STATUS - subversion - Git at Google

       * * * * * * * * * * * * * * * * * * * * * * * * * * * *
       *                                                     *
       *  THIS RELEASE STREAM IS OPEN TO BUG FIXES.          *
       *                                                     *
       * * * * * * * * * * * * * * * * * * * * * * * * * * * *

 This file tracks the status of releases in the 1.6.x line.

 See "Stabilizing and maintaining releases" in trunk/www/hacking.html for
 details on how release lines and voting work, what kinds of bugs can
 delay a release, etc.

 Status of 1.6.12:
 ================

 Candidate changes:

  * r876032, r876104, r878447, r878910, r878911, r878916
    Several fixes to the Win32 exception handler. This change also makes it
    available on Windows Vista and 7 where it was previously blocked by a
    version check.
    Justification:
      This exception handler is a valuable debugging tool for users who don't
      have development tools installed. (Most users on Windows).
    Notes:
      Most of these revisions cancel out each other. The real changes are in
      r878447 (removing version check) and r878910,r878916 (fixing formatting
      specifier). It's probably easier to review the combined patch than the
      individual revisions.
    Branch:
      ^/subversion/branches/1.6.x-r38836.
    Votes:
      +1: rhuijben, steveking
      +0: pburba - I don't have a Vista or Win7 box to try this out on, but it
          works fine on my XP box.
      +0: lgo - code reviewed and looks good, but I don't have access to a
          Windows machine.

  * r879766
    Fix issue #3498: Subversion password stores freeze Eclipse
    Justification:
      All subclipse users are in great need of this.
    Branch:
      Resolves a minor conflict in libsvn_auth_gnome_keyring/gnome_keyring.c
      where the code changed was refactored on trunk.
      ^/subversion/branches/1.6.x-r39692
    Votes:
      +1: stylesen
      +0: julianfoad (reviewed, not tested)
      -0: rhuijben (needs more explanation: The not-interactive and no
                    callback code paths should be more similar if the code
                    works correctly. I think the not interactive code path
                    needs the same fix).

  * r879902, r879916
    Disables file synchronization mutexes in fsfs on Windows. These mutexes
    were added to make POSIX systems lock files that were in use by other
    threads in the same process.

    Windows implements per filehandle locking, so it doesn't need these
    locks. We usually see the opposite problem in Subversion on Windows.
    Justification:
      This gives a measurable performance boost (5% decrease in running time
      on running the testsuite in parallel mode (on trunk). The difference
      will be bigger when measured on a real subversion server like one
      running apache or svnserve.
    Votes:
      +1: rhuijben, steveking

  * r879966, r879971
    Reduce the lifetime of the rep-sharing database handles in FSFS to avoid
    having writers blocking readers, and update that database outside the
    general FSFS write lock.  This address issue #3506.
    Justification:
      The performance problems and outright failures of the current code
      are hurting the ASF repository, and presumably other high-concurrency
      repositories, too, when large commits occur.
    Branch:
      ^/subversion/branches/1.6.x-issue3506
    Votes:
      +1: danielsh, cmpilato

  * r934599, r934603
    Fix a concurrency issue in the FSFS rep-cache code.  This indirectly
    concerns issue #3506.
    Justification:
      Opening the DB twice (and overwriting the handle) is undesired.
    Notes:
      r934599 updates the svn_atomic__init_once() interface.
      r934603 is the actual fix.
    Notes:
      Depends on the r879966 group, which does not merge cleanly.
    Branch:
      ^/subversion/branches/1.6.x-issue3506-init_once
      (based on ^/subversion/branches/1.6.x-issue3506)
    Votes:
      +1: danielsh

  * r898048
    Extend the Windows only, crypto api verification of server certificates
    to check for certificate revocation. Also simplify the code by making
    Windows do the chain verification instead of just taking an intermediate
    result.
    Justification:
      It's better to be safe, when we are talking about SSL.
    Votes:
      +1: rhuijben, steveking

  * r878590, r878607, r878625, r878626, r878627
    In trunk we optimized the common case of 'find-and-replace with same uri'
    of proxied content thanks to issue 3445 "WebDAV proxy code munging user
    data".
    r878590 is just a change that adds FIXME marker to code comment. We take it
    to avoid spurious conflicts with other revisions.
    r878607 Special cases no-op find and replace.
    "r878625, r878626, r878627" are follow-up to r878607 and the other long
    pending Master info leak to the clients.
    Justification:
      1. This group has the most common 'optimization' fix of *not* groking the
         proxied response to find and replace with same string.
      2. Fixes the master information leak via the Location header.
      3. Need this to be ported for the other defect to be ported
         without conflict.
    Votes:
      +1: kameshj
      -0: julianfoad (seem to be multiple changes here for different reasons -
            at least issue '3445 and an optimization and an information leak;
            r878607 log msg says it fixes bugs but it's not clear what bugs;
            don't know how to tell whether justification 1 is significant;
            justifications don't seem to refer to issue #3445. Please can we
            separate these changes and clearly describe each one? And update
            the r878607 log msg.)

  * r916286, r917512
    ???
    Notes:
      This backport depends on the r878590 group only for the conflict free port.
    Justification:
      Vague commit error when server has been configured with <Location /svn/>
      and write through proxy.
    Votes:
      +1: kameshj

  * r917523
    ???
    Notes:
      This backport depends on the r878590 and r916286 groups only for the
      conflict free port.
    Justification:
      If the configured slave url has the *encodable* characters *write through*
      is not happening rather commit happens in slave itself.
    Votes:
      +1: kameshj

  * r935631
    Fix one of the sanity checks in the 'svn merge --reintegrate' logic.
    Justification:
      Sanity checks should themselves be sane.
    Votes:
      +1: cmpilato

  * r935992, r935996 (issue #3636)
    Makes it possible to delete paths directly from the repository that contain
    characters that need escaping in an uri.
    Justification:
      svn rm URL should just work on 'http://server/path/with%20space'
    Notes: r935992 is a test for this issue and r935996 fixes the issue and
      extends the testcase a bit.
    Branch:
      ^/subversion/branches/1.6.x-r935996.
    Votes:
      +1 rhuijben, steveking

  * r937610
    Allow fs-pack-test 1 to pass with low file descriptor limits.
    Justification:
      Makes regression tests pass on a default OpenBSD install
      without requiring users to bump the default fd limit first.
      See http://marc.info/?l=openbsd-ports&m=127197191204545&w=2
      for problem report. (The test logs show fs-pack-test 1 failing with
      "svn_tests: Can't open file 'test-repo-fsfs-pack/transactions/
      33-x.txn/changes': Too many open files")
    Votes:
      +1: stsp

  * r939375-939376
    Fix issue #3623, a bug in foreign repository merges that causes properties
    on merge-added files to be dropped from commits of those merges.
    Justification:
      This can actually result in what is arguably a corrupt working copy,
      one that thinks it is in sync with the repository but actually is not.
    Notes:
      r939375 tweaks a test to cover this case.  There's a minor conflict
      with its merge with a single deleted function call.  (Keep 'theirs'.)
      r939376 is the fix.  It merges cleanly.
    Votes:
      +1: cmpilato

  * r944635
    On Windows, allow building subversion against Sqlite versions that have a
    fourth number in their version string, such as 3.6.23.1.
    Justification:
      We force users to run using unpatched versions without this patch.
    Votes:
      +1: rhuijben, danielsh

  * r945708
    Fix issue #3469 "tree conflict inside of directory external causes
    assertion to fail".
    Justification:
      Assertion failures are bad and tree-conflicted externals cannot
      be reverted or resolved without this fix.
    Branch:
      ^/subversion/branches/1.6.x-issue3469
      This is not a backport branch. The fix is for 1.6.x only.
      Trunk will need a separate fix as part of wc-ng.
    Votes:
      +1: stsp, julianfoad

 Veto-blocked changes:
 =====================

  * r877552, r877617, r877618, r877619, r877621
    Support using `svnversion` on files and improve some messages printed by
    `svnversion`.
    Justification:
      Some messages printed by `svnversion` are inconsistent and sometimes
      incorrect:
      $ svnversion /usr
      exported
      $ svnversion /bin/ls
      '/bin/ls' not versioned, and not exported
      $ svnversion VERSIONED_FILE
      'VERSIONED_FILE' not versioned, and not exported
      $ svnversion /nonexistent
      subversion/libsvn_wc/questions.c:90: (apr_err=2)
      svn: '/nonexistent' does not exist
      ^^^
      Messages with these revisions backported:
      $ svnversion /usr
      Unversioned directory
      $ svnversion /bin/ls
      Unversioned file
      $ svnversion VERSIONED_FILE
      NUMBER
      $ svnversion /nonexistent
      '/nonexistent' doesn't exist
    Votes:
      +1: arfrever
      -0: cmpilato (Introduces new features and changes output)
      -1: rhuijben (Introduces new features and changes output;
                    I think it should be removed from STATUS)

  * r878078, r878088, r878102, r878196
    Make KWallet window attached to terminal window.
    Justification:
      Without this change, KWallet window can appear behind terminal window
      and is left after killing terminal.
    Votes:
      +1: arfrever
      -0: peters
      -1: stsp, rhuijben ("/proc/%ld/stat" is ugly and Linux-specific -
                          would just getenv("WINDOWID") work instead?)

  * r921453, r927184, r927243
    Fix reopened issue #3020 'Reflect dropped/renumbered revisions in
    svn:mergeinfo data during svnadmin load'
    Justification:
      Prior to this fix, when loading a partial dump with mergeinfo, the
      resulting mergeinfo in the target repository could refer to non-existent
      revisions or revisions that have nothing to do with the merge source
      in the original repository.  The original fix for issue assumed that
      the dump stream was for a complete repository.
    Notes:
      r921453 and 927184 are tests, r927243 is the fix.
    Branch:
      ^/subversion/branches/1.6.x-issue3020
    Votes:
      +1: rhuijben, kameshj
      -1: pburba (There is a regression with this fix, see
          http://svn.haxx.se/dev/archive-2010-03/0716.shtml)

  * r933194
    Fix for issue #2753
    SVNListParentPath feature doesn't work when svn authz is used.
    Justification:
      This issue is waiting for resolution for more than 3 years.
    Notes:
      See discussion at http://svn.haxx.se/dev/archive-2010-04/0275.shtml
    Votes:
      -1: cmpilato (pending results of aforemented discussion)

 Approved changes:
 =================
	* * * * * * * * * * * * * * * * * * * * * * * * * * * *
	* *
	* THIS RELEASE STREAM IS OPEN TO BUG FIXES. *
	* *
	* * * * * * * * * * * * * * * * * * * * * * * * * * * *

	This file tracks the status of releases in the 1.6.x line.

	See "Stabilizing and maintaining releases" in trunk/www/hacking.html for
	details on how release lines and voting work, what kinds of bugs can
	delay a release, etc.

	Status of 1.6.12:
	================

	Candidate changes:

	* r876032, r876104, r878447, r878910, r878911, r878916
	Several fixes to the Win32 exception handler. This change also makes it
	available on Windows Vista and 7 where it was previously blocked by a
	version check.
	Justification:
	This exception handler is a valuable debugging tool for users who don't
	have development tools installed. (Most users on Windows).
	Notes:
	Most of these revisions cancel out each other. The real changes are in
	r878447 (removing version check) and r878910,r878916 (fixing formatting
	specifier). It's probably easier to review the combined patch than the
	individual revisions.
	Branch:
	^/subversion/branches/1.6.x-r38836.
	Votes:
	+1: rhuijben, steveking
	+0: pburba - I don't have a Vista or Win7 box to try this out on, but it
	works fine on my XP box.
	+0: lgo - code reviewed and looks good, but I don't have access to a
	Windows machine.

	* r879766
	Fix issue #3498: Subversion password stores freeze Eclipse
	Justification:
	All subclipse users are in great need of this.
	Branch:
	Resolves a minor conflict in libsvn_auth_gnome_keyring/gnome_keyring.c
	where the code changed was refactored on trunk.
	^/subversion/branches/1.6.x-r39692
	Votes:
	+1: stylesen
	+0: julianfoad (reviewed, not tested)
	-0: rhuijben (needs more explanation: The not-interactive and no
	callback code paths should be more similar if the code
	works correctly. I think the not interactive code path
	needs the same fix).

	* r879902, r879916
	Disables file synchronization mutexes in fsfs on Windows. These mutexes
	were added to make POSIX systems lock files that were in use by other
	threads in the same process.

	Windows implements per filehandle locking, so it doesn't need these
	locks. We usually see the opposite problem in Subversion on Windows.
	Justification:
	This gives a measurable performance boost (5% decrease in running time
	on running the testsuite in parallel mode (on trunk). The difference
	will be bigger when measured on a real subversion server like one
	running apache or svnserve.
	Votes:
	+1: rhuijben, steveking

	* r879966, r879971
	Reduce the lifetime of the rep-sharing database handles in FSFS to avoid
	having writers blocking readers, and update that database outside the
	general FSFS write lock. This address issue #3506.
	Justification:
	The performance problems and outright failures of the current code
	are hurting the ASF repository, and presumably other high-concurrency
	repositories, too, when large commits occur.
	Branch:
	^/subversion/branches/1.6.x-issue3506
	Votes:
	+1: danielsh, cmpilato

	* r934599, r934603
	Fix a concurrency issue in the FSFS rep-cache code. This indirectly
	concerns issue #3506.
	Justification:
	Opening the DB twice (and overwriting the handle) is undesired.
	Notes:
	r934599 updates the svn_atomic__init_once() interface.
	r934603 is the actual fix.
	Notes:
	Depends on the r879966 group, which does not merge cleanly.
	Branch:
	^/subversion/branches/1.6.x-issue3506-init_once
	(based on ^/subversion/branches/1.6.x-issue3506)
	Votes:
	+1: danielsh

	* r898048
	Extend the Windows only, crypto api verification of server certificates
	to check for certificate revocation. Also simplify the code by making
	Windows do the chain verification instead of just taking an intermediate
	result.
	Justification:
	It's better to be safe, when we are talking about SSL.
	Votes:
	+1: rhuijben, steveking

	* r878590, r878607, r878625, r878626, r878627
	In trunk we optimized the common case of 'find-and-replace with same uri'
	of proxied content thanks to issue 3445 "WebDAV proxy code munging user
	data".
	r878590 is just a change that adds FIXME marker to code comment. We take it
	to avoid spurious conflicts with other revisions.
	r878607 Special cases no-op find and replace.
	"r878625, r878626, r878627" are follow-up to r878607 and the other long
	pending Master info leak to the clients.
	Justification:
	1. This group has the most common 'optimization' fix of not groking the
	proxied response to find and replace with same string.
	2. Fixes the master information leak via the Location header.
	3. Need this to be ported for the other defect to be ported
	without conflict.
	Votes:
	+1: kameshj
	-0: julianfoad (seem to be multiple changes here for different reasons -
	at least issue '3445 and an optimization and an information leak;
	r878607 log msg says it fixes bugs but it's not clear what bugs;
	don't know how to tell whether justification 1 is significant;
	justifications don't seem to refer to issue #3445. Please can we
	separate these changes and clearly describe each one? And update
	the r878607 log msg.)

	* r916286, r917512
	???
	Notes:
	This backport depends on the r878590 group only for the conflict free port.
	Justification:
	Vague commit error when server has been configured with <Location /svn/>
	and write through proxy.
	Votes:
	+1: kameshj

	* r917523
	???
	Notes:
	This backport depends on the r878590 and r916286 groups only for the
	conflict free port.
	Justification:
	If the configured slave url has the encodable characters write through
	is not happening rather commit happens in slave itself.
	Votes:
	+1: kameshj

	* r935631
	Fix one of the sanity checks in the 'svn merge --reintegrate' logic.
	Justification:
	Sanity checks should themselves be sane.
	Votes:
	+1: cmpilato

	* r935992, r935996 (issue #3636)
	Makes it possible to delete paths directly from the repository that contain
	characters that need escaping in an uri.
	Justification:
	svn rm URL should just work on 'http://server/path/with%20space'
	Notes: r935992 is a test for this issue and r935996 fixes the issue and
	extends the testcase a bit.
	Branch:
	^/subversion/branches/1.6.x-r935996.
	Votes:
	+1 rhuijben, steveking

	* r937610
	Allow fs-pack-test 1 to pass with low file descriptor limits.
	Justification:
	Makes regression tests pass on a default OpenBSD install
	without requiring users to bump the default fd limit first.
	See http://marc.info/?l=openbsd-ports&m=127197191204545&w=2
	for problem report. (The test logs show fs-pack-test 1 failing with
	"svn_tests: Can't open file 'test-repo-fsfs-pack/transactions/
	33-x.txn/changes': Too many open files")
	Votes:
	+1: stsp

	* r939375-939376
	Fix issue #3623, a bug in foreign repository merges that causes properties
	on merge-added files to be dropped from commits of those merges.
	Justification:
	This can actually result in what is arguably a corrupt working copy,
	one that thinks it is in sync with the repository but actually is not.
	Notes:
	r939375 tweaks a test to cover this case. There's a minor conflict
	with its merge with a single deleted function call. (Keep 'theirs'.)
	r939376 is the fix. It merges cleanly.
	Votes:
	+1: cmpilato

	* r944635
	On Windows, allow building subversion against Sqlite versions that have a
	fourth number in their version string, such as 3.6.23.1.
	Justification:
	We force users to run using unpatched versions without this patch.
	Votes:
	+1: rhuijben, danielsh

	* r945708
	Fix issue #3469 "tree conflict inside of directory external causes
	assertion to fail".
	Justification:
	Assertion failures are bad and tree-conflicted externals cannot
	be reverted or resolved without this fix.
	Branch:
	^/subversion/branches/1.6.x-issue3469
	This is not a backport branch. The fix is for 1.6.x only.
	Trunk will need a separate fix as part of wc-ng.
	Votes:
	+1: stsp, julianfoad

	Veto-blocked changes:
	=====================

	* r877552, r877617, r877618, r877619, r877621
	Support using `svnversion` on files and improve some messages printed by
	`svnversion`.
	Justification:
	Some messages printed by `svnversion` are inconsistent and sometimes
	incorrect:
	$ svnversion /usr
	exported
	$ svnversion /bin/ls
	'/bin/ls' not versioned, and not exported
	$ svnversion VERSIONED_FILE
	'VERSIONED_FILE' not versioned, and not exported
	$ svnversion /nonexistent
	subversion/libsvn_wc/questions.c:90: (apr_err=2)
	svn: '/nonexistent' does not exist
	^^^
	Messages with these revisions backported:
	$ svnversion /usr
	Unversioned directory
	$ svnversion /bin/ls
	Unversioned file
	$ svnversion VERSIONED_FILE
	NUMBER
	$ svnversion /nonexistent
	'/nonexistent' doesn't exist
	Votes:
	+1: arfrever
	-0: cmpilato (Introduces new features and changes output)
	-1: rhuijben (Introduces new features and changes output;
	I think it should be removed from STATUS)

	* r878078, r878088, r878102, r878196
	Make KWallet window attached to terminal window.
	Justification:
	Without this change, KWallet window can appear behind terminal window
	and is left after killing terminal.
	Votes:
	+1: arfrever
	-0: peters
	-1: stsp, rhuijben ("/proc/%ld/stat" is ugly and Linux-specific -
	would just getenv("WINDOWID") work instead?)

	* r921453, r927184, r927243
	Fix reopened issue #3020 'Reflect dropped/renumbered revisions in
	svn:mergeinfo data during svnadmin load'
	Justification:
	Prior to this fix, when loading a partial dump with mergeinfo, the
	resulting mergeinfo in the target repository could refer to non-existent
	revisions or revisions that have nothing to do with the merge source
	in the original repository. The original fix for issue assumed that
	the dump stream was for a complete repository.
	Notes:
	r921453 and 927184 are tests, r927243 is the fix.
	Branch:
	^/subversion/branches/1.6.x-issue3020
	Votes:
	+1: rhuijben, kameshj
	-1: pburba (There is a regression with this fix, see
	http://svn.haxx.se/dev/archive-2010-03/0716.shtml)

	* r933194
	Fix for issue #2753
	SVNListParentPath feature doesn't work when svn authz is used.
	Justification:
	This issue is waiting for resolution for more than 3 years.
	Notes:
	See discussion at http://svn.haxx.se/dev/archive-2010-04/0275.shtml
	Votes:
	-1: cmpilato (pending results of aforemented discussion)

	Approved changes:
	=================