[libsql-abstract-perl] annotated tag upstream/1.79 created (now a0e06e4)
gregor herrmann
gregoa at debian.org
Thu Sep 25 22:43:54 UTC 2014
This is an automated email from the git hooks/post-receive script.
gregoa pushed a change to annotated tag upstream/1.79
in repository libsql-abstract-perl.
at a0e06e4 (tag)
tagging 0de190d367edc77ebbb94269008856dc49b806db (commit)
replaces upstream/1.78
tagged by gregor herrmann
on Fri Sep 26 00:26:33 2014 +0200
- Log -----------------------------------------------------------------
Upstream version 1.79
Alexander Hartmaier (1):
Add test for SQL statement that used to fail before the refactor
Arthur Axel "fREW" Schmidt (80):
Add non-equality tests for ordered lists
count(*), count(a) and count(0) should not be equivalent
Release 1.64
set version to 1.64_01 for dev users
cleaner dev version stuff; thanks to riba
new manifest file so that we do not get .svn dirs
Release 1.65
set version for dev users
this should have been release 1.66, but I forgot to commit it when I released :-(
release 1.67
for people working from repository
YASQLATBranch
break out Tree
document hopes and dreams
add format command
OO-ify and add some silly colors for fun
fix SQLATest, add more rudimentary formatting and some depth
rearrange things, looks almost good
good formatting for both subqueries and otherwise
do not indent for the first select
less warnings
initial profile and configuration support
sensible profiles and accessors for formatting
tests and slightly better profiles
test for no formatting
add html profile
add test for somewhat complex sql and add extra config for missing keywords
initial formatting script
add one more hope and dream (This should be easy to add)
test parse to tree
parse RNO, better coloring for some keywords
add space to fix stuff pressing up against last paren
test coloring of keyword
whitespace tests
roundtrip tests
add more keywords, colors, and indentation rules
add examples and DBIC::Storage::PrettyPrinter
from has to stay
fix colors for from
add some docs
a bit less indentation
fix deps
slightly nicer newline handling for stuff in parens
update changes
rename DBIC::Storage::PP and get ready to re-release
change overridden method to pretty print more
add quotes to examples to show hidden whitespace
fix pod coverage test
add placeholder support
fix oddity in initial newlines
integrate placeholder population with DBIC
in case there were no bindargs passed
merge configs with profiles
add missing dep
initial start of warn-style caller info
caller info does not belong here
docs for ::Tree
release 1.67_03
use object interface to Hash::Merge
doc additions and a method rename
placeholder coloring!
add a bit more doc, test doc, and rename a method
release 1.68
add test for parsing lists
add a couple more examples
make format-sql actually useful
s// does not return the modified string, $_ does
Changes for past two commits
fix placeholder surround and add complex test
fix tree tests
release SQL::Abstract 1.69
fix placeholder quoting
fix pagination keywords
Hide bulk inserts from DBIx::Class
Highlight transaction keywords
make console color profile a bit easier to read
Add EXECUTING for clarity of long running SQL
allow disabling of show_progress; _ some accessors
better bulk insert support
try DBIC colorizer!
Arthur Axel 'fREW' Schmidt (28):
Add "no_repeats" option to fix it such that repeated SQL may be ellided except for placeholders
test for no-repeats
squash_repeats > no_repeats
Do not show_progress by default
tests for bulk insert
fix precedence for executing in new
fixup tests
docs for pretty printer
Highlight HAVING
Add error checking on "profile" for SQLA::Tree
Fix missing doc (RT#62587)
Document repo
Leave quotes from DBIC in bindargs
better quoting logic
no raw ASCII
sort changes by importance
Release 1.70
migrate _parenthesis_unroll to SQL::Abstract::Tree
Release 1.72
take out useless _table option
Stop filling in placeholders in `format-sql`
fix regression introduced in 6f2a5b6
release 1.73_01
release 1.73
Add (now passing) reassembly test
An old (now passing) test for regression introduced in 6f2a5b6
Original (now passing) test for 08e163601
Fix the reassembler to treat RNO as the precious special snowflake it is
Ash Berlin (14):
Starting import of SQL-A
initial "import"
bumped version, added keywords
keyword cleanup
reorg
just comments and email address
changed email to CPAN address
r
added patch from Guillermo to cleanup in/between,
abstract from code imported from DBIC::SQL::Abstract plus tests
quoting tests
added patch from Guillermo to cleanup in/between,
Sanitised layout - now to start hacking
Added { -desc => 'foo' } order by support
Dagfinn Ilmari Mannsåker (19):
Fix parsing of foo.* in SQLA::Tree
Add support for NULLS FIRST/LAST in ORDER BY
Fix -nulls value case-insensitivity
Revert "Add support for NULLS FIRST/LAST in ORDER BY"
Fix typos in POD and comments (RT#87776)
Document and test non-reference { -in => ... }
Extra test from the DQ branch (passes in 1.5)
Clarify in => undef exception (based on 90c87778)
Port fce4b7a9 from the DQ branch (with warning expectation ammendments)
Actually use the descriptions in the test cases
Move simple IN/BETWEEN tests to t/05in_between.t
Parentheses are significant in IN tests
Diag input and output for unexpected success
Check for extra warnings, and diag input and output
Add basic .dir-locals.el
Escape closing quote character in table and column names
Correct comments about overload methods vs. ->can()
Fix typos in literal + bind examples
Warn if _recurse_where is called in scalar context
Guillermo Roditi (3):
bumped version but not $REVISION. small fix to order_by plus more tests. docs pending
one more test and some simple docs
fix for key => [] + tests + cleanup of 02where.t
John Napiorkowski (2):
added test for the html profile destroyed placeholder problem
examined placeholder status a bit more carefully
Justin Hunter (13):
branch to work on -and => [], -or => {}
fix for hashref $logic
patch for -between to handle [\"", \""] and \["", @bind] with accompanying tests
readd the failing test back in
fix for x => { '!=' => [ -and => (1 .. 3) ] }
remove parens from test, they do not matter
add support for order_by => [qw/colA colB/]
add in my changes
patch to suppress warnings in case first element of the arrayref is undef (NULL)
add handling for NULL when using -in
fix git url, RT67137
fix some pod per RT65556
_debug not debug, RT#67179
Laurent Dami (16):
commit 8a68b5be44b4394787aa83b8b452403b3214e16e
1.5 release candidate : filling some holes in the doc
more improvements / fixes in documentation
moved internal test module into published SQL/Abstract/Test, so that clients of SQLA can take advantage of it for writing their own tests.
fixed bug signaled by Zbigniew Lukasiak : -and => [{}, {foo => 1}]
patch by Norbert BUCHMULLER (order_by => undef)
numify version
patch by Norbert BUCHMULLER: arguments to 'where' that are blessed objects with a stringification method are treated like scalars.
DBIx::Class directly calls _recurse_where in scalar context, so we must implement it, even if not in the official API
patch from mendel, using Test::Builder
fixed _refkind for \$object, \\$object. Added tests for _refkind
fixed the failing tests
updated doc to officially support the -and/-or => [...] syntax instead of [-and/-or => ...]
doc improvements, restructuring and adding explanations for the -and/-or stuff, undocumenting the -nest[0-9] stuff
corrections/additions in POD, after remarks from ribasushi
test -nest inside an arrayref
Marc Mims (1):
fixed open outer parens in a multi-line literal
Marcus Ramberg (1):
SQLA syntax regression from 1.24
Mark A. Stratman (1):
Added (now passing) test for @@ full text search
Matt S Trout (1):
moved SQL-Abstract to new layout
Moritz Onken (2):
added failing test for -desc => \['colA LIKE ?', 'test']
order_by: added passing test
Naveed Massjouni (1):
Cleaning up docs as ribasushi suggested. Made it clear that literal sql
Nigel Metheringham (13):
New branch for -bool operator integration
Added -bool/-not_bool operators - required some refactoring
Imported POD/POD::Coverage tests
Made unary_ops a direct equivalent of special_ops with supporting documentation.
Doc typo fix
Reduced regex munging of operators and streamlined backcompat syntax implementation.
Extended -bool handling to handle more types of parameters
Removed debugging code accidently added
Documentation tweak on how you handle booleans
Added check for defined to str equality test
Added some tests for -not
Trivial doc fix (malformed pod markup)
Improvements to where clause docs for undef/NULL - tackling RT Bug #58490
Norbert Buchmuller (41):
Added two more quoting tests for the 'colA DESC' stuff - to document the expected behaviour.
Changed behaviour to not stringify blessed objects, but pass them through to the bind data structure untouched.
Made stringify_bind() work with 'bindtype' set to 'columns' and 'array_datatypes' set to true.
Added test cases for passing arrayref bind values to insert() and update() when the 'array_datatypes' constructor parameter is set to true.
Converted all test cases where the tests are queued in an array to calculate the test plan from the length of that array.
Updated svn::ignore.
Support for the { operator => \"..." } construct (to embed literal SQL).
Refactored to use _SWITCH_refkind in _where_hashpair_HASHREF.
Updated doc to refer to Test::Builder instead of Test::More where appropriate.
Support for the { operator => \["...", @bind] } construct (to embed literal SQL with bind values).
Documented { op => \'...' } and { op => \['...', @bind] } syntaxes.
Added { op => \["...", @bind] } to the changes list.
Removed a wrong test (tested for overloaded stringification).
Updated dependencies.
Updated dependencies. (made a typo in previous commit)
Updated svn:ignore.
Reimplemented SQL::Abstract::Test::eq_bind to compare the data structures instead of stringifying them.
Replaced eq_bind() implementation to use Test::Deep::eq_deeply().
Added dependency to Test::Deep.
Added myself to the contributors list.
Merged changes from the '1.50_RC-extraparens' branch.
Removed an unnecessary 'use'.
Passing a hashref as bind value in insert() does not blow up, but passes the hashref through (like older SQLA versions did) and emits a warning. Later to be changed to die (with a helpful message).
A new predicate sub: eq_sql_bind
Added test for an expected warning.
Changed author field in Makefile.PL to mst.
Added test cases to test the arrayref (\[$sql, @bind]) support if @bind is empty.
Fixed a bug in hashpair with literal SQL and bind value (ie. col => { op => \[$sql, @bind] }) when bindtype => 1.
Formatting.
Applied patch from Zbigniew Lukasiak (with slight modifications) to accept "col => { -in => \[$sql, @bind] }".
Fixed 'use constant' statement to work on Perl 5.6.
Bumped revision number.
Fixed behaviour of 'literal SQL with bind' feature (ie. \[$sql, @bind]) to expect @bind in the same format as returned by ->where().
Added documentation of the impact of bindtype on \[$sql, @bind].
Added test cases for "col => {-in => \[$sql, @bind]}" with bindtype => 'columns'.
Bumped version number to 1.49_03.
Added tests and updated POD for bindtype = 'columns' case in \[$sql, @bind] when column_meta is a reference in the [column_meta => value] format bind values.
Bumped version number to 1.49_04.
Fixed a typo in Changes.
Bumped version number to 1.49_05 and updated Changes.
Fixed my typos/wording errors in the POD.
Olaf Alders (1):
Normalize the select docs s/$table/$source/
Peter Rabbitson (227):
Update Changes
Do not version MANIFEST
Failing tests by debolaz
Add value-only insert test with a literal sql snippet
Deal properly with () in is_same_sql_bind
Cleanup test and add a failing test case against RT#43483
Clarify insanedness of test
Better handling of borked sql in tests
Fix is_* in test to return the correct test value
Enhance values test
Bump to 1.50
Failing nested modifier tests (the tests themselves are debatable - possible regression from 1.24
Correct nested modifier tests
Framework for modifiers tests
test SQLA::Test shortcommings
Massive SQLA::Test rewrite - make it handle ()s sanely, instead of pruning them at will
Comment clarify
Consider 'X BETWEEN' a statement keyword
Dummyfy test for now
Reverting massive test breakage by LDAMI - this stuff can not happen in 1.x, has to wait for 2.x (some tests changes are reasonable, acking where appropriate)
Preliminary fix by arcanez
More modifier tests, all appear to be good (more needed)
Sanify patch by arcanez - we do not gobble any -X op, just -and/-or
A yet more-smarter parser for SQLA::Test - any smarter than that and we go skynet
All functionality tests in, except for nest and opX warnings
mod_N deprecation tests
Better test exception diag
Warn about deprecation of and/or/nestX
Split test for clarity
Massive modifier test cleanup
Someone else would have to write -nest tests
Release 1.51
This todo is active on the cpan 1.51 release, still needs resolving thus leaving as failing
Remove CR/LF and exec bits
Silence warning on bleadperl (by SMPETERS)
Strip some parenthesis for easier reading
Merge 'and_or' into 'trunk'
Make POD more readable, add a (failing) multikey order_by test
Wrap up order_by saga
Refactor the parenthesis unroll SQLA::Test code
Add -nest tests by ldami
Disable t/10test.t for ordinary users (suggested by ldami)
Parenthesis do matter in -nest tests
Revert r5976, trying something better
A much better fix for the x => [ -and, ... ] corner case
Merge 'trunk' into 'and_or'
Merge 'and_or' into 'trunk'
Add myself to contributors
Release 1.52
Make sure hash key sorting is maintained throughout SQLA
Add license POD tag, MI friendly
Switch installer to MI
Tweak tester disclaimer
Release 1.53
Deprecate -nest in favor of -paren
Back out changes, needs more work, move to a branch
New branch to allow special op overriding
Allow special op handlers to be a method name alongside a coderef
Merge 'special_op_handling' into 'trunk'
Release 1.54
Test and fix for obscure where-cond modification
Release 1.55
Fix for _order_by with bind values - will not work on DBIC - needs matching changes to SQLAHacks
Do not join hash order conditions early
Release 1.56
New branch to work on SQLA::Test
Merge 'trunk' into 'bool_operator'
Final cleanups
Merge 'bool_operator' into 'trunk'
Fix Tester to deal properly with NOT and single parenthesized expressions
Release 1.58
Fix some warnings
Merge 'trunk' into 'test_refactor'
Omnipotent 'between'
IN testing support
Allow scalarref in IN and open up non-grouping parenthesis around IN arguments (saves sqlite from brain damage)
Documentation and changes
Release 1.59
fix a dumb dumb regex error, /me --
Release 1.60
New op for functionality necessary for RT#39121
Failing test
optional functionality for INSERT
Merge 'trunk' into 'test_refactor'
Enhance/detabify
Things look saner now
Merge 'test_refactor' into 'trunk'
Release ready - still need to coordinate with DBIC
Merge 'trunk' into 'arbitrary_op_nesting'
Release 1.61
Lose a couple of oddball dependencies (while moronizing the tests a bit)
Slightly change the license wording (to satisfy the M::I license guesser) while keeping the former verbosity
Factor out insert returning generator
Merge 'trunk' into 'arbitrary_op_nesting'
Looks like its finally solved
Simplify a bit
Propagate bindtype properly over nested functions
Changes
Merge 'arbitrary_op_nesting' into 'trunk'
Clean namespace
De-golf
Extra nesting test and fix
Bump trunk version
Shuffle dependencies a bit more
Saner MANIFEST.SKIP
Whoops - forgot to commit
ILIKE is common enough to be in core
Release 1.63
Fix RTs #56062 and #56258
10% speed up on quoted statement generation
Doesn't get any tighter
Stupid omission
Add missing bindtype assertions
Puke in bind-assert and rewrite test to stop T::E from puking itself
Whitespace/comment cleanup
Fix SQLA::Test problem
Bump trunk version
More forgotten changes
Fix missed preexisting borked test
No versioning before releaseChanges
Commit changes to previous erroneous commit
Merge 'sqla-tree' into 'trunk'
Move the colordefs to the profile def
Tokenizer fixed \o/
Stupid stupid typo
Reenable the tester test - it's fast enough now
Replace one more grep with a regex
Fix incomplete handling of IN/BETWEEN sub-args
Backout errnoeusly committed stuff
Fix bindtype omission in -in handling
These 'optimizations' are silly
t/01generate.t
Fix bogus test (does not pass yet)
Add list support, and various mini-fixes
Preserve \@bindargs passed to unparse()
Properly support ops containing _'s (valid in Oracle)
Proper placeholder support in the AST
Fix incorrect padding
Fix default value inserts (an op can have an empty arglist)
Add support for { -op => $foo } in UPDATE arguments
Reenable show_progress in example
gitignore
Make resulting SQL work on MySQL
Disable dbic-tests when no dbic is present
Release 1.71
Fix retardation in -between bind handling
Warnfree tests
Ask for a faster CAG, switch to MI::AI
Better diagnostics (unparse the parent of the failed chunk)
Pass options to _insert_returning and allow bind rvs
Fix stupid assumption in parenthesis unroller
Undocument -nest with extreme prejudice
Require bugfixed CAG
injection_guard implementation
Consolidate the injection_guard checks, fix bobby's name
Fix stupid not exists omission
Fix recursion warnings while parsing extremely long lists
Fix incorrect comparison of malformed lists
Add some crazy malformed SQL to make sure things do not throw
Do not import Carp functions into the namespace
Promote AS to a binop of sorts
Fix over-eager parenthesis unrolling (only legal in AND/OR)
Massively refactor arbitrary sql parser code
Cleanup parser a bit more, anchor everything correctly, etc
Migrate -ident and -value from dbic, reorganize docs
Upgrade M::I
Pull in some dependency bugfixes
Fix false-negative function comparisons
Pave the road to more global state
Drop the silly Dumpered bind comparison
Stop differentiating between ORDER BY foo and ORDER BY foo ASC by default
All roundtrip tests now look for the exact string
The ORDER BY parsing fix in 73835ff0 only worked by accident
Change parser to not eagerly slurp RHS expressions it doesn't recognize
Stop unknown token processing early in case a potential unknown function lies ahead
Make tests pass without Storable
No dynamic deps
release 1.74
Finish a RHS branch before applying ASC/DESC mods
Kill the plan
Backport undef-with-in tests from DQ, add "roadwork ahead" exceptions
Do not yell loudly on TODOified is_same_* tests
And yet more improvements in the parsing engine
We already have Storable as a dep, this is pointless
Test BETWEEN after the IN debacle (luckily it throws in all expected cases)
Make it clearer in examples that bool takes expressions
No tabs, no trailers
Extra stopword (we need to find a way to introduce dialects...)
Centralize handling of in-test dumpering
Remove many of the settled-by-time comments, modernize a bit
Normalize handling of expected warnings/exceptions in tests
Move test to a better location (no changes)
Extensive tests and deprecations of multivalue/no value inequality op calls
Reintroduce handling of IS( NOT) ops
release 1.75
Ensure that multi-nested parenthesis are never unrolled after IN
Fix regression in column level { not => undef } op
Port scalarref in array of sources test from dq (7732c37a)
release 1.77
Fix the binop leg of the parser to correctly consider only a single LHS node
Treat RNO as a unary op, and properly normalize it
Use variables to simplify grocking the parenthesis unroller rules
Port the \[] *_sql_bind unpacker from DBIC
Swap CAG with Moo, lazify a lot of the profile selection
Rewrite sql formatter script and shove it into examples until more tested
Drop reqs - works on 5.6.1 just fine
release 1.78
Fixup descriptions a bit
After 5f3fa0ac script/ is no more
Fix -values not being handled properly with NULLs
New exportable functions: is_literal_value($) and is_plain_value($)
Aha - this is why I didn't use ->can originally
{ -ident => undef } makes zero sense
Even more reuse of @_
Make sure is_plain_value returns the actual object pre-stringify
Even more versatile is_plain_value testing - feeble attempts to trip up perl
More (still passing!!!) test cases after conversation with FC
An arrayref makes sense for literals, but no sense for values
Boolification-derived-stringification is a thing... just shoot me now
Well how about them apples?! Add temporary overload-madness escape hatch
IRC meta
Truly fix literal+bind examples (build upon a3e3e6a18)
Link fixes (no other changes)
More robust Moo import block (modelled on DBIC)
Deprecate and properly handle empty lhs
Even though MRO::C is a dep of Moo, specify it explicitly
Release 1.79
Rob Kinyon (3):
Fixed the problem with values() not behaving the same as the rest of the code.
Fixed interjecting arrayrefref into a where clause
cp -r from branches/1.50_RC where svk push won't work
Robert Bohne (1):
Fixed small POD mistake
Rudolf Leermakers (2):
Minor documentation fixes; typos, POD formatting code syntax, return value explanation consistency
Updating links to repositories, added GitHub mirror
Ryan P. Kelly (1):
Document with the right name
Wes Malone (1):
Removed excess parentheses in generated SQL
gregor herrmann (1):
Imported Upstream version 1.79
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libsql-abstract-perl.git
More information about the Pkg-perl-cvs-commits
mailing list