headers have been reviewed at least once.
- The following 1.9 headers contain non-trivial changes against 1.8:
- 
-  . mod_dav_svn.h
-  . svn_cmdline.h
-  . svn_config.h
-  . svn_diff.h
-  . svn_error.h
-  . svn_io.h
-  . svn_mergeinfo.h
-  . svn_ra.h
-  . svn_ra_svn.h
-  . svn_string.h
-  . svn_version.h
-  . svn_wc.h
  
  ----
  == Reviewed But Need Further Attention ==
@@ -29, +16 @@

   * `svn_client_vacuum` — Missing documentation for arguments 
`fix_recorded_timestamps` and `vacuum_pristines`. (svn praise: r1548088 
rhuijben)
   * `svn_client_cleanup2` — Missing documentation for arguments 
`fix_recorded_timestamps`, `clear_dav_cache` and `vacuum_pristines`. (svn 
praise: r1548088 rhuijben)
   * `svn_client_cleanup`  — Missing SVN_DEPRECATED annotation.
+ 
+ 
+ === svn_diff.h ===
+ 
+  * `svn_diff_output2`
+  * `svn_diff_file_output_unified4`
+  * `svn_diff_output_binary`
+  * `svn_diff_file_output_merge3`
+  * `svn_diff_mem_string_output_unified3`
+  * `svn_diff_mem_string_output_merge3` — Document the optional cancel_func & 
cancel_baton. It may be helpful to tell when they will be invoked (once per 
file / hunk?).
  
  === svn_dirent_uri.h ===
   * `svn_relpath_limit` — Is this really a good public API, either in 
functionality or in name? I  suggest it should be a local function in the one 
file where it's used,  or else made private and named 
`svn_relpath__first_n_components`.
@@ -40, +37 @@

   * `svn_fs_lock_callback_t` — `pool` parameter: Maybe rename to 
`scratch_pool`.
   * `svn_fs_lock_many`, `svn_fs_unlock_many` — Mention in docstring that these 
are not atomic operations. If the callback returns an error, they simply return 
that error to their callers and leave part of the locks in their initial and 
the remainder in the final state. What is the lock error handling in case of 
`lock_callback==NULL`?
   * `svn_fs_unlock_many` — What is the `result_pool` being used for?
+ 
+ === svn_ra_svn.h ===
+ 
+  * `svn_ra_svn_create_conn4` — The @note is unclear. Is the file being used 
for time-out detection?
  
  === svn_repos.h ===
  
@@ -53, +54 @@

   * `svn_null_pointer_constant_stdarg_sentinel_t` — Should this be private?
     * No; the idea is that API users can use `SVN_VA_NULL`. Note that the 
definition is a forward declaration of the type for the purpose of making it a 
pointer constant that's not trivially assignable to anything else (except 
`void*`), but the type itself is never defined, on purpose. — brane
  
+ 
+ === svn_wc.h ===
+ 
+  * `svn_wc_notify_move_broken` — Has a TODO left-over from 1.8
+  * `svn_wc_conflict_description2_t.merged_file` — Clarify the ###.
+  * `svn_wc_conflict_description_create_tree2` — What is that @c 
local_node_kind that is being mentioned _twice_?
+  * `svn_wc__conflict_description2_dup` — Simply remove from API.
+  * `svn_wc_conflict_choice_t.svn_wc_conflict_choose_undefined` — If truly 
internal, we might rename it to `svn_wc_conflict_choose__undefined`.
+  * `svn_wc_add_from_disk3` — Still has TODO and ###s. Document the 
notification callback (is it optional?)
+  * `svn_wc_add_from_disk2` — What is skip_som_prop_canon?
+  * `svn_wc_queue_committed4` — Various outstanding ###s.
+  * `svn_wc_cleanup4` — Document the notification callback (is it optional?)
+ 
  ----
  == Review Completed! ==
+ 
+ These headers have been reviewed by stefan2 and where issues have been found, 
they are already fixed.  More review is welcome.
+ 
+  . mod_dav_svn.h
+  . svn_cmdline.h
+  . svn_config.h
+  . svn_error.h
+  . svn_io.h
+  . svn_mergeinfo.h
+  . svn_ra.h
+  . svn_string.h
+  . svn_version.h  
  
  The following 1.9 header contain only minor changes against 1.8.  Those 
changes seem to be o.k.  Reviewed by stefan2.
  

Reply via email to