Checkout Tools
  • last updated 33 mins ago
Constraints: committers
Constraints: files
Constraints: dates

Changeset 903934 is being indexed.

Revert a portion of r903930. On second thought (and having actually

played with the site some), having menu items link to mere

descriptions of other menu items is obnoxious and violates the

principle of least surprise. If I click "About Subversion", I wanna

read about Subversion, not about where I might learn about Subversion

elsewhere. If I click "Getting Subversion", I expect to be one step

closer to actually having Subversion in my hands. And so on.

I have faith that our site visitors aren't complete and utter idiots.

* site/publish/site-nav.html

Restore the non-linky-ness of the top-level menu items.

Use the front page of the website to give a high-level overview of the

site as a whole.

* site/publish/index.html

Add an 'About This Site' section.

* site/publish/site-nav.html

Make top-level menu items links to the 'About This Site' bits.

* site/publish/news.html

New 'News' archive page.

Suggested by: julianfoad

* site/publish/style/site.css

(:target): New style def.

* subversion/bindings/javahl/tests/org/tigris/subversion/javahl/

Remove unneeded imports.

Fix a bug in obliterate, whereby obliterating an added node could fail

because it first deleted the node's node-origin record and then tried to

update the same.

* subversion/libsvn_fs_base/dag.c

(node_origins_update): If a node-origin record is not found, just ignore

it and continue.

JavaHL: Add the implemented client class, and make the compat code build

dependent upon the existing packages.

* build.conf:

(javahl-compat): Add the dependency on javahl.

[ in subversion/bindings/javahl/ ]

* src/org/apache/subversion/javahl/


* src/org/apache/subversion/javahl/

New, copied from org.tigris.subversion, and remove all the deprecated cruft.

* src/org/apache/subversion/javahl/

Remove a few more deprecated functions.

* src/org/apache/subversion/javahl/


Fix packing of format-4 FSFS repositories.

* subversion/libsvn_fs_fs/fs_fs.c


Don't try to pack revprops unless the format allows that.

A first pass at creating a JavaHL package in the org.apache.subversion

namespace. This only migrates the interface, and creates a new sub-package

for all the callbacks.

There is also some build-system munging to get this to build. To do so, run

'make javahl-compat'. Eventually, though, it should be folded into the

proper target, and everything should be built together.


Add classpath for javahl-compat, and also ensure the source directory

is part of the classpath.

* build.conf

(javahl-compat): New.

* subversion/bindings/javahl/src/org/apache/subversion/javahl:

New package, with content largely copied (and trimmed) from


* build/generator/

(TargetJavaClasses): Handle multiple packages for the same target (needed

for the javahl-compat target).

  1. … 32 more files in changeset.
* site/publish/index.html

Teeny weeny whitespace/formatting tweak only.

* site/publish/docs/community-guide/index.html

Note that this body of knowledge is "also known as HACKING".

Suggested by: julianfoad

French translation update for r903740

* subversion/po/fr.po: po-update, 6 fuzzy and 2 new fixed.

Adjust svn patch to cope with the fix made in r903733.

* subversion/libsvn_client/patch.c

(apply_one_patch): Oh my, svn_wc_delete4() suddenly works fine without

an adm access! Don't use svn_client__make_local_parents() because

that calls svn_client_add4() which tries to lock the WC again.

Use svn_wc_add4() directly instead, re-using most of the existing

logic written for the dry-run to create missing directories.

Remove access batons from wc-to-wc copies.

* subversion/libsvn_wc/wc_db.c

(kind_map): Add "none".

* subversion/libsvn_client/copy.c

(do_wc_to_wc_copies): Split and call svn_wc__call_with_write_lock.

(do_wc_to_wc_copies_with_write_lock): New.

Don't use multiple function calls as parameters when the order of

evaluation matters.

This causes 1, 4, 5 and 6 to fail. I believe this is

a bug in the patch code since reversing the order (which from a

locking point of view is silly) makes the tests work. The failing

tests claim to require a lock but only work without a lock.

* subversion/libsvn_wc/lock.c

(svn_wc__call_with_write_lock): Ensure func is evaluated

before svn_wc__release_write_lock.

Follow-up to r903153.

* subversion/libsvn_ra_neon/util.c

(wrapper_reader_cb): Fix formatting.

Found by: gstein

Remove the word "unidiff" from patch test names and docstrings.

We do not expect svn patch to handle non-unidiff types of diffs

any time soon, and it makes the test run output a bit more pretty.

* subversion/tests/cmdline/

(patch_unidiff, patch_unidiff_absolute_paths, patch_unidiff_offset,

patch_unidiff_strip1, patch_unidiff_reject, patch_unidiff_keywords):

Rename to ...

(patch, patch_absolute_paths, patch_offset, patch_strip1, patch_reject,

patch_keywords): ... these, and adjust docstrings.

Make svn patch contract keywords on input and expand them on output.

Keywords do not interfere with the patching process anymore, and are

expanded in the patched result. Also, write the correct EOL-style

to the patched result if an svn:eol-style property is set on the target.

* subversion/include/svn_subst.h

(svn_subst_stream_translated): Document support for mark/seek feature.

* subversion/libsvn_client/patch.c

(): Include svn_props.h.

(patch_target_t): Fix error in comment. New fields KEYWORDS and EOL_STYLE.

(init_patch_target): Evaluate svn:keywords and svn:eol-style. Correctly

create the translation streams for patched result and reject file.

Move everything which needs to be done if the target already exists

into the same if-block.

(read_line): Return a plain const char * rather than a stringbuf.

Contract keywords after reading the line (wrapping a translation stream

around a stream reading from a range of a file interferes with

EOF-detection in svn_stream_readline(), so we need to contract keywords

manually -- I know this sounds really wrong but it's not trivial for me

to fix).

(seek_to_line, copy_lines_to_target): Track read_line() parameter change.

(match_hunk): As previous, but also contract keywords after reading

a line from the patch file in case the patch contains expanded keywords.

(scan_for_match): Constify a pointer. Don't allow rejected hunks to

block another hunk from matching (not that they did, but it's nice to

be explicit about it).

(get_hunk_info): Improve a comment.

* subversion/libsvn_subr/subst.c

(translated_stream_mark, translated_stream_seek): New.

(svn_subst_stream_translated): Set mark and seek handlers.

* subversion/tests/cmdline/

(patch_unidiff_keywords): New test. Tests patch application where the

target contains expanded keywords but the patch does not ('svn diff'

contracts keywords so this should be the common case).

(test_list): Add new test.

Remove an outdated comment.

* subversion/libsvn_fs_base/dag.c

(svn_fs_base__dag_delete): Remove its doc string from here, as it has a

proper one at its declaration and this one is outdated.

Remove the "run_and_verify_svn_match_any()" functions from the test harness,

in favour of using the appropriate kind of ExpectedOutput class instance.

The latter is a more generic solution; the former was only used in one place.

* subversion/tests/cmdline/

(verify_depth): Use an ExpectedOutput with run_and_verify_svn() to avoid

using run_and_verify_svn_match_any().

* subversion/tests/cmdline/svntest/

(run_and_verify_svn_match_any, run_and_verify_svn_match_any2): Remove.

Tweak some doc strings in the test harness.

* subversion/tests/cmdline/svntest/

(run_and_verify_svn2): Mention the peculiar difference between the

interpretations of the expected stdout and stderr when they are strings.

* subversion/tests/cmdline/svntest/

(TestCase): Add a doc string to __init__().

(XFail): Add a place-holder for documenting the WIP parameter of __init__().

(Skip): Fix a typo.

JavaHL: Use the client interface instead of the client class itself.

[ in subversion/bindings/javahl/ ]

* native/CreateJ.cpp:

(Property): Update the constructor reference.

* src/org/tigris/subversion/javahl/

(PropertyData): Update constructor to use the less-restrictive interface,

instead of the class.

Fix a free memory read tracked down with valgrind.

* subversion/libsvn_client/mergeinfo.c

(svn_client_mergeinfo_log): Use scratch_pool rather than iterpool

to duplicate a range that must persist during the loop.

Fix a free memory read tracked down with valgrind.

* subversion/libsvn_wc/update_editor.c

(maybe_bump_dir_info): Don't destroy pools here.

(close_directory): Destroy pools after notification.

* tools/dist/

Run the python detection script only after exporting it.

* tools/dev/unix-build/Makefile.svn: Add missing line ending chars.

Redo the merge of r891672 from ^/subversion/trunk to

^/subversion/branches/1.6.x-r891672 made in r892189.

As discussed here,, this

merge inexplicably did not set mergeinfo.

As discussed here,, the

lack of mergeinfo may have something to do with some merge problems

encountered when repeating the merge of ^/subversion/branches/1.6.x-r891672

back to ^/subversion/branches/1.6.x.

* publish/packages.html: Use   in package names to

prevent ugly wrapping.

* publish/packages.html: Say 'Binary' packages in the title.
* site/publish/faq.html

* site/publish/features.html,

Fix some broken links and links to that should be aimed at

* site/publish/style/site.css

Minor style tweaks.