Below is the file 'ChangeLog' from this revision. You can also download the file.
2006-02-23 Matt Johnston <matt@ucc.asn.au>
* enumerator.{cc,hh}: avoid transferring deltas on both sides of merge
revisions, and prefer deltas to data when both are available.
See
https://savannah.nongnu.org/bugs/?func=detailitem&item_id=15846
2006-02-21 Nathaniel Smith <njs@pobox.com>
* work.cc (detach_node): This time for sure!
2006-02-21 Nathaniel Smith <njs@pobox.com>
* work.cc (detach_node): Oops, x != y != !(x == y).
2006-02-21 Nathaniel Smith <njs@pobox.com>
* work.cc (detach_node): Check if we are passed the root dir, and
error out if so.
2006-02-21 Matt Johnston <matt@ucc.asn.au>
* commands.cc (pid_file): newline-terminate the pid
2006-02-21 Richard Levitte <richard@levitte.org>
* contrib/usher.cc (server::set_hosts): Erasing the list node that
we're iterating on, then trying to go to the next node doesn't
work. Save the iterator, then increment it before erasing the
node using the saved value. No more segfaults.
2006-02-21 Matthew Gregan <kinetik@orcon.net.nz>
* sanity.hh: Work around roster_merge.cc compilation failure with
GCC 3.4.
2006-02-20 Richard Levitte <richard@levitte.org>
* monotone.cc (cpp_main), options.hh, ui.cc (redirect_log_to),
ui.hh (struct user_interface): Add --log option, to redirect the
log lines to a file.
* monotone.texi (OPTIONS): Document it.
* po/sv.po: Translate help for this option to Swedish.
* testsuite.at, tests/t_log_to_file.at: Test it.
2006-02-20 Matt Johnston <matt@ucc.asn.au>
* database.cc (remove_version): get rid of dangling deltas,
don't try to put data or deltas if they already exist.
* tests/t_db_kill_rev_locally_2.at: un-XFAIL
2006-02-20 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Add an ADD_FILE variant that allows use of
alternate databases.
* tests/t_netsync_permissions.at: Missed some cases of the
database locking race.
2006-02-19 Nathaniel Smith <njs@pobox.com>
* ChangeLog: Fixup after xxdiff lossage.
2006-02-19 Nathaniel Smith <njs@pobox.com>
* roster_merge.cc (make_lifecycle_objs): Fix test bug.
2006-02-19 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Add a REVERT_TO variant that allows use of
alternate databases.
* tests/t_netsync_permissions.at: Attempt to avoid a database
locking race in this test that is causing spurious failures by
using the new REVERT_TO variant to cause revert to be performed
using the "client" database.
2006-02-19 Nathaniel Smith <njs@pobox.com>
* roster.cc (shallow_equal): Publically expose.
* roster.cc, roster_merge.cc: Various compile fixes.
2006-02-19 Nathaniel Smith <njs@pobox.com>
* roster_merge.{hh,cc}: Make terminology more consistent.
"marking_map" type -> "markings" name, "marking_t" type ->
"marking" name.
2006-02-19 Nathaniel Smith <njs@pobox.com>
* roster.{hh,cc} (testing_node_id_source): Make this node source
available to unit tests in other files.
* roster_merge.cc (test_roster_merge_node_lifecycle): New test.
Still quite ugly.
2006-02-18 Nathaniel Smith <njs@pobox.com>
* roster_merge.cc (roster_merge): Remove obsolete FIXME.
2005-10-19 Matthew A. Nicholson <matt@matt-land.com>
* contrib/monotone.bash_completion: Update for 0.25.
2006-02-19 Matthew Gregan <kinetik@orcon.net.nz>
* cset.hh (struct editable_tree): Add commit() member function to
editable_tree.
* cset.cc (cset::apply_to): Call editable_tree::commit() after
applying any other changes.
* roster.hh, roster.cc: Empty implementation of
editable_roster_base::commit().
* work.hh, work.cc: Implementation of
editable_working_tree::commit() that ensures all detached nodes
have been reattached.
* work.hh (struct editable_working_tree): Add map for tracking
path name mappings across node detach operations.
* work.cc (editable_working_tree::detach_node): Insert path name
mappings into map.
(editable_working_tree::drop_detached_node,
editable_working_tree::attach_node): Report add/drop/rename
operations during workspace updates.
* lua.cc: Use the safer luaL_check* rather than lua_to* in
monotone_*_for_lua functions.
2006-02-18 Markus Schiltknecht <markus@bluegap.ch>
* tests/t_cvsimport_branch.at, testsuite.at: New XFAIL test for
cvs_import branch reconstruction.
2006-02-18 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_log_nofiles_merges.at: Add test for the log options
--no-files and --merges.
* testsuite.at: Add t_log_nofiles_merges.at.
2006-02-13 Nathaniel Smith <njs@pobox.com>
* roster_merge.cc (log_conflicts): Tweak string.
Add list of tests needed.
2006-02-13 Nathaniel Smith <njs@pobox.com>
* roster_merge.cc (is_clean): Simplify.
(add_roster_merge_tests):
* unit_tests.cc (init_unit_test_suite):
* unit_tests.hh (add_roster_merge_tests): Add unit test
boilerplate.
2006-02-18 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_db_kill_rev_locally_2.at: Add an XFAIL test for a
kill_rev_locally bug reported by Daniel Carosone.
* testsuite.at: Add t_db_kill_rev_locally_2.at.
* sqlite/parse.h: Regenerated parse.h from pristine SQLite 3.3.4
source. The version committed in the 3.3.4 import had a bunch of
duplicate entries.
* commands.cc (CMD(identify)): This isn't really a "workspace"
command--stick it under "debug" for lack of a better place.
* commands.cc (CMD(refresh_inodeprints)): Check for a valid
workspace rather than failing with an invariant when run outside
of a workspace.
* tests/t_revert_new_project.at: Add an XFAIL test for a bug where
reverting a file added in a new project will leave the workspace
in a bad state until MT/work is removed manually.
* testsuite.at: Add t_revert_new_project.at.
* app_state.cc, app_state.hh, commands.cc, monotone.cc,
options.hh: Add '--no-files' option to log to allow users to
exclude the list of files changed in each revision from the log
output.
* monotone.texi: Document '--no-files', and '--next' and '--diffs'
while there.
2006-02-17 Matthew Gregan <kinetik@orcon.net.nz>
* lua.cc, lua.hh, monotone.texi, std_hooks.lua, test_hooks.lua,
testsuite.at: Remove unused non_blocking_rng_ok hook.
* sqlite/*: Import SQLite 3.3.4.
2006-02-16 Patrick Mauritz <oxygene@studentenbude.ath.cx>
* netsync.cc (handle_new_connection): Netxx::Address.get_name()
returns NULL every now and then. if so, continue with "" instead
* sqlite/parse.c: move #line under all #include directives so
the compiler can't be confused by it.
2006-02-14 Richard Levitte <richard@levitte.org>
* Makefile.am (htmldir): Add variables so monotone.html is created
and installed automatically.
This is prompted by debian/monotone.html, which indicates
monotone.html should be available.
* netsync.cc (serve_connections): Correct spelling.
2006-02-13 Matthew Gregan <kinetik@orcon.net.nz>
* sanity.cc (sanity::dump_buffer): Fix a SEGV when we're in an
error unwind and about to ask the user to mail us the crash
log--we must use FL() rather than F() here, since by the time this
is called we can't rely on the i18n infrastructure being alive.
2006-02-12 Nathaniel Smith <njs@pobox.com>
* netsync.cc (serve_connections): Revert garbage that I
accidentally checked in last time...
2006-02-12 Nathaniel Smith <njs@pobox.com>
* NEWS: Add things done since last time I did this...
2006-02-12 Matthew Gregan <kinetik@orcon.net.nz>
* {cset,paths,revision,roster,sanity,vocab}.{cc,hh}: GCC 4.1
compile fixes.
2006-02-11 Richard Levitte <richard@levitte.org>
* NEWS: Removed my notice about netsync, as I just noticed it
was already mentioned under Bugs:.
2006-02-11 Nathaniel Smith <njs@pobox.com>
* configure.ac, debian/changelog, monotone.spec:
* win32/monotone.iss: Bump version to 0.26pre2.
2006-02-11 Nathaniel Smith <njs@pobox.com>
* NEWS: Add mention of validate_commit_message.
2006-02-11 Blake Kaplan <mrbkap@gmail.com>
* monotone.texi (Hooks): Added new subsection about validation
hooks, and describe validate_commit_message in it.
* std_hooks.lua (validate_commit_message): Change the second
argument to match the documentation.
2006-02-11 Matt Johnston <matt@ucc.asn.au>
* database.hh: increase checkpoint batch size from 100 to 1000
2006-02-11 Matt Johnston <matt@ucc.asn.au>
* NEWS: Fix rename example.
2006-02-11 Richard Levitte <richard@levitte.org>
* NEWS: Update with the netsync change.
2006-02-11 Nathaniel Smith <njs@pobox.com>
* NEWS: Draft for 0.26pre2.
2006-02-11 Richard Levitte <richard@levitte.org>
* netsync.cc (serve_connections): Enclose more or less everything
in a try-catch block to catch if using IPv6 failed, and to try
with just IPv4 in that case. This is important for those who
copy a IPv6-enabled binary to a system that doesn't use IPv6.
* po/sv.po: Adapt translation to the newly changed messages.
2006-02-10 Derek Scherger <derek@echologic.com>
* tests/t_drop_missing.at:
* tests/t_rename.at:
* work.cc (visit_file): attempt to improve a couple of messages;
remove some unrequired \n's
2006-02-10 Derek Scherger <derek@echologic.com>
* netsync.cc (process_anonymous_cmd, process_auth_cmd): don't
report misleading permission denied errors for branches that are
not being served
* paths.cc (find_and_go_to_workspace): delete stale comment
* tests/t_netsync_permissions.at: add test pull of branch that is
not served
2006-02-11 Timothy Brownawell <tbrownaw@gmail.com>
Adding your db is silly and confusing (what should revert do?).
So, it's not allowed any more (the db file is ignored, regardless of
what the ignore hook says).
* tests/t_add_owndb.at: remove XFAIL, use 'ls known' instead of
'ls unknown'
* testsuite.at: don't put the db in the ignore hook
* database.{cc,hh}: is_dbfile(), check if a path is the database file
* work.cc: check is_dbfile where we check the ignore hook when
walking the filesystem
* tests/t_mt_ignore.at: fix for having the db always be ignored
2006-02-10 Richard Levitte <richard@levitte.org>
* monotone.texi (Hooks): Change the example for
get_revisions_cert_trust to check "branch" certs instead of
"ancestor" ones, and thereby match the effect of the "approve"
command.
2006-02-10 Matt Johnston <matt@ucc.asn.au>
* commands.cc (CMD(checkout)): wrapping in a transaction makes
a big difference.
2006-02-09 Nathaniel Smith <njs@pobox.com>
* Makefile.am (SQLITE_SOURCES): Remove header files lost in
latest SQLite upstream import.
2006-02-09 Graydon Hoare <graydon@pobox.com>
* lua.cc (hook_validate_commit_message): make validated the
default.
2006-02-09 Richard Levitte <richard@levitte.org>
* tests/t_cvsimport.at, tests/t_cvsimport3.at,
tests/t_cvsimport_deleted_invar.at,
tests/t_cvsimport_manifest_cycle.at, tests/t_cvsimport_samelog.at,
tests/t_singlecvs.at: Changed to cope with the strictness of
CVSNT. It doesn't create a CSVROOT/history file automagically,
but CVSROOT/modules is created automatically both by the older CVS
and by CVSNT. You can't check out a group of files using revision
numbers with CVSNT. You MUST stand in the work directory for some
command with CVSNT. Finally, with CVSNT, 'cvs init' generates the
following message if you're a normal user, at least on Debian:
cvs init: Unable to register repository.
cvs init: Your login may not have sufficient permissions to modify the
cvs init: global server settings.
cvs init: Repository /home/levitte/cvsfoo initialised
2006-02-09 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Remove duplicate line.
2006-02-08 Matthew Gregan <kinetik@orcon.net.nz>
* sqlite/*: Import SQLite 3.3.3.
* Makefile.am: Adjust for new and removed files in import.
* NEWS: Make a note of the SQLite database format change.
* database.cc (database::load): Rewrite so that we don't need any
local changes to SQLite.
2006-02-08 Richard Levitte <richard@levitte.org>
* testsuite.at, tests/t_ls_changed.at: News test, for "list
changed".
* monotone.texi, monotone.1: Document "list changed".
* po/sv.po: Correct translations of changed messages, translate
new messages.
* commands.cc (ls_changed, CMD(list)): Add a new command, "list
changed", to list changed files, always sorted in lexical order.
2006-02-06 Blake Kaplan <mrbkap@gmail.com>
* commands.cc CMD(commit): Call a new lua hook to validate the commit
message. Don't ignore -m "" when it's passed on the command line.
* lua.cc, lua.hh: Add a new hook that validates a given commit message
and passes in the added files, deleted files, and modified files.
* std_hooks.lua: Give a default hook to validate commit messages. This
currently disallows empty messages, as monotone currently does.
2006-02-05 Benoît Dejean <benoit@placenet.org>
* ui.cc (tick_write_count::write_ticks): Reverted lexical_cast,
back to F().
2006-02-05 Benoît Dejean <benoit@placenet.org>
* ui.cc (tick_write_count::write_ticks): Fixed utf8 handling.
boost::format + locale sucks, it ouputs "1\u+ffff24" for
"%d" % 1024.
2006-02-02 Emile Snyder <emile@alumni.reed.edu>
* merge.cc (resolve_merge_conflicts): If the merge has non-content
conflicts warn the user and abort before making them do content
merges. Move logging/warning of conflicts to methods in the
roster_merge_result struct.
* roster_merge.{cc,hh} (log_conflicts,
warn_non_content_conflicts): New methods to do the busy work of
logging or warning.
2006-02-02 Emile Snyder <emile@alumni.reed.edu>
* netsync.cc (run_netsync_protocol): Print a warning if either the
include or the exclude branch pattern have a ' or a " character
anywhere in them. Hopefully will help users running in a Windows
shell when they do: monotone sync myserver 'mybranch*' and nothing
happens because the ' characters are passed in to monotone.
2005-01-30 Emile Snyder <emile@alumni.reed.edu>
* annotate.cc: (do_annotate) Add ticker for revisions processed to
annotate command.
2006-01-29 Richard Levitte <richard@levitte.org>
* po/sv.po: Update a few translations, leave a number of fuzzy
ones until we've decided how "workspace" should be translated.
2006-01-27 Matthew Gregan <kinetik@orcon.net.nz>
* *: Use the term 'workspace' consistently throughout monotone for
the concept we previously described interchangably using the two
terms 'working copy' and 'working directory'. This change has
been made everywhere except in historical documentation (NEWS and
ChangeLog).
2006-01-27 Richard Levitte <richard@levitte.org>
* monotone.texi (Generating Keys): Correct small type, the keys
are not stored in /home/jim/.monotone/monotonerc.
2006-01-26 Derek Scherger <derek@echologic.com>
* app_state.{cc,hh}:
* commands.cc (log):
* monotone.cc:
* options.hh: allow --next to view upcoming changes
* tests/t_log_depth.at: rename to ...
* tests/t_log_last_next.at: ... this since log now uses
--last/--next and not --depth
* testsuite.at: rename t_log_depth.at to t_log_last_next.at
2006-01-27 Matthew Gregan <kinetik@orcon.net.nz>
* monotone.texi: Clean up 'serve' syntax in a couple of places.
2006-01-26 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Generating Keys, Network Service Revisited):
Clean up some key-related stuff -- genkey doesn't need a db, so
don't confuse the reader by giving it one, and update the sample
output too.
2006-01-27 Stéphane Gimenez <dev@gim.name>
* sanity.hh: fix scoping issue revealed by gcc-4.1.
2006-01-26 Richard Levitte <richard@levitte.org>
* po/sv.po: One more translation.
2006-01-25 Matthew Gregan <kinetik@orcon.net.nz>
* netsync.cc (session::rebuild_merkle_trees): Ticker header
consistency tweak.
2005-01-25 Matt Johnston <matt@ucc.asn.au>
* HACKING: add some notes about compiling - precompiled headers,
-O0, and ccache.
* INSTALL: mention --enable-pch
2006-01-25 Richard Levitte <richard@levitte.org>
* po/sv.po: Fix all fuzzy translations (some needed no fix).
Translate the last untranslated ones.
2005-01-24 Timothy Brownawell <tbrownaw@gmail.com>
Make a netsync client crash not hang the testsuite.
* tests/t_netsync_single.at: use netsync macros
* testsuite.at: Make NETSYNC_SETUP set a trap on exit to kill
any servers.
2005-01-24 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (update): Remove fixme comment.
2005-01-24 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (update): Allow backwards/sideways updates.
tests/t_update_to_revision.at: remove XFAIL
2006-01-23 Nathaniel Smith <njs@pobox.com>
* database.{cc,hh}: Clean up code formatting a bit, rename some
variables, rely more on automatic conversion.
2006-01-24 Vinzenz Feenstra <evilissimo@c-plusplus.de>
* query_args.hh: Introduced struct query_args and struct
query_args_param. Used to typesafe arguments for database::execute
and database::fetch
* database.cc/.hh: Adjusted database to the new fetch and execute
argument style via operator% which is more typesafe.
2006-01-23 Nathaniel Smith <njs@pobox.com>
* database.cc (assert_sqlite3_ok): Improve the hint message given
for SQLITE_ERROR's.
2006-01-23 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_database_check_normalized.at: update included database to
use rosters
2006-01-23 Timothy Brownawell <tbrownaw@gmail.com>
New ChangeLog utilities in contrib/ .
* contrib/ChangeLog.sh: Script that takes the date, author, and
changelog certs from the last n revisions and prints them in standard
ChangeLog format.
* contrib/edit_comment_from_changelog.lua: an edit_comment hook that
takes the initial commit message from the most recent ChangeLog entry
2006-01-23 Henry Nestler <henry@bigfoot.de>
* monotone.texi, monotone.1: Mode none for --ticker.
2006-01-23 Nathaniel Smith <njs@pobox.com>
* commands.cc (update): Hopefully improve wording of help a bit.
2006-01-21 Nathaniel Smith <njs@pobox.com>
* netcmd.cc (read): Expand the error message we issue on protocol
version mismatch to be a bit more informative.
2006-01-21 Matthew Gregan <kinetik@orcon.net.nz>
* configure.ac: Fix up Windows and IPv6 tests after the change
from AC_TRY_RUN to AC_TRY_COMPILE. Also a couple of other minor
cleanups.
* Makefile.am: Remove win32/wcwidth.c from WIN32_PLATFORM_SOURCES.
2006-01-21 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_db_execute.at: Play with the files table instead of the
revisions table, so we don't trigger the has-revisions-but-no-rosters
check.
2006-01-21 Richard Levitte <richard@levitte.org>
* po/sv.po: Sort the translation guidelines a little more.
2006-01-21 Joel Rosdahl <joel@rosdahl.net>
* po/sv.po: Corrections in translation guidelines.
2006-01-21 Joel Rosdahl <joel@rosdahl.net>
* AUTHORS: Added myself as a translation contributor.
* po/sv.po: Second round of review of Swedish translation.
2006-01-20 Richard Levitte <richard@levitte.org>
* po/sv.po: Added a \n at the end of a msgstr that was missing
it.
2006-01-20 Joel Rosdahl <joel@rosdahl.net>
* po/sv.po: Review and suggested corrections of Swedish
translation.
2006-01-20 Timothy Brownawell <tbrownaw@gmail.com>
* netsync.cc: Make our sockets non-blocking.
2006-01-20 Matt Johnston <matt@ucc.asn.au>
* commands.cc CMD(setup): default to current dir
* monotone.texi: update
* t_setup_existing_path.at: add test
2006-01-19 Emile Snyder <emile@alumni.reed.edu>
Add a --brief mode to the annotate command which prints more
informative annotations rather than just the raw revision ids.
* commands.cc (CMD(annotate)): Add --brief option to the annotate
command and remove obsolete comment lines.
* annotate.cc (dump): Pass in app_state to allow access to db for
cert lookups on revisions. Honor new --brief
flag by printing <short id>.. by <author> <date>: as the
annotation rather than the raw revision id, and only printing the
annotation for the first line of each block of lines from the same
revision.
(cert_string_value): Given a set of certs from a revision, a cert
name, and some detail of what part of the value we want, find and
return that part of the cert.
(build_revisions_to_annotations): After we finish with the
annotations pass, build up a mapping of revision id to annotation
string.
* monotone.texi: Add --brief flag and description to the annotate
section of the command reference.
2006-01-19 Matt Johnston <matt@ucc.asn.au>
* work.{cc,hh}, commands.cc: add "rename src1 [src2 ...] dst/"
syntax.
* monotone.texi: update
* testsuite.at, tests/t_rename_destdir: new test (is incomplete).
* tests/t_no_rename_overwrite.at: syntax should now succeed.
* vocab.cc: add hexenc<id> dump() instantiation.
2006-01-19 Matt Johnston <matt@ucc.asn.au>
* HACKING: escape the colon in the cino vim modeline option.
2006-01-19 Nathaniel Smith <njs@pobox.com>
* database.cc (check_format): Small cleanups to previous change.
2006-01-19 Nathaniel Smith <njs@pobox.com>
Fix bug reported by Henry Nestler, where 'serve' did not detect
that the user had not run 'db rosterify'.
* database.cc (check_rosterified): Rename to...
(check_format): ...this. Detect unchangesetified dbs too.
(sql, ensure_open_for_format_changes, get_roster_id_for_revision):
* revision.cc (build_roster_style_revs_from_manifest_style_revs)
(build_changesets_from_manifest_ancestry): Adjust accordingly.
* tests/t_check_db_format.at, testsuite.at: New test.
2005-01-18 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_lua_privkey.at: Remove; this tests a hook that was removed.
* testsuite: adjust accordingly
2006-01-19 Matthew Gregan <kinetik@orcon.net.nz>
* keys.cc (get_passphrase): Update std::map correctly--remove
existing cached passphrase for a given keyid, then store the new
one using safe_insert().
(make_signature): Fix crash when running monotone with persistent
passphrases disallowed--remove shared_ptr that was going out of
scope too early, and shadowed and existing shared_ptr in the
correct scope.
* tests/t_no_persist_phrase.at: New test to check that basic no
persistent passphrase functionality works.
* testsuite.at: Include new test.
2006-01-18 Timothy Brownawell <tbrownaw@gmail.com>
* merkle_tree.{cc,hh}: Add a locate_item() function to find the
node and slot in a merkle_table that a given item is in.
* refiner.cc (refiner::process_refinement_command): If they have
a leaf and we have a subtree then if their leaf is in our subtree,
we need to tell them.
2006-01-18 Nathaniel Smith <njs@pobox.com>
* configure.ac: Remove some dead comments.
2006-01-18 Nathaniel Smith <njs@pobox.com>
* configure.ac: s/TRY_RUN/TRY_COMPILE/ a few places, for general
cleanliness.
2006-01-18 Richard Levitte <richard@levitte.org>
* po/sv.po: Better translation of "branch". Inspired from
subversion.
2006-01-18 Nathaniel Smith <njs@pobox.com>
* database.cc (begin_transaction): Turn an E() into an I().
2006-01-15 Richard Levitte <richard@levitte.org>
* po/sv.po: Almost all translations done. I've left a few for
lack of inspiration. Someone, PLEASE REVIEW!
2006-01-15 Nathaniel Smith <njs@pobox.com>
* AUTHORS: Add Richard as a translator.
2006-01-15 Nathaniel Smith <njs@pobox.com>
* roster.cc (union_corpses): New function.
(unify_roster_oneway, unify_rosters): Remove unused new_ids
argument. Add call to union_corpses. Add big comment explaining
what's going on.
(test_unify_rosters_end_to_end_ids)
(test_unify_rosters_end_to_end_attr_corpses): Split and improve
tests.
2006-01-15 Richard Levitte <richard@levitte.org>
* po/sv.po: More translations done.
2006-01-15 Richard Levitte <richard@levitte.org>
* LINGUAS: Change sv_SE to just sv.
* po/sv_SE.po: Rename ...
* po/sv.po: ... to this, as there aren't so many dialects. Update
with more strings to translate.
* std_hooks.lua: xgettext whines about a missing quote, so add
one.
2006-01-15 Nathaniel Smith <njs@pobox.com>
* roster.cc (test_unify_rosters_end_to_end): Add failing test for
unify_roster's handling of attr corpses.
2006-01-15 Nathaniel Smith <njs@pobox.com>
* cert.cc (load_key_pair):
* keys.cc (require_password): Use accessor method, so as to
actually compile...
2006-01-15 Matt Johnston <matt@ucc.asn.au>
* ui.cc: make tickers saner (203 K vs 0.2 M).
Print full byte counts below 1 K.
2006-01-14 Nathaniel Smith <njs@pobox.com>
* lua.cc (hook_get_key_pair): Remove.
* keys.cc (require_password): Adjust accordingly.
* cert.cc (priv_key_exists, load_key_pair): Likewise.
2006-01-14 Richard Levitte <richard@levitte.org>
* configure.ac: Make sure there's an empty xgettext.opts
when xgettext accepts --flag. This avoids getting an error
message later on when cat can't find it to append to
po/Makefile.in.
* po/LINGUAS, po/sv_SE.po: New Swedish translation. Needs
reviewing at this stage.
2006-01-13 Richard Levitte <richard@levitte.org>
* monotone-notify.pl: Make it possible for the user to say what
the monotone binary is called or where it is. Adapt to revision
output changes and to changes in the monotone UI.
2006-01-12 Nathaniel Smith <njs@pobox.com>
* po/Makevars (XGETTEXT_OPTIONS): This time for sure!
2006-01-13 Nathaniel Smith <njs@pobox.com>
* database.cc (check_rosterified): Clarify message emitted when db
has not been rosterified.
2006-01-12 Nathaniel Smith <njs@pobox.com>
* po/Makevars (XGETTEXT_OPTIONS): More fiddling to make
intltool-update happy.
2006-01-13 Bruce Stephans <monotone@cenderis.demon.co.uk>
* lua.cc (shared_trust_function_body): Index lua arrays
from 1, not 0.
2006-01-12 Nathaniel Smith <njs@pobox.com>
* sanity.cc (dump_buffer): Add forgotten newline.
2006-01-12 Nathaniel Smith <njs@pobox.com>
* roster.cc (make_roster_for_merge): Add more MM()'s.
2006-01-12 Matthew Gregan <kinetik@orcon.net.nz>
* contrib/{color-logs.{conf,sh},monotone.el}: Changes to handle
renaming of --no-merges.
* app_state.{cc,hh}, commands.cc, monotone.cc, options.hh: Make
'log' default to not displaying merges. Rename '--no-merges'
option to '--merges'.
2006-01-11 Nathaniel Smith <njs@pobox.com>
* commands.cc (diff): gettext'ify the "no changes" tag.
2006-01-11 Richard Levitte <richard@levitte.org>
* Makefile.am: Add contrib/usher as an extra program. This works
as a reminder that this program exists, and provides supported
ways to build it.
* contrib/usher.cc: Add a -m switch, to be able to specify the
name or complete path of/to the monotone binary, in case there
are several floating around on the system.
2006-01-11 Matt Johnston <matt@ucc.asn.au>
* Makefile.am: pch.hh should be included
2006-01-11 Nathaniel Smith <njs@pobox.com>
* sanity.hh (struct i18n_format): Make internationalized format
strings and non-internationalized format strings type-distinct.
(This hopefully sets up for fixing some of the charset mess later
down the road.)
(FL): New macro; like F but without the i18n glue.
Use FL with L, and F/FP with everything else.
* *.hh, *.cc: Adjust existing F/L/etc. usage accordingly.
2006-01-11 Nathaniel Smith <njs@pobox.com>
* HACKING: Add section listing all the single-character macros.
2006-01-11 Nathaniel Smith <njs@pobox.com>
* database.cc (fetch): Don't log all SQL unless requested by
user.
2006-01-11 Nathaniel Smith <njs@pobox.com>
* po/Makevars (XGETTEXT_OPTIONS):
* configure.ac: Tweak xgettext configuration to make intltool
happier.
2006-01-11 Nathaniel Smith <njs@pobox.com>
* netsync.cc (process_confirm_cmd): Remove.
2006-01-11 Matthew Gregan <kinetik@orcon.net.nz>
* platform.hh, win32/wcwidth.c: Remove local wcswidth
implementation--it's no longer needed.
2006-01-10 Derek Scherger <derek@echologic.com>
* tests/t_revert_unchanged.at: new test to check that reverting
nothing but unchanged files doesn't end up with an empty
restriction and revert everything
* testsuite.at: call it
2006-01-10 Nathaniel Smith <njs@pobox.com>
* packet.cc (packet_roundabout_test): Test rdata packets too.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* netsync.cc (session):
* commands.cc (read): Adjust to match previous change.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* packet.{hh,cc} (struct packet_db_writer): Remove old guard
against accidentally ingesting public keys -- this has not been
useful for some years.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* diff_patch.hh (struct content_merger): Oops, forgot to remove
the merge2 prototype.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* std_hooks.lua:
* lua.cc (hook_merge2):
* diff_patch.cc (try_to_merge_files): Remove merge2 code, since we
no longer do merge2's. (We can always revive it if we add suture
support.)
2006-01-10 Nathaniel Smith <njs@pobox.com>
* packet.cc (consume_file_delta): Remove unused 4-argument form.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* lua.{cc,hh} (hook_resolve_file_conflict)
(hook_resolve_dir_conflict): Remove dead code.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* tests/t_selectors_b_h.at, testsuite.at: New test.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* database.cc (manifest_cert_exists, put_manifest_cert):
(get_manifest_cert, get_manifest_certs): Remove unused methods.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* database.cc (put_reverse_version, put_file_reverse_version):
Whoops, missed some dead code.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* revision.cc (build_roster_style_revs_from_manifest_style_revs):
Remove unused variable; silence gcc warning.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* packet.cc (consume_file_reverse_delta)
(struct feed_packet_consumer): Remove support for nonexistent
"frdelta" packet type.
* database.{hh,cc} (put_reverse_version)
(put_file_reverse_version): Remove unused methods.
* key_store.cc (struct keyreader): Remove obsolete methods.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* database.cc (set_filename): Simplify slightly.
(delta_exists): Remove unused 3-argument version.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* commands.cc (reindex): Remove.
* database.cc (database::rehash): Likewise.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* change_set.{cc,hh}: Remove.
2006-01-10 Nathaniel Smith <njs@pobox.com>
* contrib/ciabot_monotone.py (Monotone.get_revision): Update to
use 'automate get_revision' instead of 'cat revision'.
2006-01-10 Richard Levitte <richard@levitte.org>
* lua.cc (monotone_parse_basic_io_for_lua): Don't trust the
returned value from lua_tostring() to stick around. Instead, copy
it into a regular std::string.
* std_hooks.lua (get_netsync_read_permissions,
get_netsync_write_permissions): Make sure to properly close the
permission files.
2006-01-09 Richard Levitte <richard@levitte.org>
* contrib/usher.cc, monotone.1, monotone.cc, po/fr.po, po/ja.po,
po/pt_BR.po, testsuite.at: Change the default port from 5253 to
4691 (assigned to us by IANA).
2006-01-07 Patrick Mauritz <oxygene@studentenbude.ath.cx>
* schema_migration.cc: add safe-guard against passing a null pointer
into string handling. (necessary with sun studio)
2006-01-05 Derek Scherger <derek@echologic.com>
* app_state.{cc,hh} (set_restriction): remove "respect_ignore"
flag and don't complain about ignored files
* commands.cc (status, list, diff): allow --exclude option for
restrictions
* restrictions.hh: add a comment about restricted command
consistency
* tests/t_restricted_commands_consistent.at: new test to give that
bark some bite
* testsuite.at: call it
2005-12-29 Nathaniel Smith <njs@pobox.com>
* NEWS: Write up for 0.25.
2005-12-29 Matthew Gregan <kinetik@orcon.net.nz>
* sqlite/{pager.c,sqlite3.h}: Upgrade to SQLite 3.2.8.
2005-12-29 Nathaniel Smith <njs@pobox.com>
* NEWS: Notes on what to write up for 0.25.
* win32/monotone.iss, monotone.spec, debian/changelog:
* configure.ac, UPGRADE: Bump.
2005-12-26 Matt Johnston <matt@ucc.asn.au>
* commands.cc (CMD(revert)): revert with no args prints usage,
mention "." for entire working copy.
2005-12-21 Matt Johnston <matt@ucc.asn.au>
* commands.cc (dump_difs, CMD(log)): only print --diffs for the file
of interest.
2005-12-21 Matthew Gregan <kinetik@orcon.net.nz>
* botan/es_capi.{cpp,h}: Windows compile fixes for Botan 1.4.10.
2005-12-21 Matt Johnston <matt@ucc.asn.au>
* upgrade to Botan 1.4.10
2005-12-20 Daniel Carosone <dan@geek.com.au>
* monotone.texi: Rearrange the description of monotone serve into
two sections; the first one describing basic steps necessary to
get Jim and the reader through to the next sections, and a later
more advanced section that discusses some of the subleties and
choices and moves them on to running a dedicated server. Other
small tweaks in nearby sections while here.
2005-12-18 graydon hoare <graydon@pobox.com>
* AUTHORS: Add Roland.
2005-12-17 Roland McGrath <roland@redhat.com>
* work.hh (struct file_itemizer): Remove extraneous qualifier on
visit_file declaration.
* app_state.hh (class app_state): Remove extraneous qualifier on
set_restriction declaration.
2005-12-14 Matthew Gregan <kinetik@orcon.net.nz>
* win32/fs.cc: Handle the fact that the MoveFileEx symbol might
exist in kernel32.dll even though it doesn't actually work. It
turns out that Win9x implements a bunch of NT-only symbols as
stubs that return ERROR_CALL_NOT_IMPLEMENTED, so it's not
sufficient to detect feature availability by doing symbol lookups
at runtime. Also add a missing parameter to the final error
message. Fixes #15063.
2005-12-13 Derek Scherger <derek@echologic.com>
* commands.cc (update): remove \n from F() string
(revert): display "reverting..." messages similar to update
* tests/t_revert.at: allow output from revert
2005-12-13 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Use SIGTERM rather than SIGSEGV to close down
monotone servers.
* HACKING, tests/README: Remove references to use of SEGV for
terminating monotone servers.
2005-12-13 Timothy Brownawell <tbrownaw@gmail.com>
* constants.cc: New netsync is incompatible, bump protocol version.
* constants.hh: Change default port to the one IANA assigned us.
2005-12-10 Timothy Brownawell <tbrownaw@gmail.com>
* roster.cc (make_roster_for_merge): unshadow the node_id_source
argument. Resolve usages to the argument or the local by educated guess.
2005-12-10 Timothy Brownawell <tbrownaw@gmail.com>
* paths.{cc,hh}: operator<() for split_path's
2005-12-10 Timothy Brownawell <tbrownaw@gmail.com>
* sanity.{cc,hh}: Fix musing dumps so that the header line always gets
printed, even if printing the body throws an exception.
* database.cc, packet,cc, revision.cc: sprinkle MMs
* cset.cc: sanity check print_cset to match parse_cset
2005-12-08 Matthew Gregan <kinetik@orcon.net.nz>
* po/pt_BR.po: Add missing newline.
* AUTHORS: Add Alex Queiroz to list of translators.
2005-12-08 Alex Queiroz <asandroq@gmail.com>
* po/pt_BR.po: Beginnings of the Brazilian Portuguese translation.
* po/LINGUAS: Add it.
2005-12-06 Timothy Brownawell <tbrownaw@gmail.com>
* packet.cc (packet stream reading): Fix handling of privkey packets.
It would convert pubkey+privkey to keypair, but not change the private
key to the new encryption type. Now uses migrate_private_key to
convert privkey->keypair. Reading privkey packets from stdin only works
if the get_passphrase hook is defined for that key.
Fix packet_roundabout_test; the send+receive part was using an empty
stream, rather than the packet sequence generated in the first part.
* packet.hh, commands.cc, key_store.cc: read_packets now takes an
app_state argument.
* tests/t_dropkey_2.at: use genkey instead of reading packets
* tests/t_read_privkey.at: check that reading a privkey packet works
* testsuite.at: add read_privkey test
2005-12-06 Matthew Gregan <kinetik@orcon.net.nz>
* ui.cc (write_ticks): Move CR out of tickline2; output it where
required instead. Fixes brokeness of 'count' ticker on Win32 (and
perhaps elsewhere) that was causing each line redraw to be written
on a new line.
* contrib/usher.cc (fork_server): Avoid getting stuck in an
infinte loop attempting to read() from a stream that has
reached EOF before we've read anything from it. This can occur if
the fork()+exec() of an usher managed monotone server fails.
* ui.cc (write_ticks): Don't call display_width() repeatedly for
the same string--get the result once and reuse it where necessary.
* netxx/sockopt.cxx, netxx/sockopt.h (set_ipv6_listen_for_v6_only):
New member function to set the IPV6_V6ONLY flag on a socket.
* netxx/serverbase.cxx (bind_to): Call set_ipv6_listen_for_v6_only
on IPv6 sockets.
* app_state.cc (allow_working_copy): Use keydir path from
MT/options if set.
2005-12-05 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_database_sig_cleanup.at: Mark this as NOT_ON_WIN32; as
per the MSDN link I've added to the test, we can't handle
SIGINT/SIGTERM gracefully on Windows.
2005-12-04 Nathaniel Smith <njs@pobox.com>
* paths.cc (file_path): MM() the path we're validating.
2005-12-04 Nathaniel Smith <njs@pobox.com>
* schema_migration.cc: Modernize error checking a bit. (Use E()
and I() instead of throwing runtime_errors.)
2005-12-03 Nathaniel Smith <njs@pobox.com>
* sanity.cc (dump_buffer): If we dumped debug info, write out a
note asking that it be included in bug reports.
2005-12-03 Nathaniel Smith <njs@pobox.com>
* schema_migration.cc (logged_sqlite3_exec): New function.
Use everywhere in this file. Enables logging of migration
commands for debugging purposes.
2005-12-03 Nathaniel Smith <njs@pobox.com>
* tests/t_mixed_case_pwd.at, testsuite.at: New test.
2005-12-02 Matthew Gregan <kinetik@orcon.net.nz>
* Makefile.am: Make Win32 build link against shfolder.lib to get
SHGetFolderPathA on older Windows platforms (we had previously
been relying on it to be available in shell32.lib) .
* packet.cc (feed_packet_consumer::feed_packet_consumer,
extract_packets): Move duplicated sets of regexs for key IDs,
certs, etc. into constants.cc.
* constants.{cc,hh}: Add regex strings extracted from packet.cc. Also
fix the key ID regex to accept a few characters we allowed in
legal_key_name_bytes already.
* tests/t_genkey.at: Test good and bad keys with all characters we
allow in the user portion of a key id.
2005-12-01 Matthew Gregan <kinetik@orcon.net.nz>
* database.cc: Revert last change; db version and db migrate need
to be able to open databases with old schema versions. Calling
sql() directly broke this. Moved a db existence check scattered
through the code into db_check_exists(), and added calls to this
into ::version and ::migrate.
* database.cc (database::version, database::migrate): Improve
error handling where user has supplied the path to a non-existant
database file. Submitted by Neil Conway.
* win32/fs.cc (rename_clobberingly_impl): Improve error handling
around LoadLibrary call.
* lua/{lvm.c,lgc.c,lapi.c}: Lua post-5.0.2 bugfixes from
http://www.lua.org/bugs.html. Submitted by Alex Queiroz.
2005-11-29 Matt Johnston <matt@ucc.asn.au>
* Makefile.am: add PCH_FLAGS only when compiling monotone and
unit_tests objects, not for libs.
* pch.hh: define __STDC_CONSTANT_MACROS as required for UINT32_C.
2005-11-28 Nathaniel Smith <njs@pobox.com>
* txt2c.cc (main): Work correctly even on empty files.
2005-11-28 Nathaniel Smith <njs@pobox.com>
* schema_migration.cc: Provide more feedback while 'db migrate' is
running.
2005-11-28 Matthew Gregan <kinetik@orcon.net.nz>
* HACKING: Make a note of the fact that we kill 'monotone serve'
processes with a SIGSEGV.
* tests/README: Minor cleanup, mention SIGSEGV thing here too.
2005-11-27 Nathaniel Smith <njs@pobox.com>
* std_hooks.lua: Make regexp matching error slightly more
user-friendly.
2005-11-27 Nathaniel Smith <njs@pobox.com>
* commands.cc (diff): Remove OPT_BRANCH from diff's options.
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* configure.ac: Do not use test's == operator because it is not
compatible with many implementations; use = instead.
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* configure.ac, ui.cc: Windows does have sync_with_iostream; the
problem is that it does not behave correctly on some MinGW versions.
Properly detect this condition. Thanks to Matthew Gregan for the
explanation.
2005-11-27 Matthew Gregan <kinetik@orcon.net.nz>
* netxx/address.cxx (parse_uri): Reset URI parser state after
checking for IPv6 address.
2005-11-27 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_netsync_largish_file.at: Move LARGISH_FILE_CREATE
definition from here...
* testsuite.at: ...to here.
* tests/t_dump_load.at: Use LARGISH_FILE_CREATE to create a 1MB
file so that the db dump output is larger than a few kB. This is
needed to catch the iostreams-returning-EOF-early bogosity under
MinGW when sync_with_stdio(false) has been called on the iostream.
2005-11-27 Nathaniel Smith <njs@pobox.com>
* monotone.cc: Include "i18n.h" instead of "libintl.h" directly,
to theoretically play better with autoconf stuff.
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* tests/t_automate_stdio.at: Don't expect perl to be in /usr/bin;
this assumption doesn't hold breaks on several systems (such as
NetBSD).
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* tests/t_checkout_options.at: Use the [] operator instead of {} in
a shell pattern so that it works with interpreters other than bash
(e.g., NetBSD's sh).
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* testsuite.at, tests/t_empty_env.at, tests/t_i18n_file.at,
tests/t_netsync_sigpipe.at, tests/t_update_1.at, tests/t_update_2.at:
Do not use test's == operator because it is not portable; use =
instead.
2005-11-27 Matt Johnston <matt@ucc.asn.au>
* Makefile.am, configure.ac, pch.hh: add --enable-pch configure
flag to enable precompiled boost headers. Based on LyX automake/autoconf
rules, Zbynek Winkler suggested that just boost headers would be
beneficial.
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* app_state.cc: No need to include headers for chdir() any more.
This function was replaced by change_current_working_dir.
* monotone.cc: Unconditionally include libintl.h as this file uses
gettext functions.
* configure.ac, rcs_import.cc: Check for strptime(3) and use the
check results instead of assuming that only Windows lacks this
function.
* configure.ac, ui.cc: Check whether C++ streams support the
sync_with_stdio method and use the check results instead of assuming
that only Windows lacks this function.
Addresses part of bug #12086.
2005-11-27 Julio M. Merino Vidal <jmmv@NetBSD.org>
* app_state.cc: Correctly detect when MT/options does not have a
database property instead of printing the cryptic message "misuse:
invalid path ''".
2005-11-27 Grahame Bowland <grahame@angrygoats.net>
* configure.ac: expose --enable-ipv6 in config.h
* monotone.cc (cpp_main): parse IPv6 addresses correctly
in --bind arguments
* netsync.cc (call_server,serve_connections): if USE_IPV6 defined,
create netxx Address instances with ipv6 enabled.
* netxx/address.cxx (parse_uri): parse IPv6 addresses as hostnames
* closes bug #14446
2005-11-27 Grahame Bowland <grahame@angrygoats.net>
* automate.cc (automate_stdio_read): remove check for EINTR
that was breaking win32 build.
2005-11-27 Grahame Bowland <grahame@angrygoats.net>
* automate.cc (automate_certs,automate_keys): use non-exclusive
transaction guard
* commands.cc (ls_certs,ls_keys,cat): use non-exclusive
transaction guard
* database.cc (begin_transaction): used BEGIN DEFERRED (only attempt
exclusive lock once a write is attempted on DB) when not exclusive
(transaction_guard::transaction_guard): new argument exclusive indicating
whether an immediate exclusive lock is requested, defaults to true
* database.hh: update prototype for transaction_guard
2005-11-26 Nathaniel Smith <njs@pobox.com>
* NEWS: Fill in date for 0.24 release.
2005-11-26 Nathaniel Smith <njs@pobox.com>
* UPGRADE: Add note about serve changing syntax.
2005-11-26 Nathaniel Smith <njs@pobox.com>
* UPGRADE: Update for 0.24.
* configure.ac, win32/monotone.iss, monotone.spec:
* debian/changelog: Bump version number.
2005-11-26 Nathaniel Smith <njs@pobox.com>
* ChangeLog: fixup after merge.
2005-11-27 Daniel Carosone <dan@geek.com.au>
* monotone.texi, std_hooks.lua: sync hook definitions with
std_hooks.lua, and group the descriptions into @subsections,
reordering a few items accordingly.
2005-11-26 Nathaniel Smith <njs@pobox.com>
* Makefile.am: Remove botan/{omac,fips_rng}.h.
2005-11-26 Julio M. Merino Vidal <jmmv@NetBSD.org>
* AUTHORS, ChangeLog: Update my email address; I don't read the
other one any more and the new one represents better my limited
involvement in Monotone.
2005-11-26 Julio M. Merino Vidal <jmmv@NetBSD.org>
* configure.ac, Makefile.am, po/Makevars: Avoid using xgettext's
--flag option on versions that do not support it.
2005-11-26 Grahame Bowland <grahame@angrygoats.net>
* automate.cc (automate_stdio) add wrapper function to read()
so that loops will not wind backwards possibly overrunning buffers
if an error occurs. Hopefully fixes some of the strange edge cases
seen using automate. (Closes #15062 in bug tracker)
2005-11-26 Matthew Gregan <kinetik@orcon.net.nz>
* botan/mem_pool.cpp (Pooling_Allocator::allocate): Botan's
Pooling_Allocator assumes that Buffers stored in free_list are
sorted, and uses std::inplace_merge() in deallocate() to ensure
the recently freed Buffer is positioned in free_list
appropriately. This same check was not being performed when a
Buffer was added to free_list in allocate() (it's not safe to
assume that the address of a newly allocated Buffer will always be
greated than existing Buffers).
2005-11-25 graydon hoare <graydon@pobox.com>
* database.cc (get_version): Another important fix: make sure to
cancel overlapping paths, to avoid exponential memory requirement
in deep, heavily-forked-and-merged storage paths.
2005-11-25 graydon hoare <graydon@pobox.com>
* database.cc (get_version): Another important fix: make sure to
cancel overlapping paths, to avoid exponential memory requirement
in deep, heavily-forked-and-merged storage paths.
2005-11-25 graydon hoare <graydon@pobox.com>
* database.cc (get_version): Crucial fix! Rewrite the
delta-reconstruction algorithm to work with a DAG storage topology,
rather than assuming an inverted tree. Counterpart to the
delta-storage optimization made by Timothy Brownawell on
2005-11-05.
2005-11-23 Matt Johnston <matt@ucc.asn.au>
* botan/*: import of Botan 1.4.9
2005-11-23 Grahame Bowland <grahame@angrygoats.net>
* automate.cc (automate_stdio): Fix partial reads cloberring
start of buffer.
2005-11-22 Matthew Gregan <kinetik@orcon.net.nz>
* database.cc (database::dump): Don't include SQLite internal
tables in the dump.
* tests/t_dump_load.at (database dump/load): Run a db analyze
before dumping the db--catches cases where we dump internal db
tables that we can't reload.
2005-11-21 Henry Nestler <Henry@BigFoot.de>
* contrib/mtbrowse.sh: Version 0.1.13
Handle authors without '@'. Minor box sizing.
Get Version of monotone before starts anything.
DEPTH_LAST detects for version >0.19, no save to config.
'cat revision' replaced with 'automate get_revision'.
ANCESTORS "M" changed to "A". Default "L" (monotone log --brief).
Warn user about misconfigurations. Temp files without tailing dot.
New option: Sort by Date/Time, up or down.
Fix: Parents of merge, if certs list from cache (ncolor).
2005-11-21 Matthew Gregan <kinetik@orcon.net.nz>
* NEWS: Minor spelling tweaks.
* platform.hh: Add get_default_confdir().
* app_state.cc (app_state::app_state): Call get_default_confdir()
to initialize confdir.
* unix/fs.cc (get_default_confdir): Use the home directory as the
base of the default configuration dir.
* win32/fs.cc (get_default_confdir): Use either %APPDATA% or the
result of querying SHGetFolderPath() for CISDL_APPDATA as the base
of the default configuration dir.
(get_homedir): More consistent method for deciding what the user's
home directory is. Behaviour based on the documentation for Qt's
QDir::home().
2005-11-21 Nathaniel Smith <njs@pobox.com>
* NEWS: Initial draft of 0.24 release notes.
2005-11-14 Benoît Dejean <benoit@placenet.org>
* po/fr.po: Updated French translation.
* netsync.cc: Removed many i18n strings.
2005-11-14 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Historical records): Clarify wording, based on
feedback from Daniel Phillips.
2005-11-13 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Automation): Document sort order for 'automate
select'.
2005-11-10 Richard Levitte <richard@levitte.org>
* monotone.texi (Hooks): Small correction so the text matches the
example.
2005-11-10 Timothy Brownawell <tbrownaw@gmail.com>
* lua.cc, std_hooks.lua, monotone.texi: basic_io parser for lua is
now called parse_basic_io
2005-11-10 Timothy Brownawell <tbrownaw@gmail.com>
* monotone.texi: document parse_basicio in the
"Additional Lua Functions" section
2005-11-10 Timothy Brownawell <tbrownaw@gmail.com>
* testsuite.at: Several MINHOOKS_* macros that use the standard hooks
and only load predefined hooks about passphrase and rng.
* tests/t_netsync_permissions, tests/t_netsync_read_permissions: use
the standard permission hooks and define {read,write}-permissions files.
Check both the permissions mechanism and the standard hooks at once.
2005-11-10 Timothy Brownawell <tbrownaw@gmail.com>
* monotone.texi: update example project with latest read-permissions
format. Mention comment lines in the description in the hooks section
2005-11-08 Timothy Brownawell <tbrownaw@gmail.com>
Make a basic_io parser available to Lua.
* lua.cc: new function, monotone_parse_basicio_for_lua
* std_hooks.lua: use it in get_netsync_read_permitted
2005-11-05 Timothy Brownawell <tbrownaw@gmail.com>
Make sure that all new revisions added to the db deltify as much as
possible.
* commands.cc, diff_patch.cc: deltify both sides when doing a merge
* database.{cc,hh}: new function, database::deltify_revision()
call it from put_revision, so all new revisions will be deltified
2005-10-30 Nathaniel Smith <njs@pobox.com>
* std_hooks.lua (edit_comment): Insert a blank line if there is no
user log message, so the user doesn't have to.
2005-10-25 Emile Snyder <emile@alumni.reed.edu>
Fix bug reported on mailing list by Wim Oudshoorn and Tom Koelman
where 'disapprove REV' inappropriately uses your working copy
branch for the disapproved revision.
* app_state.{cc,hh} (set_is_explicit_option, is_explicit_option):
New methods to determine if a given option value was set via
command line flag or not.
* cert.cc (guess_branch): Only accept explicit --branch values in
preference to the branch the given revision lives on.
* commands.cc (CMD(commit)): Explicitly use app.branch_name() if
it exists before calling guess_branch, so pick up MT/options
setting.
* monotone.cc (cpp_main): Set the explicit_option_map entries for
select command line flag processing.
2005-10-26 Matt Johnston <matt@ucc.asn.au>
* INSTALL: mention that zlib is required
* debian/control: monotone doesn't work over plain network protocols
2005-10-25 Timothy Brownawell <tbrownaw@gmail.com>
* lua.cc: make the globish matcher available to lua
* std_hooks.lua: new (better) format for read-permissions
* monotone.texi: update documentation
2005-10-24 Benoît Dejean <benoit@placenet.org>
* commands.cc: Merged 2 i18n strings.
2005-10-23 Timothy Brownawell <tbrownaw@gmail.com>
* std_hooks.lua: new default get_netsync_*_permitted hooks, which
read permissions setting from $confdir/{read,write}-permissions
* monotone.texi: document the new default hooks
2005-10-23 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: new option "-p <pidfile>", catch SIGTERM and SIGINT
and exit cleanly
2005-10-22 Timothy Brownawell <tbrownaw@gmail.com>
* constants.cc: increase log_line_sz so that the new unified
"SERVER IDENTIFICATION HAS CHANGED" message doesn't get truncated.
It used to be multiple print statements but was merged into one,
apparently to help with translation.
2005-10-22 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: Support friendly shutdown (no new connections, but
don't kill existing ones) of local servers and the usher as a whole.
Simple admin interface to start/stop or get the status of individual
servers or the whole usher (listens for connections on a separate
port, only enabled if specified on the command line).
2005-10-22 Richard Levitte <richard@levitte.org>
* tests/t_database_sig_cleanup.at: Two more sleeps that I forgot.
2005-10-20 Nathaniel Smith <njs@pobox.com>
* cert.hh (find_common_ancestor): Remove dead declaration.
* database.cc (check_schema): Improve schema mismatch error
message.
2005-10-20 Richard Levitte <richard@levitte.org>
* tests/t_database_sig_cleanup.at: Sleep for a couple of seconds,
so monotone has a chance to clean up and die properly, even on a
slower machine.
2005-10-19 Nathaniel Smith <njs@pobox.com>
* tests/t_update_switch_branch.at, testsuite.at: New test.
2005-10-19 Nathaniel Smith <njs@pobox.com>
* commands.cc (update): Make branch sticky even if update target
== current rid.
2005-10-19 Matt Johnston <matt@ucc.asn.au>
* main.cc, database.{cc,hh}: SIGINT and SIGTERM handlers
exit gracefully, and try to ROLLBACK+close any databases to clean up
.db-journal files. Added new database::close() method to be used
rather than sqlite_close() directly
* monotone.{cc,hh}, sanity.{cc.hh}: move clean_shutdown flag to
global_sanity
* tests/t_database_sig_cleanup.at: test it
* keys.cc: don't L() private key
2005-10-19 Matthew A. Nicholson <matt@matt-land.com>
* std_hooks.lua: Minor correction to vim warning during 3-way merge.
2005-10-18 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: Update comment about client versions
2005-10-18 Timothy Brownawell <tbrownaw@gmail.com>
* netsync.cc netcmd.{cc,hh}: usher_reply_cmd now has both who we
connected to *and* what pattern we asked for
* contrib/usher.cc: new file format, allow to key servers on either
hostname or pattern, hostname checked first
reload config file on SIGHUP
2005-10-18 Timothy Brownawell <tbrownaw@gmail.com>
* netsync.cc (session::process_usher_cmd): reply with who we connected
to (hostname or hostname:port), not what pattern we asked for.
* contrib/usher.cc: Update comment.
2005-10-18 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: support dynamic local servers
These are started when a client attempts to connect, and killed at
a set time interval after the last client has disconnected.
2005-10-15 Nathaniel Smith <njs@pobox.com>
* manifest.cc (build_restricted_manifest_map): Mention the new
--missing option in the missing files error message hint.
2005-10-17 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (serve): check that the db is valid before beginning
service (would previously not know until someone connected)
* netsync.cc (serve_connections): don't say "beginning service" until
after opening the socket.
2005-10-16 Timothy Brownawell <tbrownaw@gmail.com>
* app_state.{cc,hh}, key_store.cc: change how the keystore directory
is initialized; was using app.confdir before it was ready
2005-10-16 Timothy Brownawell <tbrownaw@gmail.com>
Make the configuration directory configurable and available to lua
* lua.cc: export new function to lua, "get_confdir"
* app_state.{cc,hh} monotone.cc options.hh: new option --confdir,
make the configuration directory something other than ~/.monotone
* lua.{cc,hh}: make the app_state availabe to lua callbacks
* key_store.cc: use get_confdir instead of hardcoded
* lua.cc: use get_confdir instead of hardcoded
* tests/t_config_confdir.at: test --confdir and lua get_confdir
* testsuite.at: add it
2005-10-16 Timothy Brownawell <tbrownaw@gmail.com>
Teach client to optionally push unused keys; new pubkeys can now be
given to a server without restarting it.
* app_state.{cc,hh}, monotone.cc, options.hh:
new command-specific-option --key-to-push=<key> , used to sync/push
a key that hasn't signed anything
* netsync.cc: make it work
* commands.cc: push and sync take it
* tests/t_netsync_pubkey.at: test it
2005-10-16 Matthew Gregan <kinetik@orcon.net.nz>
* configure.ac: Don't add boost_unit_test_framework to global
LIBS, we will link against it directly when needed.
* Makefile.am: Only allow unit_tests to be built if
boost_unit_test_framework was available at configure time.
2005-10-15 Matthew Gregan <kinetik@orcon.net.nz>
* configure.ac: Use boost_filesystem as the library to test for
when trying to guess a suitable BOOST_SUFFIX. Make a missing
boost_unit_test_framework library a non-fatal error.
2005-10-14 Emile Snyder <emile@alumni.reed.edu>
* commands.cc: Fix breakage I introduced with the --unknown flag.
Rename find_unknown to find_unknown_and_ignored, since that's what
it's actually doing. Likewise ls_unknown ->
ls_unknown_or_ignored, and use find_unknown_and_ignored correctly.
Fixes t_mt_ignore.at failures.
2005-10-14 Nathaniel Smith <njs@pobox.com>
* key_store.{cc,hh} (get_key_dir): New method.
* schema_migration.cc (migrate_client_to_external_privkeys): Tell
user that we're moving their keys.
* commands.cc (genkey): Refer to keystore by the directory, not as
"keystore".
(ls_keys): Likewise.
2005-10-14 Timothy Brownawell <tbrownaw@gmail.com>
Add .mt-ignore, ignore most generated files, except for the *m4 files.
2005-10-14 Emile Snyder <emile@alumni.reed.edu>
* app_state.{cc,hh}: new 'unknown' class member for --unknown flag.
* commands.cc (CMD(add), ls_unknown, find_unknown): use new
--unknown flag to add any files in the working copy that monotone
doesn't know about (and isn't ignoring).
* monotone.cc (coptions, cpp_main): add the --unknown flag handling.
* options.hh: add OPT_UNKNOWN.
* monotone.texi: document it.
* tests/t_add.at: test it.
2005-10-14 Emile Snyder <emile@alumni.reed.edu>
* database.cc (complete): enhance h: and b: to mean "current
branch" (as defined by your working copy) when given empty.
* monotone.texi: document it.
2005-10-14 Matthew Gregan <kinetik@orcon.net.nz>
* database.cc: Exclusively lock the database when performing a 'db
dump'.
* schema_migration.cc: Run an analyze at the end of a database
migration to help out the SQLite query optimizer. Also change
schema id calculation to match 'sqlite_stat%' rather than the
specific table 'sqlite_stat1'--the SQLite docs suggest that more
tables will be created in the future.
2005-10-14 Matthew Gregan <kinetik@orcon.net.nz>
* database.cc, schema.sql, schema_migration.cc: Use SQLite 3's
exclusive locking whereever we start a new transaction; allows
monotone to report database locking errors earlier if a second
monotone process attempts to perform database operations.
2005-10-14 Matthew Gregan <kinetik@orcon.net.nz>
* commands.cc (message_width): Refactor into display_width() and
remove message_width().
* transforms.{cc,hh} (display_width): Rename length() to
display_width() and use the code from message_width().
* ui.cc, commands.cc: Change calls to length() to display_width().
* netsync.cc: Remove length() calls, test for string emptiness
instead.
2005-10-13 Emile Snyder <emile@alumni.reed.edu>
* commands.cc (CMD(revert)): use --missing for revert a'la drop.
* monotone.texi: document it.
* tests/t_revert_restrict.at: test it.
2005-10-13 Emile Snyder <emile@alumni.reed.edu>
* app_state.{cc,hh}: new missing class member for --missing flag.
* commands.cc (CMD(drop), ls_missing, find_missing): use new
--missing flag to drop any files already deleted in the working copy.
* monotone.cc (coptions, cpp_main): add the --missing flag handling.
* options.hh: add OPT_MISSING.
* monotone.texi: document it.
* tests/t_drop_missing.at: test it.
2005-10-13 Emile Snyder <emile@alumni.reed.edu>
* database.cc (complete): implementation for h:branch selector to
find heads of a branch.
* selectors.{cc,hh}: add sel_head with selector character 'h'.
* monotone.texi: document it.
2005-10-12 Nathaniel Smith <njs@pobox.com>
* revision.hh: Oops, missed a bit.
2005-10-12 Nathaniel Smith <njs@pobox.com>
* revision.cc (add_node_for_old_revision): Rename to...
(add_node_for_oldstyle_revision): ...this.
(build_changesets_from_existing_revs): Rename to...
(build_roster_style_revs_from_manifest_style_revs): ...this.
* commands.cc (db): Rename "rebuild" to "rosterify".
2005-10-08 Nathaniel Smith <njs@pobox.com>
* revision.cc (analyze_manifest_changes): Remove.
2005-10-08 Nathaniel Smith <njs@pobox.com>
* revision.cc (calculate_change_sets_recursive):
(find_subgraph_for_composite_search)
(calculate_composite_change_set, calculate_arbitrary_change_set):
Remove.
(construct_revision_from_ancestry): Likewise.
2005-10-08 Nathaniel Smith <njs@pobox.com>
* revision.cc (check_sane_history): Remove.
2005-10-13 Matt Johnston <matt@ucc.asn.au>
* botan/mem_pool.cpp: fix bug preventing remove_empty_buffers()
from being called, gives significant improvements for long-running
processes.
2005-10-12 Emile Snyder <emile@alumni.reed.edu>
* monotone.texi: add njs's lucid definition of what the disapprove
actually means in terms of the revision graph to it's section in
the documentation.
2005-10-12 Matt Johnston <matt@ucc.asn.au>
* monotone.cc: return with exit code of 2 on usage output
* testsuite.at: turn "cat_foo" into "automate get_foo"
* tests/t_empty_env.at, tests/t_null.at, tests/t_at_sign.at,
tests/t_unreadable_db.at, tests/t_cmdline_options.at: expect 2 exit code
* tests/t_cross.at: update to the changeset paradigm of immutable
ancestors.
2005-10-11 Emile Snyder <emile@alumni.reed.edu>
* app_state.{cc,hh}: new bind_address and bind_port class members.
* options.hh: new OPT_BIND for serve command.
* monotone.cc: handle OPT_BIND.
* commands.cc (process_netsync_args, CMD(serve)): use values from
the new --bind=[addr]:port option for 'monotone serve' rather than
the positional argument. default to binding all interfaces.
* monotone.texi: document new serve syntax
* netsync.cc (serve_connections): use empty address argument to
mean bind all interfaces.
* netxx/address.cxx (Address::add_all_addresses): set port_ member
from passed in port argument.
* testsuite.at, tests/t_netsync_single.at: use new --bind syntax
for serve commands.
2005-10-11 Matthew Gregan <kinetik@orcon.net.nz>
* commands.cc (message_width, explain_usage): New function to
calculate display width of translated strings. Use it in
explain_usage.
Patch from Thomas Moschny <thomas.moschny@gmx.de>
* certs.cc (get_user_key): Correct logic in N() test for multiple
keys. Resolves bug reported by Malte Thoma.
2005-10-10 Timothy Brownawell <tbrownaw@gmail.com>
Make netsync shutdown work properly.
* netsync.cc (session::which_events): Only ask to read if not armed.
A read failure will discard all read-but-not-processed commands.
* netsync.cc (session::maybe_say_goodbye): Only say goodbye once.
Repeated goodbyes means that one side *will* have a full write-queue
when the connection is torn down.
2005-10-10 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_automate_stdio.at: Add comment here too.
2005-10-10 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/get_stdio.pl: Add comments. I'm told it was "obtuse",
hopefully this will help somewhat.
2005-10-10 Matt Johnston <matt@ucc.asn.au>
* configure.ac, Makefile.am: add {MONOTONE,LIB3RDPARTY}_C{,XX}FLAGS
variables.
2005-10-10 Benoît Dejean <benoit@placenet.org>
* commands.cc: s/key store/keystore/g.
2005-10-10 Matt Johnston <matt@ucc.asn.au>
* netsync.cc, database.{cc,hh}: don't include unwanted branch
certs when building the merkle tree.
* merkle.hh: add variant of insert_into_merkle_tree
taking non-raw id.
* markle.cc: comment out unused bits
2005-10-10 Nathaniel Smith <njs@pobox.com>
* cert.{cc,hh} (get_user_key): New function, replacing
guess_default_key.
* revision.cc (build_changesets_from_existing_revs)
(build_changesets_from_manifest_ancestry):
* rcs_import.cc (import_cvs_repo):
* commands.cc (internalize_cert_name, push, sync, serve):
* cert.cc (make_simple_cert, cert_revision_author_default): Use
it.
2005-10-10 Matthew Gregan <kinetik@orcon.net.nz>
* netsync.cc: Clean up another case where port numbers were being
munged by localized iostreams.
* tests/t_netsync_largish_file.at: Tweak the random file
generation so that it's a bit faster on platforms with slow awk
implementations.
* netsync.cc: Remove comment stating that our official IANA port
should be renamed from 'netsync' to 'monotone'; Tomas sorted this
out a while ago.
* lua.cc (Lua::ok): Only log the 'Lua::ok() failed' message if
there has been a failure.
* monotone.texi: Fix a typo.
* win32/fs.cc (get_homedir): Clarify comment.
2005-10-09 Matthew Gregan <kinetik@orcon.net.nz>
* contrib/usher.cc (main): Reset parser state for config file
parser.
* win32/monotone.iss: Reduce minimum OS version for installer.
* configure.ac: Another Boost library suffix.
2005-10-08 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_automate_keys.at: Portability fixes.
2005-10-08 Benoît Dejean <benoit@placenet.org>
* commands.cc: One more i18n string.
2005-10-08 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_automate_keys.at: Don't rely on extended sed either!
2005-10-08 Nathaniel Smith <njs@pobox.com>
* database.{hh,cc}: Clean-up previous change -- switch to using a
cleanup_ptr to hold sqlite3_stmt handles, for proper exception
safety.
2005-10-08 Nathaniel Smith <njs@pobox.com>
* database.cc (fetch): Do not insert prepared statements into the
statement table until they are successfully created, to avoid
segfaults on database teardown.
2005-10-08 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_automate_keys.at: Use sed rather than relying on an
extended grep being available.
2005-10-08 Nathaniel Smith <njs@pobox.com>
* contrib/ciabot_monotone.py (send_change_for): Handle branch-less
revisions.
2005-10-08 Nathaniel Smith <njs@pobox.com>
* paths.cc: Make initial_rel_path an fs::path. Adjust code
correspondingly. Add tests that calling monotone from inside MT
should work.
* tests/t_add_inside_MT.at, testsuite.at: New test.
2005-10-08 Matthew Gregan <kinetik@orcon.net.nz>
* win32/fs.cc: Don't consider USERPROFILE as a possible home
directory; APPDATA is more appropriate. Add a second method to
get the APPDATA path. Always use HOME if it's set (not just under
Cygwin/MinGW).
* tests/t_netsync_absorbs.at: Add missing CANONICALISE().
2005-10-07 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (pubkey): don't insist on having a database
* tests/t_key_management_without_db.at: check it
2005-10-07 Timothy Brownawell <tbrownaw@gmail.com>
* file_io.cc, vocab.cc, Makefile.am: fix merge
* botan/algolist.cpp, botan/def_alg.cpp: fix merge
2005-10-07 Timothy Brownawell <tbrownaw@gmail.com>
* automate.cc: add "automate keys". like "ls keys", except it
uses basic_io (since "ls keys" is now slightly irregular and not
very well suited to machine interpretation).
* monotone.texi: document it
* tests/t_automate_keys.at: test for it
* testsuite.at: add test
* basic_io.{cc,hh} (stanza): add push_str_multi, for entries
that can take a variable number of string arguments.
* commands.cc: add "automate keys" to automate help
misc minor whitespace cleanup
2005-10-06 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (ls_keys): don't insist on having a db
* tests/t_key_management_without_db.at: check ls keys
* monotone.texi: update for new key storage
* concepts :: storage and workflow
* tutorial :: generating keys
* tutorial :: committing work
* command reference :: informative
* command reference :: key and cert
* command reference :: packet I/O
2005-10-06 Timothy Brownawell <tbrownaw@gmail.com>
* commands.cc (chkeypass): doesn't use the db, so we don't need a guard
* commands.cc (dropkey): don't insist on having a db
* tests/t_genkey_without_db.at: rename (s/genkey/key_management) and
add checks for chkeypass and dropkey without a db
2005-10-06 Timothy Brownawell <tbrownaw@gmail.com>
* database.{cc,hh}: new function, database::database_specified()
return whether we were given a database (doesn't care about validity,
only that we were told to use one)
* commands.cc (genkey): don't insist on having a db, but still
check it for duplicates if we're given one.
* tests/t_genkey_without_db.at: make sure it works
* testsuite.at: add test
2005-09-28 Timothy Brownawell <tbrownaw@gmail.com>
* keys.cc (get_private_key): don't use lua hook passwords for
chkeypass.
2005-09-28 Matt Johnston <matt@ucc.asn.au>
* keys.cc (get_private_keys): don't use lua hook passwords for
chkeypass.
2005-09-28 Timothy Brownawell <tbrownaw@gmail.com>
* key_store.cc: fix delete_key
* tests/t_database_check_normalized: needs 'db migrate'
* tests/t_lua_privkey.at: update for new key location
* tests/t_netsync_absorbs.at: use genkey instead of packets
* tests/t_netsync_checks_server_key.at: same
* tests/t_netsync_pubket.at: update key packet to new format
2005-09-27 Timothy Brownawell <tbrownaw@gmail.com>
Replace extract_keys command with a database migrator.
* database.{cc,hh}: Remove last traces of private keys in db.
* schema.sql: Remove private_keys table
* commands.cc (dropkey): private keys can no longer be in the db.
* schema_migration.cc: new migrator
* commands.cc: remove extract_keys
* tests/t_migrate_schema.at: update
* keys.hh, cert.cc, key_store.cc: keys_match doesn't work well on
privkeys. Don't use it.
* key_store.cc: Allow duplicate insertions, if they match.
2005-09-27 Timothy Brownawell <tbrownaw@gmail.com>
* schema_migration.hh: migrate_depot_schema() doesn't really exist
* database.cc, schema_migration.{cc,hh}: migrate_monotone_schema
now takes an app_state * , so data can be moved to/from the database.
2005-09-27 Matt Johnston <matt@ucc.asn.au>
* botan/pkcs8.cc, keys.cc: fix monotone-specific pkcs8 key
parsing.
* commands.cc: ls keys shows keystore-only pubkeys as well
* packet.cc: fix unit tests
* testsuite.at: use keypair packet rather than privkey/pubkey
packets.
2005-09-26 Matt Johnston <matt@ucc.asn.au>
* commands.cc, keys.cc, others: use standard
"PBE-PKCS5v20(SHA-1,TripleDES/CBC)" encoding rather than
raw arc4. extract_keys will re-encode keys, still
need to work on packet.cc encoding.
* botan/{des*,pkcs5*,algolist.cpp}: required files
from Botan 1.4.6
2005-09-25 Timothy Brownawell <tbrownaw@gmail.com>
Move private keys outside the database.
They're now stored as keypairs in ~/.monotone/keys
Details:
* keystore.{cc,hh}: handle storage of keypairs
* tests*: Specify keystore location when running the testsuite.
This prevents it from polluting the user's keystore.
* database{cc,hh}: remove ability to put privkeys. They can still
be listed, retrieved, and deleted, to support
* commands.cc: New command, extract_keys. Copies private keys from
the database to the keystore.
* vocab.hh: add keypair type, like pair<pubkey, privkey>, except that
the members are named pub and priv.
* packet.cc: Clean up read_packets.
* packet.{cc,hh}: privkey packets no longer exists, handle keypair
packets instead.
* file_io.{cc,hh}: add read_directory because tree_walker needs to
be in the wc. add a write_data that works outside the wc.
* Makefile.am: add keystore.{cc,hh}
* others: update to work with new key storage
2005-10-06 Matthew Gregan <kinetik@orcon.net.nz>
* std_hooks.lua (merge3_emacs_cmd): Minor cleanup.
* std_hooks.lua: Pass '--eval' rather than '-eval' to Emacs; the
CVS emacsclient does not consider '-eval' as valid.
* platform.hh, unix/fs.cc, win32/fs.cc (rename_clobberingly): New
function for best-effort atomic renames. This will probably never
be atomic on Win32.
* file_io.cc (write_data_impl): Append process ID to temporary
file name so that multiple monotone processes operating on the
same working copy are less likely to clobber each other. Use
rename_clobberingly().
* ui.cc (user_interface): Allow std::cout to raise
ios_base::failure exceptions if stream goes bad (e.g. stream
receives EPIPE).
* monotone.cc (cpp_main): Catch ios_base::failure exceptions and
exit cleanly.
* main.cc (main_with_signal_handlers): Install handler to ignore
SIGPIPE.
* Makefile.am: Remove platform_netsync.cc references.
* {unix,win32}/platform_netsync.cc: Remove files.
* platform.hh, netsync.cc (run_netsync_protocol):
Remove {start,end}_platform_netsync() calls.
2005-10-04 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Automation): Document the mysterious columns of
numbers in 'automate inventory' output format.
2005-10-04 Matt Johnston <matt@ucc.asn.au>
* botan/, Makefile.am: update to Botan 1.4.7
2005-10-03 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Automation): Clean up formatting and description
of 'automate inventory' sample output.
2005-10-02 Marcel van der Boom <marcel@hsdev.com>
* commands.cc (CMD(checkout)): Check for non-existant checkout dir
earlier, so we can fail earlier.
2005-09-30 Nathaniel Smith <njs@pobox.com>
* Makefile.am (AM_CFLAGS, AM_CXXFLAGS): Add
-DBOOST_SP_DISABLE_THREADS (disables thread-safe shared_ptr
reference counting).
2005-09-30 Nathaniel Smith <njs@pobox.com>
* debian/README.Debian, debian/TODO.Debian: Add debian files
left out of repository.
2005-09-30 Nathaniel Smith <njs@pobox.com>
* NEWS: Set date for 0.23.
2005-09-30 Nathaniel Smith <njs@pobox.com>
* testsuite.at: Whoops, forgot to hit save. Remove
t_diff_changeset_bug.at.
2005-09-30 Grahame Bowland <grahame@angrygoats.net>
* paths.cc: remove unit tests for
file_path_internal_from_user
2005-09-30 Patrick Mauritz <oxygene@studentenbude.ath.cx>
* configure.ac: only add gnu specific CFLAGS and CXXFLAGS for
gnu compilers
2005-09-30 Nathaniel Smith <njs@pobox.com>
* tests/t_log_outside_working_dir.at: Fix summary line.
* tests/t_diff_changeset_bug.at: Remove redundant test
(cf. t_restricted_diff_unchanged.at).
2005-09-30 Nathaniel Smith <njs@pobox.com>
* Makefile.am (BOTAN_SOURCES): Add botan/hash_id.h.
2005-09-30 Nathaniel Smith <njs@pobox.com>
* NEWS: More additions (mention MM and the problems paths.cc will
cause for colinux folks).
2005-09-30 Nathaniel Smith <njs@pobox.com>
* Makefile.am (MOST_SOURCES): Add hash_map.hh.
(BOTAN_SOURCES): Add botan/charset.h.
2005-09-29 Nathaniel Smith <njs@pobox.com>
* NEWS: Fix typos pointed out by Matthew Gregan.
2005-09-29 Nathaniel Smith <njs@pobox.com>
* UPGRADE, configure.ac, monotone.spec, win32/monotone.iss:
* debian/changelog: Bump version number.
2005-09-30 Grahame Bowland <grahame@angrygoats.net>
* paths.hh: update file_path documentation,
remove file_path_internal_from_user
* paths.cc: remove file_path_internal_from_user
* commands.cc CMD(cat): change file_path_internal_from_user
to file_path_external
* tests/t_diff_outside_working_dir.at: check diff works outside
of a working dir (with a file specified)
* tests/t_log_outside_working_dir.at: check log works outside
of a working dir (with a file specified)
* tests/t_diff_changeset_bug.at: mysterious changeset bug
* testsuite.at: add new tests
2005-09-29 Nathaniel Smith <njs@pobox.com>
* AUTHORS: Add Grahame and Marcel.
* NEWS: Write 0.23 entry.
* monotone.texi (Reserved Files): Mention in .mt-ignore docs that
the patterns are regexes.
2005-09-30 Grahame Bowland <grahame@angrygoats.net>
* paths.cc (file_path::file_path): Allow external paths outside of
a working directory, so long as we don't enter a working directory
later.
(access_tracker) add new very_uninitialized state which
prevents subsequent initialization
(test_access_tracker) check the very_uninitialized state works
* paths.hh: correct mention of monotone cat file to new syntax
2005-09-29 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc: Restore two unit tests on Win32.
2005-09-29 Richard Levitte <richard@levitte.org>
* testsuite.at: Removed a spurious G...
2005-09-29 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc, win32/fs.cc: Fixes to allow unit tests to pass on Win32.
2005-09-29 Marcel van der Boom <marcel@hsdev.com>
* netsync.cc (rebuild_merkle_trees): only get matched branch
certs, not all of them
2005-09-28 Nathaniel Smith <njs@pobox.com>
* tests/t_rename_diff_names.at, testsuite.at: New test.
2005-09-27 Richard Levitte <richard@levitte.org>
* monotone.texi (Tutorial): It seems like texi2pdf gets quite
confused when a @chapter has a @subsection with a @section in
between. The resulting PDF had a Tutorial that was a indexed as a
subsection under Concepts/Branches instead of being indexed as the
chapter it is, and most confusingly, it's last subsection
(Branching and Merging) ended up as a a separate chapter of it's
own...
2005-09-27 Matt Johnston <matt@ucc.asn.au>
* netsync.cc (ancestry_fetcher): new approach to requesting file
and manifest deltas and full data. Tries to be more efficient
for the vcache of recontstructed data, and conceptually simpler
* tests/t_netsync_unrelated.at: passes, remove XFAIL
2005-09-26 Benoît Dejean <benoit@placenet.org>
* database.cc: Merged 3 strings.
2005-09-26 Matt Johnston <matt@ucc.asn.au>
* commands.cc: add 'help' command
* monotone.cc: add -h alias for --help
2005-09-25 Matt Johnston <matt@ucc.asn.au>
* netsync.cc: use lexical_cast on port numbers to avoid
strange-looking "port 5,253"
2005-09-25 Matthew Gregan <kinetik@orcon.net.nz>
* sqlite/*: Update in-tree SQLite from 3.2.6 to 3.2.7.
2005-09-25 Benoît Dejean <benoit@placenet.org>
* netsync.cc: 3 more i18n strings.
2005-09-25 Benoît Dejean <benoit@placenet.org>
* cert.cc (cert_signable_text): F() vs. boost::format.
* lua.cc (monotone_guess_binary_file_contents_for_lua),
(monotone_include_for_lua), (monotone_includedir_for_lua):
* rcs_import.cc (import_cvs_repo): Merged some strings.
2005-09-25 Benoît Dejean <benoit@placenet.org>
* lua.cc (monotone_guess_binary_file_contents_for_lua):
Fixed signednes warning and istream usage.
2005-09-25 Matt Johnston <matt@ucc.asn.au>
* ui.cc: fallback to the default locale if locale("") fails.
2005-09-25 Matt Johnston <matt@ucc.asn.au>
* botan/arc4.cpp: increase maximum keylength to 256 bytes.
2005-09-24 Benoît Dejean <benoit@placenet.org>
* ChangeLog: Fixed.
I don't understand how xxdiff works :/ Please help.
2005-09-24 Satoru SATOH <ss@gnome.gr.jp>
* po/ja.po: Updated Japanese translation.
2005-09-23 Benoît Dejean <benoit@placenet.org>
* std_hooks.lua: More i18n strings.
I don't know how to handle lua
"multiline" ..
"strings" because gettext doesn't join them.
2005-09-22 Benoît Dejean <benoit@placenet.org>
* sanity.{cc,hh}: Uninlined F() and FP().
This happily saves ~100KB of .text on my ppc.
2005-09-22 Benoît Dejean <benoit@placenet.org>
* netsync.cc: Reverted changes
from f0a632bf14468b6e03a488d6f1a64ef18b61d04c
to 4f7f638954c79c54460d04c3be111acad8b26dd3
2005-09-22 Benoît Dejean <benoit@placenet.org>
* lua.cc (monotone_gettext_for_lua): New function.
* po/POTFILES.in: Added std_hooks.lua.
* std_hooks.lua: One more string for i18n.
2005-09-22 Benoît Dejean <benoit@placenet.org>
* ui.{cc,hh} (get_user_locale): Returns a const &.
2005-09-21 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: better error checking
2005-09-20 Timothy Brownawell <tbrownaw@gmail.com>
* restrictions.cc: Make calculate_unrestricted_revision work
with --exclude
* tests/t_restriction_with_exclude_iprint.at: remove XFAIL
2005-09-20 Patrick Mauritz <oxygene@studentenbude.ath.cx>
* configure.ac: solaris needs librt for fdatasync
* sqlite/parse.c: #line and #include interactions aren't defined
2005-09-20 Matt Johnston <matt@ucc.asn.au>
* database.cc (space_usage): workaround the issue that in sqlite
3.2.6, SUM({empty set}) returns NULL. (this is part of the sql spec,
see http://www.sqlite.org/cvstrac/tktview?tn=1413 and the comment for
SUM() in sqlite docs).
2005-09-20 Matt Johnston <matt@ucc.asn.au>
* ui.cc, ui.hh, sanity.hh: replace ui.user_locale with
get_user_locale(), so that we can guarantee that it will be
initialised when the global_sanity object (and probably other things)
are instantiated.
2005-09-20 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_restriction_with_exclude_iprint.at: New test, variant of
t_restriction_exclude.at, but with inodeprints enabled. Mark as
XFAILed.
* testsuite.at: Add new test.
* sqlite/*: Update in-tree SQLite from 3.2.2 to 3.2.6.
* Makefile.am: Add sqlite/{analyze,vbdefifo,complete}.c to
SQLITE_SOURCES
* schema_migration.cc (calculate_schema_id): Explicitly exclude
sqlite_stat1 table from schema ID calculation. This is a new
magic table created by the 'analyze' command in SQLite >= 3.2.3.
2005-09-18 Nathaniel Smith <njs@pobox.com>
* po/ja.po: New translation from Satoru SATOH <ss@gnome.gr.jp>.
* AUTHORS: Add Satoru SATOH.
2005-09-18 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Tweak default 'eveything' netsync glob to
something that works around the MinGW/Win32 globbing issues for
now.
* tests/t_log_brief.at: Canonicalise output on Win32.
2005-09-17 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: Remove stray "new int[2];' statements from
debugging. Remove connections from list when finished.
2005-09-16 Timothy Brownawell <tbrownaw@gmail.com>
* Makefile.am (EXTRA_DIST): add contrib/monotone-cvs-ignore.lua,
contrib/ciabot_monotone_hookversion.py, and contrib/usher.cc .
also remove duplicate of contrib/monoprof.sh
2005-09-16 Timothy Brownawell <tbrownaw@gmail.com>
* netsync.cc: finish renaming things, so it compiles again...
2005-09-16 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/usher.cc: A simple usher/proxy server. It asks connecting
clients for their include pattern, and then forwards the connection
to an appropriate (as given in a config file) monotone server. Note
that all servers operating behind one usher need to have the same
server key.
2005-09-16 Timothy Brownawell <tbrownaw@gmail.com>
* netcmd.{cc,hh}, netsync.cc: new netcmd types: usher_cmd and
usher_reply_cmd. They are not included in the HMAC, and do not
occur during normal communication. Purpose: running multiple servers
from the same port. This allows a special server to ask for the
client's include pattern, and then forward the connection to a real
monotone server.
2005-09-16 Matt Johnston <matt@ucc.asn.au>
* botan/pkcs8.cpp: re-add the monotone-specific code for guessing if
a key is DER encoded or not.
2005-09-16 Matt Johnston <matt@ucc.asn.au>
* botan/*: update to Botan 1.4.6
* Makefile.am: ditto
2005-09-15 Timothy Brownawell <tbrownaw@gmail.com>
* app_state.{cc,hh}: restrictions now understand --exclude
* commands.cc: commit and revert now take OPT_EXCLUDE
* monotone.cc: update description of --exclude
* tests/t_restriction_with_exclude.at: new test
* testsuite.at: add it
2005-09-14 Timothy Brownawell <tbrownaw@gmail.com>
* contrib/ciabot_monotone_hookversion.py: CIA bot client script
meant to be called from the note_netsync_revision_received hook.
* lua.{cc,hh}: hook_note_commit and hook_note_netsync_revision_received
now take the text of the revision as an argument.
* netsync.cc, commands.cc: Give hooks new argument.
* monotone.texi: Update documentation for those hooks.
2005-09-11 Benoît Dejean <benoit@placenet.org>
* database.cc
* database_check.cc
* netsync.cc
* po/POTFILES.in
* po/POTFILES.skip
* rcs_import.cc: More i18n strings.
2005-09-07 Jim Meyering <jim@meyering.net>
* tests/t_rename_dir_cross_level.at: When invoking mv to rename
a directory, do not include a trailing slash on the target -- that
is not portable.
* Makefile.am (TESTS_ENVIRONMENT): Ensure that PATH starts with
the current directory, so we test the just-built monotone binary,
not some older version.
2005-09-07 Benoît Dejean <benoit@placenet.org>
* std_hooks.lua: Don't spawn emacs with '-no-init-file'.
2005-09-09 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc (save_initial_path): Default Boost's fs::path to
fs::native grammar.
* tests/t_unreadable_{db,MT}.at: Disable on Win32 for now.
* paths.cc: Consistency--use WIN32 rather than _WIN32.
* file_io.cc (walk_tree): Correct test for file existence.
2005-09-08 Matthew Gregan <kinetik@orcon.net.nz>
* Makefile.am: Add '-f' argument to 'mv'; avoids build waiting on
user confirmation in some cases.
2005-09-08 Nathaniel Smith <njs@pobox.com>
* monotone.texi (Certificates): Remove mention of fcerts and
mcerts.
2005-09-07 Benoît Dejean <benoit@placenet.org>
* ui.{cc,hh}: Added user_inferface::user_locale.
* sanity.hh: Made F() and FP() locale aware.
2005-09-06 Benoît Dejean <benoit@placenet.org>
* monotone.cc: One more i18n string.
2005-09-06 Benoît Dejean <benoit@placenet.org>
* po/fr.po: Updated French translation.
2005-09-06 Benoît Dejean <benoit@placenet.org>
* commands.cc: No i18n for cert_revision_changelog.
2005-09-06 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_netsync_read_permissions.at: Tweak tests so we aren't
trying to serve a bare '*'; works around test hangs due to glob
expansion sillyness on MinGW.
* tests/t_netsync_globs.at: Ditto.
* tests/t_netsync_exclude.at: Ditto.
* std_hooks.lua (ignore_file): Add Mac OS X (.DS_Store) and
Windows (desktop.ini) per-directory browser configuration files.
2005-09-05 Benoît Dejean <benoit@placenet.org>
* commands.cc: Fixed some strings (added ' around revisions).
Removed some whitespaces.
No i18n in diff output.
2005-09-05 Benoît Dejean <benoit@placenet.org>
* sanity.{cc,hh}: boost::format vs. F(). Merged boost::format
and moved non-template code to sanity.cc.
2005-09-05 Matthew Gregan <kinetik@orcon.net.nz>
* win32/terminal.cc (have_smart_terminal): We were returning false
in almost all circumstances; changed logic so that we at least
work when running in a cmd.exe window.
2005-09-05 Matt Johnston <matt@ucc.asn.au>
* commands.cc (dump_diffs): don't use the terminal width to
print ===== seperators.
2005-09-05 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc (find_and_go_to_working_copy): Create root and bookdir
paths as fs::native.
* main.cc: Tweak #ifdef to avoid exposing some unused SEH handling
on MinGW.
* configure.ac: Minor cleanup to Win32 configure test.
2005-09-04 Nathaniel Smith <njs@pobox.com>
* monotone.cc (options): Remove default from the help string for
--count, since none of the options listed are actually the
default.
2005-09-04 Nathaniel Smith <njs@pobox.com>
* tests/t_unreadable_db.at, testsuite.at: New test.
2005-09-03 Nathaniel Smith <njs@pobox.com>
* po/Makevars (XGETTEXT_OPTIONS): N_ != ngettext.
Add c-format flags on F() and FP() calls (only partially
successful on latter, because of bug in xgettext).
2005-09-04 Grahame Bowland <grahame@angrygoats.net>
* commands.cc: siplify the monotone cat command
to "monotone cat [-r] FIELNAME" (as in bug #12597)
* monotone.texi: update documentation of "monotone cat"
* tests/t_add_edge.at, tests/t_cat_file_by_name.at,
tests/t_change_empty_file.at, tests/t_cvsimport.at,
tests/t_cvsimport_deleted_invar.at, tests/t_cvsimport_drepper.at,
tests/t_cvsimport_drepper2.at, tests/t_cvsimport_manifest_cycle.at,
tests/t_cvsimport_samelog.at, tests/t_database_check.at,
tests/t_db_kill_rev_locally.at, tests/t_empty_id_completion.at,
tests/t_epoch.at, tests/t_epoch_server.at, tests/t_erename.at,
tests/t_i18n_file.at, tests/t_import.at, tests/t_merge_add_del.at,
tests/t_movedel.at, tests/t_movepatch.at, tests/t_netsync.at,
tests/t_netsync_exclude.at, tests/t_netsync_exclude_default.at,
tests/t_netsync_globs.at, tests/t_netsync_nocerts.at,
tests/t_netsync_permissions.at, tests/t_netsync_read_permissions.at,
tests/t_netsync_single.at, tests/t_normalized_filenames.at,
tests/t_persistent_server_revision.at, tests/t_remerge.at,
tests/t_rename.at, tests/t_renamed.at, tests/t_scan.at,
tests/t_set_default.at, tests/t_singlecvs.at,
tests/t_update_with_pending_add.at, tests/t_update_with_pending_drop.at,
tests/t_update_with_pending_rename.at, tests/t_versions.at:
use automation interface rather than "monotone cat"
2005-09-04 Grahame Bowland <grahame@angrygoats.net>
* ChangeLog: fix up screwed up three-way merge
2005-09-04 Grahame Bowland <grahame@angrygoats.net>
* automate.cc, commands.cc: add "automate get_file",
"automate get_revision" and "automate get_manifest" to
automation interface.
* monotone.texi: document new automation commands
* tests/t_automate_get_file.at, tests/t_automate_get_revision_at,
tests/t_automate_get_manifest.at: trivial testing of new
automation commands for output as specified, make sure they
do not complete IDs, make sure invalid IDs are caught.
* testsuite.at: add new tests
2005-09-03 Matthew Gregan <kinetik@orcon.net.nz>
* tests/t_persistent_server_keys_2.at: 'commit' needs a commit
message. Un-XFAIL.
* tests/t_netsync_unrelated.at: Fix 'setup' syntax.
* tests/t_add_vs_commit.at: BASE_REVISION needs to be in the root
of a working copy to work. Un-XFAIL.
* tests/t_add_stomp_file.at: 'add' does not take a --branch
argument. BASE_REVISION needs to be in the root of a working copy
to work.
* annotate.cc (build_parent_lineage): Don't access uninitialized
memory--use resize() rather than reserve().
2005-09-02 Matthew Gregan <kinetik@orcon.net.nz>
* monotone.cc: Use consistent case in option descriptions.
2005-09-02 Nathaniel Smith <njs@pobox.com>
* paths.{hh,cc}: Add split_path typedef. Use it.
2005-09-02 Matt Johnston <matt@ucc.asn.au>
* lua.cc (monotone_guess_binary_file_contents_for_lua): use a
temporary char* buffer rather than &string[], extra copying seems
to have negligible performance impact.
* tests/perf-test.sh: change path from tests/ to contrib/, make
executable.
* tests/parse-accounting.pl: make executable.
2005-09-01 Timothy Brownawell <tbrownaw@gmail.com>
* lua.cc, std_hooks.lua: use proper regexes for .mt-ignore
taken from a patch from Martin Dvorak
* contrib/monotone-cvs-ignore.lua: New file, from the same patch.
supports .cvsignore files
* tests/t_mt_ignore.at: check that a missing .mt-ignore
doesn't cause problems
2005-09-01 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_mt_ignore.at: use RAW_MONOTONE instead of ugly --rcfile
Also actually do "mtn add" this time.
2005-09-01 Timothy Brownawell <tbrownaw@gmail.com>
* tests/t_mt_ignore.at: new test, checks that .mt-ignore works
* testsuite.at: add it
2005-09-01 Timothy Brownawell <tbrownaw@gmail.com>
* std_hooks.lua: support .mt-ignore
* monotone.texi: mention .mt-ignore and MT/wanted-testresults under
"Existing control files"
* .mt-ignore: ignore testsuite.dir
2005-09-03 Benoît Dejean <benoit@placenet.org>
* commands.cc (ls_certs):
* netsync.cc (load_data): Merged strings.
2005-09-01 Benoît Dejean <benoit@placenet.org>
* commands.cc: Merged one more "no such revision '%s'" string.
2005-09-01 Benoît Dejean <benoit@placenet.org>
* commands.cc: Merged all "no such revision '%s'" strings.
(string_to_datetime): Merged catch blocks in order to merge error
messages.
2005-09-01 Benoît Dejean <benoit@placenet.org>
* ChangeLog: Fixed.
2005-09-01 Matthew Gregan <kinetik@orcon.net.nz>
* ui.cc (user_interface): Avoid calling sync_with_stdio(false) on
Win32 for now to work around a bug in MinGW where unsynchronized
std::cin returns EOF earlier when reading a stream with DOS
newlines. Resolves 'db load' failure reported by Howard Spindel.
* database.cc (load): Don't bother executing an empty string.
* commands.cc (ALIAS(import, setup)): Remove alias.
2005-09-01 Matt Johnston <matt@ucc.asn.au>
* schema.sql: add BEGIN, COMMIT to make it a single transaction,
improves db init performance significantly on OS X (avoids many
disk-cache flushes).
2005-09-01 Matthew Gregan <kinetik@orcon.net.nz>
* testsuite.at: Increase entropy used to generate port numbers
where we can and increase range of port numbers generated.
* monotone.texi: Fix a couple of minor typos.
2005-09-01 Matthew Gregan <kinetik@orcon.net.nz>
* monotone.texi: Update 'setup' documentation and tutorial to
reflect new usage. Also update much of the monotone output in the
tutorials to reflect the output of more modern versions of
monotone. Correct some minor errors and typos while here.
* commands.cc (CMD(setup)): Require database and branch arguments.
(ALIAS(import,setup)): Add setup alias.
* testsuite.at, tests/*.at: Update 'setup' usage.
2005-08-31 Richard Levitte <richard@levitte.org>
* lua.cc, std_hooks.lua: Rename
monotone_guess_binary_filename_for_lua and guess_binary_filename
to monotone_guess_binary_file_contents_for_lua and
guess_binary_file_contents.
2005-08-31 Benoît Dejean <benoit@placenet.org>
* basic_io.cc (basic_io::input_source::err): Merged strings.
2005-08-31 Nathaniel Smith <njs@pobox.com>
* file_io.cc (set_char_is_binary, guess_binary): static_cast chars
to uint8_t before using as array indices. Also replace some ints
with size_t's to quiet g++ warnings.
2005-08-30 Benoît Dejean <benoit@placenet.org>
In function void set_char_is_binary(char, bool)
133: warning: array subscript has type char
In function void init_char_is_binary()
147: warning: comparison between signed and unsigned integer expressions
In function bool guess_binary(const std::string&)
160: warning: comparison between signed and unsigned integer expressions
162: warning: array subscript has type char
2005-08-30 Benoît Dejean <benoit@placenet.org>
* file_io.cc (walk_tree): Fixed format.
2005-08-31 Marcel van der Boom <marcel@hsdev.com>
* std_hooks.lua (execute_confirm): New function.
(merge2_opendiff_cmd, merge3_opendiff_cmd): Add.
2005-08-31 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc (test_bookkeeping_path, test_system_path): Second
attempt at compile fixes; this time the unit tests actually pass
too.
2005-08-30 Matthew Gregan <kinetik@orcon.net.nz>
* paths.cc (test_bookkeeping_path, test_system_path): Shift object
instantiation around a little to work around what seems to be a
bug in the gcc 3.3 parser.
* win32/inodeprint.cc (inodeprint_file): Update to use new path
handling code.
* win32/fs.cc (tilde_expand): Compile fix.
2005-08-30 Petr Baudis <pasky@suse.cz>
* std_hooks.lua: Simple support for merging using merge(1) and vim.
2005-08-30 Benoît Dejean <benoit@placenet.org>
* po/fr.po: Updated French translation.
2005-08-30 Benoît Dejean <benoit@placenet.org>
* commands.cc: Merged some error messages.
2005-08-30 Benoît Dejean <benoit@placenet.org>
* commands.cc: Merged complete(..., file_id) and
complete(..., manifest_id) into template complete(..., ID).
2005-08-30 Benoît Dejean <benoit@placenet.org>
* commands.cc (ls_certs): Reworked for i18n.
(CMD(commit)): Merged 2 strings.
2005-08-30 Matthew Gregan <kinetik@orcon.net.nz>
* revision.cc (ensure_parents_loaded): Don't reuse an iterator
after we've invalidated it. Fixes 'diff' crash reported by Howard
Spindel.
2005-08-30 Matt Johnston <matt@ucc.asn.au>
* botan/allocate.cpp: avoid string comparison when looking up the
default allocator
* monotone.cc (cpp_main): set a default allocator
2005-08-28 Nathaniel Smith <njs@pobox.com>
* tests/t_attributes.at: Delete checkout dir in between
checkouts.
2005-08-28 Matt Johnston <matt@ucc.asn.au>
* keys.cc (keys_match): new function to compare whether two keys
match (ignoring whitespace as the database does, etc).
* packet.cc, keys.cc: use it for existing-key-comparison.
2005-08-27 Nathaniel Smith <njs@pobox.com>
* commands.cc (checkout): Special-case "checkout ."
* tests/t_checkout_dir.at: Test it.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* file_io.hh: Remove comment describing old path types.
* paths.hh: Add comment describing new path types.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* app_state.cc (create_working_copy): Remove
fs::filesystem_exception catching.
* file_io.hh (mkdir_p): Remove comment noting app_state.cc's
dependence on a boost-based implementation.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.cc: Include <string>. Helps build on g++ 3.3?
2005-08-26 Nathaniel Smith <njs@pobox.com>
* commands.cc (get_log_message): Make the log message commentary a
little more descriptive for people who may not know what a log
message is...
(commit): When canceling a commit due to empty log message, say
so.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* std_hooks.lua: Check for both "vi" and "notepad.exe" as fallback
editors. If no editor was found, print a helpful message instead
of just running "vi" anyway. Print a message if the editor exited
with error.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* file_io.cc (mkdir_p, make_dir_for): Increase error checking.
* commands.cc (checkout): Make sure that checkout target directory
does not already exist. Also use system_path more uniformly.
* tests/t_checkout_dir.at: Test.
* tests/t_setup_existing_path.at: New test.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* commands.cc (read): Optionally take files on command line.
* tests/t_read_from_file.at, testsuite.at: New test.
* monotone.texi (Network Service): Show Jim using this.
(Packet I/O, Commands): Document.
* monotone.1: Likewise.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* change_set.cc (move_files_from_tmp_top_down): Typo again.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* database.cc (open): Convert stray line to paths.cc.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* change_set.cc (move_files_from_tmp_top_down): Typo.
* file_io.cc (move_path): New function.
(move_file, move_dir): Minor cleanup -- use
require_path_is_nonexistent.
* work.cc (build_deletions): Use delete_file, rather than unlink.
If file is already non-existent, do nothing.
(build_rename): Use move_path, rather than rename. If file
already appears to have been renamed, do nothing.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* app_state.cc (allow_working_copy): Make logging more sensible.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* transforms.cc (length):
* database.cc (sql, load, open):
* commands.cc (rename, attr):
* change_set.cc (move_files_to_tmp_bottom_up)
(move_files_from_tmp_top_down): Merge fixups.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* database_check.cc: Track and report on manifest and revision
parseability.
* tests/t_database_check_normalized.at: Update to expect "not
parseable" messages rather than "not normalized" messages.
All tests pass.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* tests/t_unreadable_MT.at: This test was called "do not fail on
unreadable MT/options". I do not know why we wanted such
behavior. I am making it "fail cleanly on unreadable
MT/options".
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.cc (check_fp_normalizes_to, test_file_path_internal):
Oops, there were more places testing for non-brokenness; break
them too.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_split_join): Test that you cannot create a path
in MT by joining.
(file_path): Implement it.
(split): Break, to match broken behavior of old splitter (easier
than fixing change_set.cc...)
(file_path_internal): Test for brokenness accordingly.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* commands.cc (cat): Hack so that 'cat file REV PATH' works
correctly both inside and outside of working copy, interpreting
path slightly differently in each case.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* tests/t_normalized_filenames.at: Internal unnormalized pathnames
are no longer silently normalized, but rather a hard error.
Adjust test accordingly.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.hh (file_path_internal_from_user): New constructor.
* paths.cc (test_file_path_internal): Test it.
(file_path::file_path): Implement it.
* commands.cc (cat): Use it to create/validate passed in
filenames.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_system_path): Require that system_path normalize
out ..'s.
(system_path): Do so.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* work.cc (build_additions): Remove redundant (and now wrong)
code.
Test 53 now passes.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* file_io.cc (make_dir_for): Oops, this doesn't need a
fs::native.
Test 37 now passes.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* file_io.cc (mkdir): New function. Now with extra brain-eating
power.
(mkdir_p, make_dir_for, delete_file, delete_dir_recursive)
(move_file, move_dir, write_data_impl): Use it, to make all
fs::path's native and disable boost's random rejection of paths.
2005-08-26 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_file_path_external_prefix_a_b)
(test_file_path_external_no_prefix, test_file_path_internal): Test
for validity of more strange characters (,+@*%#$=).
2005-08-26 Nathaniel Smith <njs@pobox.com>
* file_io.cc (ident_existing_file): Remove accidentally-left-in
code. Test 26 now passes.
* lua.cc (monotone_includedir_for_lua, load_rcfile): Add
fs::native's.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.hh (system_path::system_path): Add new boolean argument
controlling some access_tracker behavior.
* app_state.cc (allow_working_copy): Use it.
* paths.cc (system_path): Implement it.
(test_system_path): Test it.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (walk_tree): Return properly.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_file_path_internal): Add tests for
file_path.empty().
* file_io.cc (walk_tree_recursive): Add explicit fs::native.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc: Many small changes. Unit tests now pass. 72
unexpected autotest failures.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (file_path): Fix up error reporting in external path
normalization.
(test_file_path_external_no_prefix): "" is always an invalid
path.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* database.cc (sql): Only check schema version when db actually
exists.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (read_data_for_command_line): We are given a
system_path.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc: Fix all unit test failures, except for two mysterious
boost::too_few_args exceptions.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc, unix/fs.cc: Many, many fixes and some new tests too.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (struct access_tracker): Doh, initializer should set
'initialized'...
(test_file_path_internal): It's valid for a split file
to have a null component if the file is "".
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (in_bookkeeping_dir): Last change didn't work out so
well; let's remove some negatives and see if I can understand what
the code does this way...
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (not_in_bookkeeping_dir): Handle "MT" case.
Update tests to make sure it sticks...
2005-08-25 Nathaniel Smith <njs@pobox.com>
* unit_tests.{cc,hh} (init_unit_test_suite): Remove
path_component_tests.
* unix/fs.cc (tilde_expand): Another compile fix.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* {unix,win32}/fs.cc: Misc. compile fixes.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* Makefile.am (UNIX_PLATFORM_SOURCES): Add unix/fs.cc
(WIN32_PLATFORM_SOURCES): Add win32/fs.cc
* paths.hh (bookkeeping_path): Implement default constructor.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* rcs_import.cc (import_cvs_repo):
* lua.cc (default_rcfilename):
* diff_patch.cc (get_version): Small compile fixes.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh} (is_bookkeeping_path): New static method.
* file_io.cc (walk_tree_recursive): Use it. Now compiles.
paths.cc and file_io.cc now compile.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (delete_dir_recursive): Implement.
Misc. compile fixes.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (test_book_keeping_file): Remove.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (walk_tree_recursive, walk_tree): Implement.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (const_system_path): Do tilde expansion.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (read_localized_data, read_data_for_command_line)
(write_localized_data, write_data, write_data_impl): Implement.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (read_data): Implement. Remove the base64<gzip<>>
versions.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (move_file, move_dir): Implement.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (assert_path_is_nonexistent, assert_path_is_file)
(assert_path_is_directory, require_path_is_nonexistent)
(require_path_is_file, require_path_is_directory)
(ident_existing_file, mkdir_p, make_dir_for, delete_file)
(delete_dir_recursive): Implement.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* Audit uses of 'file_exists', because its semantics have changed;
it now checks to see if a path exists and is a regular file,
rather than that it simply exists. A fair amount of code already
thought it meant that... other places now use 'path_exists'.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (path_exists, directory_exists, file_exists):
Implement.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* platform.hh (get_path_status): New function.
* unix/fs.cc (get_path_status): Implement.
* win32/fs.cc (get_path_status): Implement inefficiently (does
win32 have stat?)
2005-08-25 Nathaniel Smith <njs@pobox.com>
* file_io.cc (get_homedir, tilde_expand, book_keeping_file)
(book_keeping_dir): Remove.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* platform.hh (get_homedir): New function.
* {win32,unix}/fs.cc (get_homedir): Expose.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* Minor compile fixes.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* platform.hh (tilde_expand): New function.
* win32/fs.cc, unix/fs.cc: Implement it.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc: Many more fixes. Now compiles with and without unit
tests.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc: Lots of compile fixes for unit tests.
Add a test for access_tracker.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc: Many fixes. Now compiles.
2005-08-25 Nathaniel Smith <njs@pobox.com>
* paths.cc (system_path): Implement.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (fully_normalized_path): Use find_first_of.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (find_and_go_to_working_copy, save_initial_path)
(go_to_working_copy): Use new checked structure.
(operator <<): Make sure we can log our access_tracked values
without marking them as used.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (struct access_tracker): Add invariant checking on
lifetime usage of path roots.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (any_path::operator =): return *this.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh}: More fixes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh}: Reorganize a bit. Implement file_path and
bookkeeping_path.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (file_path): Implement basic constructor.
Misc compile fixes.
Add single-character names to tests.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (go_to_working_copy): New function. Implement.
* app_state.cc (create_working_copy): Adjust accordingly.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (find_and_go_to_working_copy): Implement.
* app_state.cc (allow_working_copy): Adjust accordingly.
(relative_directory): Remove.
* file_io.cc (find_working_copy): Remove.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (save_initial_path): Update for previous changes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_system_path): Add tests for the
from-any_path constructor. Add test for "~foo" handling.
Start cleaning up path roots.
* platform.hh: Note that get_current_working_dir() is
charset-broken (i.e., operations started inside non-utf8
directories are probably broken).
2005-08-24 Nathaniel Smith <njs@pobox.com>
* app_state.cc (allow_working_copy):
* change_set.cc (print_insane_change_set): Two more small compile
fixes.
All remaining compile errors are localized to unimplemented
paths.cc/file_io.cc functionality.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* database.cc (initialize): Missing ;.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* monotone.cc (cpp_main): Handle message_file right.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* change_set.cc (print_insane_path_rearrangement):
* database.cc (initialize):
* monotone.cc (cpp_main): More small compile fixes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* file_io.hh
({assert,require}_path_is_{nonexistent,file,directory}): New
functions.
Use them everywhere.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* basic_io.hh: #include "paths.hh".
* monotone.cc (add_rcfile): Remove obsolete absolutification, etc.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (system_path): Add a from-any_path constructor.
* Makefile.am (MOST_SOURCES): Remove path_component.{cc,hh}.
* basic_io.hh (push_file_pair): New method.
* change_set.cc (print_insane_change_set)
(print_insane_path_rearrangement): Use it.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (any_path::as_internal): On second thought, return a
std::string, not a utf8 -- utf8 would be better, but should wait
for some more general charset handling cleanup.
* Adjust other files accordingly.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* More compile fixes. All remaing compile errors are real
problems, yay.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* Lots and lots more compile fixes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh, monotone.cc, app_state.hh, app_state.cc:
* unix/inodeprint.cc: More compile fixes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* manifest.hh: Include paths.hh.
* file_io.hh: Fix syntax errors, and fixup interface.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh, sanity.hh: Compilation fixes.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.cc: Update tests to use path_state_* and pass utf8
objects.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (file_path_external): Take a utf8() object, always.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (class file_path): Make "convenience functions"
required.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* Switch rest of instances to using convenience functions.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* Switch many instances to using convenience functions.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* paths.hh (file_path_internal, file_path_external): Define
convenience functions.
(file_path, bookkeeping_path, system_path): Add default
constructors.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* app_state.cc, change_set.cc, change_set.hh, commands.cc:
* inodeprint.cc, manifest.cc, rcs_import.cc, restrictions.cc:
* work.cc: Audit all calls to file_path() to add internal/external
notation.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* app_state.cc: More paths.hh conversion.
(app_state::prefix): Remove.
* commands.cc: Remove uses of app.prefix.
* automate.cc (automate_attributes): Likewise.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.hh (any_path::as_internal): On second thought, return a
utf8 object.
* app_state.cc (set_database): Take a system_path.
(set_pidfile): Likewise.
* monotone.cc (cpp_main): Pass one.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* file_io.hh (get_homedir): Return a system_path.
* app_state.hh (app_state): Make pidfile a system_path.
* sanity.hh (sanity::filename): Make a system_path.
* monotone.cc (cpp_main): Adjust accordingly.
* paths.hh (any_path): Add as_internal() to interface.
* paths.cc: Add roundtripping tests.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* platform.hh, unix/fs.cc, win32/fs.cc
(change_current_working_dir): Take an any_path, not a string.
* rcs_import.{cc,hh}: Convert to paths.hh.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* commands.cc (pid_file): Remove fs::path.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* mkstemp.cc (monotone_mkstemp): Remove references to fs::path.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* change_set.cc (apply_rearrangement_to_filesystem): Oops, missed
some local_path's.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* path_component.{cc,hh}: Delete.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* change_set.cc (move_files_to_tmp_bottom_up)
(move_files_from_tmp_top_down): Convert to paths.hh.
Whole file: stop using path_component.hh.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.cc (test_bookkeeping_path): Oops, "" is an invalid
bookkeeping_path.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* lua.cc, paths.cc: Few more tweaks for previous change.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh}: Add / operators. Make that the usual way to use
bookkeeping_path's.
* work.cc: Adjust accordingly.
* lua.cc (working_copy_rcfilename): Likewise.
* commands.cc (update): Likewise.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh} (operator <<): Implement for any_paths.
* paths.hh (class bookkeeping_path): Note that current design is
bogus to remind myself to fix it tomorrow...
2005-08-23 Nathaniel Smith <njs@pobox.com>
* work.{hh,cc}: Convert to paths.hh.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* lua.{cc,hh}: Mostly convert to paths.hh. (Still uses boost::fs
internally for some directory iteration.)
* app_state.cc (load_rcfiles): Update accordingly.
* file_io.hh (path_state): De-templatify; take any_path instead of
a T.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.cc (any_path): New base class.
(file_path, bookkeeping_path, system_path): Inherit from it.
* transforms.{hh,cc} (utf8_to_system): Actually, always take a
utf8 after all (but still have two return types).
2005-08-23 Nathaniel Smith <njs@pobox.com>
* transforms.cc: Convert to paths.hh.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* transforms.{cc,hh} (localized, localized_as_string): Remove.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* transforms.cc (utf8_to_system): Make fast.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* transforms.hh (utf8_to_system): Add a string->string version.
* transforms.cc (utf8_to_system): Implement it.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.cc (localized_path_str): New function.
Fix some tests.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* commands.cc: Convert to paths.hh.
* mkstemp.cc (monotone_mkstemp): Likewise.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* vocab_terms.hh, vocab.cc: Remove file_path, local_path.
* database.{hh,cc}, monotone.cc: Convert to paths.hh.
* file_io.{hh,cc}: Start to convert to paths.hh.
2005-08-23 Nathaniel Smith <njs@pobox.com>
* paths.{cc,hh} (fully_normalized_path): Implement.
(external_path): Rename to system_path.
Misc. other updates.
2005-08-21 Eric Anderson <anderse-monotone@cello.hpl.hp.com>
* file_io.cc, file_io.hh, lua.cc, std_hooks.lua: determine if a
file is binary by looking at it incrementally, rather than reading
it in entirely. Prepare for making it possible to control what
characters are considered "binary"
2005-08-20 Nathaniel Smith <njs@codesourcery.com>
* paths.cc (is_absolute): New function.
(file_path::file_path(vector<path_component>))
(file_path::split): Implement.
2005-08-20 Nathaniel Smith <njs@pobox.com>
* interner.hh (interner): Add a scary constructor that lets us
insert an initial value and assert that we already knew that the
value assigned to it would be. (This lets us make it an inlined
constant.)
2005-08-20 Nathaniel Smith <njs@pobox.com>
* paths.cc: Yet more tests.
2005-08-20 Nathaniel Smith <njs@pobox.com>
* paths.cc (save_initial_path): Implement.
Add more tests.
2005-08-20 Nathaniel Smith <njs@codesourcery.com>
* paths.{cc,hh}: New files.
* Makefile.am (MOST_SOURCES): Add them.
* unit_tests.hh (add_paths_tests): Declare.
* unit_tests.cc (init_unit_test_suite): Add them.
* platform.hh (get_current_working_dir)
(change_current_working_dir): New functions.
* {unix,win32}/fs.cc: New files.
2005-08-19 Nathaniel Smith <njs@codesourcery.com>
* monotone.texi (Tutorial): Tweak wording, use --db at more
appropriate places.
2005-08-26 Richard Levitte <richard@levitte.org>
* database.cc (version): Revert the change done earlier, as it
aborted if the schema isn't the current one, rendering this method
useless.
2005-08-26 Matt Johnston <matt@ucc.asn.au>
* change_set.cc (check_depth, confirm_proper_tree):
more efficient algorithm to check for no loops
* constants.hh: new constant max_path_depth to limit
recursion in check_depth.
2005-08-26 Benoît Dejean <benoit@placenet.org>
* po/fr.po: Updated French translation.
2005-08-26 Richard Levitte <richard@levitte.org>
* options.hh, monotone.cc: Add the '--execute' command-specific
option.
* monotone.cc (cpp_main): ... and process it.
* app_state.hh (class app_state): Add the 'execute' boolean.
* app_state.cc (app_state): Initialise it.
* commands.cc (CMD(drop)): Add '--execute' capability.
* commands.cc (CMD(rename)): Add '--execute' capability. Pass
'app' to build_rename.
* work.hh, work.cc (build_deletions, build_rename): Do the actual
work. This required the addition of an app_state parameter to
build_rename.
* tests/t_drop_execute.at, tests/t_rename_execute.at: New tests.
* testsuite.at: Add them.
2005-08-26 Benoît Dejean <benoit@placenet.org>
* mt_version.cc (print_full_version): Merged strings.
* change_set.cc: No i18n for unittests. Wow, this saves
21 strings (total 781).
2005-08-25 Benoît Dejean <benoit@placenet.org>
* commands.cc (safe_gettext): New function.
(explain_usage): Used there to avoid _("").
2005-08-25 Benoît Dejean <benoit@placenet.org>
* sanity.{cc,hh} (sanity::do_format): Merged code from
sanity::{log, warning, progress} in order to merge
strings. Fixed exception rethrowing.
2005-08-25 Benoît Dejean <benoit@placenet.org>
* commands.cc (CMD(lca)): One more string for i18n.
(CMD(trusted)): Merged all strings.
2005-08-25 Benoît Dejean <benoit@placenet.org>
* po/fr.po: Updated French translation.
2005-08-25 Benoît Dejean <benoit@placenet.org>
* database.cc (database::version): Marked string for i18n
and simplifed.
(database::info): Reworked to merge all strings for i18n.
2005-08-25 Benoît Dejean <benoit@placenet.org>
* database.{cc,hh} (database::open, database::check_filename):
New functions to avoid error handling code and string duplicates.
2005-08-25 Matt Johnston <matt@ucc.asn.au>
* transform.cc ({int,ext}ernalize_rsa_keypair_id): don't
convert the username portion of key ids to/from ACE.
* tests/t_genkey.at: check that foo+bar@example.com works
and foobar@exam+ple.com doesn't.
2005-08-24 Benoît Dejean <benoit@placenet.org>
* database.cc (assert_sqlite3_ok): Somehow merged error messages.
2005-08-24 Benoît Dejean <benoit@placenet.org>
* change_set.cc (move_files_to_tmp_bottom_up): Better strings.
* keys.cc (generate_key_pair): Merged 2 strings.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* database.cc (assert_sqlite3_ok): Remove accidentally-left-in
format string argument.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* revision.cc (check_sane_history): Add MM's for calculated
changesets.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* database.cc (assert_sqlite3_ok): Don't print the raw sqlite
error code. Do add some auxiliary information when sqlite errors
are confusing.
2005-08-24 Nathaniel Smith <njs@pobox.com>
* Back out most changes since
b580c6ac5bf8eea1f442b8bddc60283b047ade1e. Handling charsets
properly by working in utf8 and then converting sucks. Problems
include
- gettext hates you (wants to return stuff in local charset)
- strerror hates you (same reason, but you can't turn it off)
- can't report charset conversion errors
We thus return to our "sorta-correct, by accident" status quo.
Only change left in is signedness fix in
5548868ab56d939c1fd8713aa2ac8caacd1184a1.
2005-08-23 Nathaniel Smith <njs@pobox.com>