Peter Rabbitson [Wed, 24 Jun 2015 11:04:28 +0000]
Remove no longer needed try{} block in related rel instantiation
This try{} is no longer necessary after
3b4c4d72, but was missed due to
its purpose not being readily apparent. The only known case where this
exception could be raised and be silenced is if an uninserted object had
some filter relations populated directly via {_inflated_column} *before*
related_resultset() has ever been called. This would result in a cascade
of failures during the relationship resolution, and bubble up here.
Starting with
3b4c4d72 the empty related resultset is instantiated ahead
of time, and everything works properly from there on.
In the unlikely case this does cause issues downstream, a way will be
found to address this ( worst case scenario by reintroducing the try{} )
Peter Rabbitson [Wed, 24 Jun 2015 07:43:19 +0000]
Remove dead code from CDBICompat::ImaDBI
This is a forgotten piece from back in
902133a3, overtaken by a local copy of
Class::DBI::SQL::Transformer. Nothing called/exercised it since then
Peter Rabbitson [Tue, 23 Jun 2015 11:04:49 +0000]
Simplify and rename _resolve_prefetch
No functional changes, just logic and arglist cleanup - a couple of them were
no longer used: $order was forgotten in
d4d8e97b, and the alias-prefix went
unused somewhere around that time as well
Peter Rabbitson [Tue, 23 Jun 2015 10:16:54 +0000]
Move prefetch selector injection a bit further down in rsattr resolution
No functional changes, just a c/p
Dagfinn Ilmari Mannsåker [Wed, 3 Jun 2015 20:40:06 +0000]
Stop importing all the POSIX functions
The only one used is called fully-qualified anyway
Dagfinn Ilmari Mannsåker [Wed, 27 May 2015 07:39:58 +0000]
Fix t/storage/quote_names.t failures, caused by lazily built sqlmaker
While the entire test is written badly (it ought to use the OptDeps system)
it would be too much churn to fix it properly for little benefit. Additionally
there is a very good chance that the shortener will become a core dep in the
near future, hence the entire point will soon be moot anyway --ribasushi
Peter Rabbitson [Wed, 20 May 2015 14:11:39 +0000]
Add extra (passing) test for operations on sourceless results
This should have been committed with
84b1ec108, reevaluated while reading
through
cb909f39c of GH#78
Peter Rabbitson [Wed, 20 May 2015 13:25:06 +0000]
Replace all instances of ; in the rowparser loop with , ( ~15% speedup )
The functional diff best examined via `... --color-words -w lib ...`
This is another microoptimization, which is nevertheless measurable on 1000+
row resultsets. Investigation into this was prompted by a curious comment made
by bulk88[1] in response to Glenn Golden[2]:
Every ";" is an op that executes that adjusts current line number
Given that we already generate all this code, and it is not intended to be
read by a human, swapping ; with , is only logical. A (very very crude) check
with the included benchmarker indeed shows a repeatable difference on 5.16.2
[1] http://www.nntp.perl.org/group/perl.perl5.porters/2015/05/msg228074.html
[2] http://www.nntp.perl.org/group/perl.perl5.porters/2015/05/msg228068.html
Peter Rabbitson [Tue, 3 Mar 2015 23:06:47 +0000]
Rewrite collapsing parser while loop (mere op-level golf)
Peter Rabbitson [Thu, 26 Feb 2015 13:22:39 +0000]
The encode-and-substitute in the generator is not necessary since
a8f62ee08
Peter Rabbitson [Thu, 26 Feb 2015 12:51:49 +0000]
Rewrite the collapsing rowparser pigeonholing logic
Instead of dynamically determining whether to use the data as a key directly,
always compose a list of copied 'row ids' and use that in all operations.
This is mainly done to stop dualvar-ing of numeric variables returned by the
engine, thus confusing JSON and similar serializers down the road
Peter Rabbitson [Sun, 7 Jun 2015 11:49:50 +0000]
(travis) Add Schmorp's "stable perl" fork to the smoke rosters
A note on motivation: I personally fully support Marc Lehmann's assertion
that the current direction of the perl5 development effort leaves a lot to
be desired (put in the mildest terms possible). The amount of changes[1]
necessary to get a *very* widely used module working again, and the failure
of the original authors to abandon the problematic direction *on their own*
is nothing short of an embarrassment. I fully recognize the reason why Marc
elected to attempt to maintain a fork (a nontrivial long-term commitment),
instead of engaging in another round of politicking and downright begging.
The inclusion of "stability perl" in the DBIC full-test matrix is the very
least I can do to support this effort.
--ribasushi
[1] http://stableperl.schmorp.de/dist/stableperl-5.22.0-1.001.diff
Georgina Thevenet [Fri, 27 Mar 2015 18:35:54 +0000]
(travis) Add explicit sudo:true to .travis.yml
The 'sudo' setting now defaults to false on establishing of new github=>travis
linkages, so the CI runs will fail for anyone who wants to smoke their own
fork.
Peter Rabbitson [Tue, 19 May 2015 03:49:48 +0000]
Make sure tests pass without a compiler present (another step to RT#74706)
Peter Rabbitson [Wed, 10 Jun 2015 14:18:01 +0000]
Make sure tests still pass in a fork-limited environment
Inspired by a temporarily stuck smoker
http://www.cpantesters.org/cpan/report/
751da1f2-e3ff-11e4-a1d1-8536eb4f9f07
Read under -w
Peter Rabbitson [Fri, 3 Apr 2015 11:30:27 +0000]
Make sure plain install with DBICTEST_VIA_REPLICATED=1 still works
This is an omission from
fd2c6658, unnoticed all the way until previous commit
Peter Rabbitson [Thu, 2 Apr 2015 01:04:24 +0000]
(travis) Multiple fixups to the CI lifted from blead:
- Much lighter CI runs when smoking a PR
- Revert backout of ODBC firebird testing
aa0e82a2
- Restore and enhance post-success to run a full CPAN install cycle
- Cut down the amount of APT repos during updates
- *FINALLY* remove Module::Build::Tiny front-loading
- Double check M::B::T is not pulled in
- Peg Module::Install - new versions deps on autodie, which is 5.8.4+
- Enhance perl sterilization to somewhat work on < 5.12 as well
- Correctly detect CI on github forks
- Improve the "Did it really install?" check
- Better shuffling of the prereq installation order
- Fix silencer changes (
f8200928) which were practically undone by 8b60b921
Peter Rabbitson [Thu, 30 Apr 2015 16:53:25 +0000]
This is no longer necessary after
4068e05f
Peter Rabbitson [Fri, 3 Apr 2015 22:18:39 +0000]
Fix missing handling of on_(dis)connect* failures
Time for a short story (see last paragraph for TL;DR): Some hours ago an
innocent question was asked in the #dbix-class channel, apparently
related to a mysterious failure to determine the RDBMS version. Before I
was able to properly investigate the problem, someone else piped up with
"Oh I've seen this before, and I worked around it". That was of course
without telling anyone else </headdesk>
After a little back and forth it became apparent that if the on_connect
settings could not be executed for whatever reason, the result may be an
implicitly failed connection attempt which is *entirely undetected* as
it leaves the original $dbh in place with some or all of the on_connect
instructions not having executed as expected.
To reiterate: if one manages to:
* supply a malformed on_connect*
and then
* call as the first order of business some codepath that can mask away
the connection failure (say ->deploy under sqlite)
then the result is a $schema with a proper $dbh in the storage instance,
but *without* some (or all) of the on_connect instructions having been
executed. As a bonus the actual call to get e.g. the RDBMS metadata (like
version etc) will have failed completely silently as well. Arguably a
rather problematic bug.
The above and by extension an issue with silent on_disconnect* failures
are bith fixed by this commit. This was possible *only* because the
second person came into the channel and reported it, instead of silently
fixing things and moving on to whatever they were doing.
Moreover while preparing the (relatively modest) patch to fix this issue,
a small refactor revealed a pretty serious bug in the XS accessor provider
Class::XSAccessor: https://rt.cpan.org/Ticket/Display.html?id=103296
Yes, it took about 3 hours to diagnose this and hunt down an isolated
failcase, but the bug severety (and its elegance) are totally worth it.
Moral of the story: *PLEASE* report issues upstream. Even if you figured
out a quick workaround, the devs still need to know about the problem you
did encounter in the first place. Bugs that nobody knows about can *NOT*
be fixed, and our software commons can not improve. So *PLEASE* speak up.
Peter Rabbitson [Fri, 3 Apr 2015 22:14:00 +0000]
Minor code reformat, to shorten the diff of next commit
Zero functional changes, read under -w
Peter Rabbitson [Mon, 27 Apr 2015 10:23:34 +0000]
Fix corner case of stringify-only overloaded values
Just a trivial cleanup, uncovered during the de-Path::Class work
Peter Rabbitson [Sun, 5 Apr 2015 11:52:11 +0000]
Rewrite dependency lister - now produces *much* easier to read output
Knowing where a module came from is quite valuable, but the segmentation done
in
cebc0cc8 makes it rather hard finding a particular module. Instead add
an extra prefix with shorthand "dirs of interest" names, and putput the entire
list in ci-alphabetical order.
Additionally make sure we explicitly print out our versions under CI (a silly
omission from earlier), and explicitly omit loading dist_* optdeps
The rough list of changes:
- Explicitly report expected but failed-to-load modules
- Add knowledge of './t' => T, under same rules as LIB
- Add support for (site|vendor)lib_stem (whatever that is)
- Add support for (site|vendor)prefix
- Add support for ./blib/(lib|arch)
- Add support for an explicit CWD marker
- When no sourcing is found - report the @INC index if possible
- If a module came from {SVP}{AL}, list it even if no version is defined
- Explain missing checksums
- Display the contents of @INC, making it even easier to follow sourcing
While the diff is massive, the actual logic did not change in any significant
way.
Peter Rabbitson [Fri, 3 Apr 2015 12:55:24 +0000]
Update Changes with stuff already shipped to CPAN as v0.082810 and v0.082820
Kent Fredric [Thu, 2 Apr 2015 11:54:19 +0000]
Do not load PadWalker at all on plain installs
Avoids a SEGV on perls with unusual compilation options
Dagfinn Ilmari Mannsåker [Wed, 1 Apr 2015 17:52:22 +0000]
Work around Firebird/InterBase/ODBC crash in tests
If DBD::Firebird and DBD::InterBase are used in the same process, the
last one loaded leaves an active statement handle around and later
segfaults at some point in the tests. Also, if DBD::ODBC is used
before either of them, they leak the handle but don't segfault.
Thus, to avoid breaking installing DBIC if more than one of
DBICTEST_FIREBIRD(_INTERBASE|ODBC)?_DSN is set, skip Interbase if any
of the others are set, and make sure to run the ODBC test last.
However, keep the leak-triggering order for author, CI and smoker
runs, so we don't forget that it needs fixing.
Peter Rabbitson [Sun, 5 Apr 2015 12:41:56 +0000]
Add an explicit DateTime::TimeZone::OlsonDB optdep
It happens to lazy-load a lot of extra modules, which both hides them from the
t/00describe_encironment lister *and* messes with preforking.
Peter Rabbitson [Mon, 4 May 2015 13:44:00 +0000]
(optdeps) Explicitly force group names to lower-case (for now)
Peter Rabbitson [Mon, 27 Apr 2015 09:01:29 +0000]
(optdeps) Rename an inconsistent 'icdt' marker to 'ic_dt' before we've shipped it
Peter Rabbitson [Mon, 27 Apr 2015 08:54:21 +0000]
(optdeps) Skip-spec brainfart in offline mysql/pg test specs
The _-prefix played a role in an earlier deficient implementation, which
was later thrown away. Make the relevant tests require-lines saner
Peter Rabbitson [Wed, 1 Apr 2015 20:52:41 +0000]
(optdeps) Fix incorrect optdep output introduced in
34d2deae / 31c31b8d
Too many thinkos for a simple feature like this. When given "foo" cpanm
actually tries to find '"foo"' on CPAN, and nothing works. The original
thinking was that the > in ~>= needs escaping when copy/pasted by a user, so
the ""-enclosing seemed like a valid approach. Of course this does not work
at all when the output is piped into cpanm directly:
perl -Ilib -MDBIx::Class::Optional::Dependencies=-list_missing,deploy | cpanm
Moreover, all of the above is *utterly unnecessary*, as Foo~>=2 is entirely
equivalent to Foo~2, which requires no escaping (thus quoting) whatsoever.
Hopefully this is the end of the embarrassing flailing on this feature :(
Alex Beamish [Wed, 4 Sep 2013 17:45:39 +0000]
Examples and documentation update and cleanup
This version updates the original POD with slightly more information
so that more DBIx::Class functionality is revealed. The source code is
removed from POD, and instead the reader is directed to the examples
directory.
It also changes the example schema from silly auto-inferred column
names to proper *id-suffixed ones.
Alastair McGowan-Douglas [Fri, 7 Nov 2014 17:00:21 +0000]
Inflators are not respected by find() (being a ResultSet method)
Adjust docs and add an extra IC::DT-specific FAQ section
Karen Etheridge [Wed, 8 Oct 2014 19:42:49 +0000]
Cross-reference alias with current_source_alias
(slightly modified
4e4b848d)
Mike Wisener [Tue, 30 Sep 2014 03:58:59 +0000]
Clarify load_namespaces call -- more than one call can be made
Robert Stone [Sat, 27 Dec 2014 19:19:10 +0000]
Added add_columns documentation about controlling float precision
C.J. Adams-Collier [Mon, 2 Mar 2015 19:41:31 +0000]
Linked "prefetch" in columns documentation
Ingy döt Net [Tue, 3 Mar 2015 19:29:39 +0000]
Example doesn't work without 'year' column
Henry Van Styn [Wed, 29 Oct 2014 13:25:33 +0000]
Fix uninitialized warnings on empty hashes passed to join/prefetch
Peter Rabbitson [Wed, 18 Mar 2015 12:15:45 +0000]
(travis) Better handle ENV poisoning, get jobs under 4mb limit again
Peter Rabbitson [Tue, 17 Mar 2015 15:31:56 +0000]
Run the entire test suite under replicated SQLite on DBICTEST_VIA_REPLICATED
There isn't real replication, the reader and writer in fact talk to the same
on-disk file. Still this is effective enough to put into CI.
Peter Rabbitson [Tue, 17 Mar 2015 15:21:47 +0000]
Fix exception on complex update/delete under a replicated setup
THIS COMMIT IS A HACK!!! The actual fix is migration of the logic to the
SQLMaker subclass and deal with it there. However the entire class hierarchy
is currently in flux, so go with the least invasive change until the storage
rewrite settles.
Testing for this will show up in the next commit
Peter Rabbitson [Tue, 17 Mar 2015 15:09:44 +0000]
Some stylistic test changes in preparation for next commits
There are zero functional changes - everything has to do with making sure
we are not blindly poking into storage objects which may be changed to
delegates AND to make sure we do not leave any open cursors, which would show
up as deadlocks on multiple connections to the same SQLite file
Read under -w
Peter Rabbitson [Tue, 17 Mar 2015 12:06:13 +0000]
Fix test suite to work again with DBICTEST_SQLITE_USE_FILE
It was never included as part of POISON_ENV and subsequently lapsed. Sigh.
Peter Rabbitson [Mon, 16 Mar 2015 14:56:53 +0000]
'undef =>' isn't what one would want
Peter Rabbitson [Fri, 13 Mar 2015 12:43:08 +0000]
Relax sanity check in _resolve_relationship_condition
It doesn't buy much and breaks legitimate class redirection techniques
(some of them even documented in the cookbook)
Peter Rabbitson [Thu, 12 Mar 2015 09:27:36 +0000]
Workaround for double-call of destructors (based on
3d56e026 and e1d9e578)
Silently fixing this up is nothing short of irresponsible, hence the
elaborate detection and alert mechanism
Instead of just marking all my DESTROYs with the function, also add it to a
base class, thus capturing *everything* there is to find. Yes it will be a
tad slower. And yes it will have a massive benefit to any DBIC user happening
to be caught in an unfriendly app-space
Peter Rabbitson [Mon, 9 Mar 2015 10:25:42 +0000]
Remove erroneous undefer introduced in
8d73fcd4 (confuses old perls in END)
While it is important to check this during namespace cleansing checks, it
actually makes no difference whatsoever during leaktests - things will either
be closed over and thus seen, or they won't be
Peter Rabbitson [Thu, 26 Feb 2015 22:25:39 +0000]
Get the leaktest to work with Test::More-(hopefully-not)-to-be
Peter Rabbitson [Thu, 12 Mar 2015 09:36:47 +0000]
Fix brainfart from
cb551b07 - 'if' is still a module, don't need it
Also rewrite the load pattern test to be saner
Peter Rabbitson [Thu, 5 Mar 2015 23:39:00 +0000]
More desc_env fixes: Win32 lib matching, and better formatting of final diag
This is what the slash-situation is on a typical strawberry:
'inc' => [
'C:/Strawberry/perl/site/lib/MSWin32-x64-multi-thread',
'C:/Strawberry/perl/site/lib',
'C:/Strawberry/perl/vendor/lib',
'C:/Strawberry/perl/lib',
'.'
],
'archlib' => 'C:\\STRAWB~1\\perl\\lib',
'archlibexp' => 'C:\\STRAWB~1\\perl\\lib',
'privlib' => 'C:\\STRAWB~1\\perl\\lib',
'privlibexp' => 'C:\\STRAWB~1\\perl\\lib',
'sitearch' => 'C:\\STRAWB~1\\perl\\site\\lib',
'sitearchexp' => 'C:\\STRAWB~1\\perl\\site\\lib',
'sitelib' => 'C:\\STRAWB~1\\perl\\site\\lib',
'sitelibexp' => 'C:\\STRAWB~1\\perl\\site\\lib',
'vendorarch' => 'C:\\STRAWB~1\\perl\\vendor\\lib',
'vendorarchexp' => 'C:\\STRAWB~1\\perl\\vendor\\lib',
'vendorlib' => 'C:\\STRAWB~1\\perl\\vendor\\lib',
'vendorlibexp' => 'C:\\STRAWB~1\\perl\\vendor\\lib'
Instead of trying to wade through it with File::Spec, just turn short names
into longnames and every \ into / and hope it works.
Also make it so that the MD5 line will align correctly on an 80-col terminal
(on wider terminals it won't matter either way), and show *all* the hashes
at all times (omitting the information makes little sense).
In addition some clarification to the undef/empty ->VERSION return error
and the diag contents themselves (Mithaldu++)
Peter Rabbitson [Fri, 6 Mar 2015 16:08:17 +0000]
Rewire OptDeps to not attempt any module loads under missing envvars
The thinking is that if some envvars are missing, it is of secondary concern
to provide the exact list of missing module reqs to the user, while the
downside can be rather severe in some cases (broken DBD::Oracle on windows is
particularly obnoxious). Thus the logic for req_missing_for switched to only
look for "definitely missing" modules via a simple @INC-walk whenever it is
already clear that the req group failed
Peter Rabbitson [Thu, 5 Mar 2015 19:14:12 +0000]
Fix t/54taint.t failures on perl-in-hell-in-space
Given its a core bug, and the test is not critical, just skip the whole thing
Peter Rabbitson [Wed, 4 Mar 2015 01:30:48 +0000]
Add comprehensive 'report-prereqs'-like tooling
Unlike what is currently available in various Dzil plugins, this is more of
a descendent of Test::DiagINC - it tries to load as much as possible and
subsequently reports on whatever can be found in the symtable.
There is no (and never will be any) attempt made to validate if all modules
listed in the various METAs are in fact available (this is what the rest of
the test suite is for).
Module versions are broken up into groups depending on their place in @INC,
without leaking the contents of @INC itself.
Peter Rabbitson [Tue, 3 Mar 2015 17:07:48 +0000]
Extract LeakTracer symtable visitor into DBICTest::Util
Minor rewrapping and renaming of variables as we copy it
No functional changes at all, verified to produce same number of visits
Peter Rabbitson [Thu, 5 Mar 2015 17:59:48 +0000]
(travis) Make a separate is_ci runmode, separated from is_smoker
Peter Rabbitson [Thu, 5 Mar 2015 18:11:31 +0000]
(travis) shunt intermittent failures stopping postgres
Peter Rabbitson [Thu, 5 Mar 2015 15:07:13 +0000]
Stop noise from a particular combination of DateTime and DateTime::TimeZone
While things went quiet for a while after
556c4fe6, it seems that another
refactor made things go awry again. I can't actually reproduce this on any
of my machines, including Travis. Besides this isn't anything to worry much
about anyway - just kill the test
Peter Rabbitson [Thu, 26 Feb 2015 16:47:29 +0000]
At last a safer Moo (lots of kludges undone)
Peter Rabbitson [Mon, 2 Mar 2015 06:08:00 +0000]
Remove warning introduced in
75a1d824d
Peter Rabbitson [Mon, 2 Mar 2015 06:02:24 +0000]
(travis) Only consider the runmode a 'Smoker' when we CI ourselves
Peter Rabbitson [Sat, 28 Feb 2015 15:57:39 +0000]
(travis) Update apt cache when doing a large set of installs
Started failing without it:
https://s3.amazonaws.com/archive.travis-ci.org/jobs/
52536716/log.txt
Peter Rabbitson [Thu, 26 Feb 2015 18:46:47 +0000]
Move a number of tests to xt, restructure extra lists
Instead of the rigid t / xt separation, add more nuance and move some heavy
tests around, in order to improve regular user test times, without
sacrificing cpantesters coverage
Additionally makes it more obvious which tests are decidedly not being run
on user-side installs
Minimal changes: read diff under -C -C -M
Peter Rabbitson [Thu, 26 Feb 2015 18:22:22 +0000]
Factor out author count anouncement, making it seen on plain installs
Peter Rabbitson [Thu, 26 Feb 2015 16:18:30 +0000]
This hasn't failed in over a year, enable for plain installs
Peter Rabbitson [Sun, 15 Feb 2015 10:08:01 +0000]
Rewrite hot-ish _dbi_attrs_for_bind codepath in a leaner way
Document various caveats better
Kevin L. Kane [Tue, 18 Nov 2014 13:02:33 +0000]
Fix updating multiple CLOB/BLOB columns on Oracle
The genric _dbi_attrs_for_bind caches the attribute hashrefs by data
type, so we can't modify them directly with column-specific data.
Instead, copy it and add the ora_field attribute to the copy.
Peter Rabbitson [Tue, 10 Feb 2015 18:16:03 +0000]
(travis) Run a couple tests in single-threaded mode
Semi-dogfooding - the massive paralellism hides the pain of how long exactly
does it take for a common user to install DBIC
(FWIW *dependency* installer timeout had to be bumped from 9 to 15 *MINUTES*)
Peter Rabbitson [Tue, 10 Feb 2015 18:12:15 +0000]
(travis) Temporarily kill Firebird ODBC testing
Sourceforge no longer facilitates downloads without a javascript enabled
browser (WAT?!), so the Odbc driver download fails. Disable the testing for
the time being, to be reinstated with the DB2 testing work
Peter Rabbitson [Tue, 10 Feb 2015 15:17:55 +0000]
Switch CDBICompat and its tests to OptDeps
This work is a direct descendant of ilmari's
c8dc7d33, even though it does
not contain almost any of the original code. While the original approach was
a big improvement, this commit is taking advantage of all improvements that
took place in OptDep handling in earlier commits. The result (at the expense
of a much much larger changeset) is a skip time of < 0.5s for the entire cdbi
set
ilmari++ # this revamp would not have taken place without your first nudge
Peter Rabbitson [Tue, 10 Feb 2015 13:24:27 +0000]
Remove redundant skip checks in CDBICompat tests
f56e59cd added the extra module, but nothing in the tests actually loads it
under Film (there likely was a switch during dev, but it was never committed).
Just rip it out entirely.
Peter Rabbitson [Sun, 25 Jan 2015 10:35:23 +0000]
Remove useless INIT blocks from CDBI tests - no changes
A use() is compile-time even when it is in an INIT block, hence removing the
INITs is safe. Do not do it for require() invocations, as it is not clear if
the tests rely on the specific order or not (more sleeping dogs left to lie)
Read under -w
Peter Rabbitson [Tue, 10 Feb 2015 12:45:32 +0000]
Introduce ad hoc requirements and add skip_without method to optdeps
On rare occasions tests need a specific module version, but nothing else in
the dist needs this. Additionally only parts of a test may rely on this extra
requirement, making the "all or nothing" approach of -skip_all_without not
very attractive.
Hence introducing ability to supply arbitrary module specifications in lieu of
group names. Each specification is checked against a special group 'test_adhoc'
which ensures that optional deps are not "forgotten" within the suite (causing
a test to never run in practice).
Peter Rabbitson [Tue, 10 Feb 2015 08:38:51 +0000]
Finally implement compound OptDep group augmentation
A group can specify not only what it requires (with includes), but can also
request a specific set of modules IFF another group is required within the
same call. This allows to do e.g. IC::DT+RDBMS dependencies in a sane manner.
Delete test_dt* groups from our optdeps (they were never documented, and a
cpangrep does not show any use) - instead we replace them with an elaborate
network of augments.
Peter Rabbitson [Fri, 6 Feb 2015 11:37:29 +0000]
Rewrite internals of the optdep include mechanism (introduced in
e163f484b)
No functional changes, "simply" moves the include-parsing logic and the env
checks to an earlier point, in order to get a complete list of all groups
mentioned (facilitating the augment functionality in the next commit)
Peter Rabbitson [Sun, 8 Feb 2015 11:33:28 +0000]
Introduce (empty for now) intermediate optdep groups related to IC::DT
They will be filled in during the next commits, done separately to reduce
diff-noise
Peter Rabbitson [Fri, 6 Feb 2015 18:37:54 +0000]
Relax the optdep naming criteria, introduce _-prefixed intermediate groups
Peter Rabbitson [Mon, 9 Feb 2015 23:05:01 +0000]
Move the SQLite test dependency specification to optdeps
Even though this is a hard test_requires (for now), move things to a proper
specification within OptDeps like all other RDBMS (will make icdt specs
much easier).
The OptDep module is very lightweight anyway, and loading it on all installs
will be a good exercise before spinning it off.
Peter Rabbitson [Fri, 6 Feb 2015 10:39:44 +0000]
Stabilize and further test complex optdep include scenarious
Peter Rabbitson [Fri, 6 Feb 2015 10:06:01 +0000]
Clarify optdep API signatures to be explicitly single-argument
Peter Rabbitson [Sat, 7 Feb 2015 16:02:02 +0000]
Remove DateTime requirement from tests that do not rely on it
The deterministic_value test checks how stringification behaves wrt inherited
values from a resultset. Rename test while we are at it.
The plus_select.t checks behavior of InflateColumn combined with get_columns.
In both cases the column in question (year) is declared as VarChar and is not
connected with IC::DT in any way
Read under -C -C -M -w
Peter Rabbitson [Sat, 7 Feb 2015 15:02:26 +0000]
Move some ICDT dependent (sub)tests around (no functional changes)
Read diff/blame with -C -C -M -w
Peter Rabbitson [Sun, 8 Feb 2015 10:11:00 +0000]
Fix overspecified msaccess test optdep (should not include icdt)
The test_rdbms_msaccess_* groups are only referenced in two tests. The basic
one does not exercise (doesn't even load) any of the ICDT handling components
The ICDT test already explicitly requires test_dt, so no changes are needed
within the actual tests
The erroneous requirement was introduced in
726c8f65 and never revised since
Peter Rabbitson [Tue, 10 Feb 2015 13:05:40 +0000]
Remove obsolete startup check (introduced in
5e724964, superseded by a5a7bb73)
Peter Rabbitson [Fri, 17 Oct 2014 09:35:46 +0000]
Add import-time-skip support to OptDeps, switch most tests over to that
This is the meat of this refactor, inspired by multiple parties complaining
(with good reason) about the sluggishness of our skip_all()s, and by groundwork
in both
c8dc7d333 and 2c2bc4e5 (both never applied - the former is superseded
by a more direct approach and the latter being too intrusive with too little
benefit after the refactor)
Only tests with unambiguous skips were switched over - later commits will
address multi-rdbms loops in a different manner.
The loss of skip-time warnings "this test creates tables blah and blah" was
deliberate, and due to several reasons
*) They are displayed at the wrong time
*) They are not well sycnrhonized with what the actual test does
*) Some time this year all table names will switch to dbictest_-prefixing
All in all the thinking is that nothing of value is lost, and the changeset
is greatly simplified as a result.
The loading before strict/warning in tests is actually significant. It appears
that in "all deps fail" cases the speedup of not loading strictures is quite
significant (demonstrated below difference on WiP CDBICompat streamlining):
strictures first:
Files=43, Tests=0, 1 wallclock secs ( 0.14 usr 0.08 sys + 0.35 cusr 0.15 csys = 0.72 CPU)
optdeps first:
Files=43, Tests=0, 1 wallclock secs ( 0.13 usr 0.06 sys + 0.14 cusr 0.11 csys = 0.44 CPU)
Peter Rabbitson [Sun, 18 Jan 2015 13:33:28 +0000]
Add import-time action stub to OptDeps, switch distbuild checks to it
More in the following commit
Peter Rabbitson [Tue, 20 Jan 2015 16:09:44 +0000]
Introduce basic optdeps group inclusion mechanism, simplify many declarations
No logical changes: verified before and after the commit via:
perl -MData::Dumper::Concise -Ilib -MDBIx::Class::Optional::Dependencies -e '
my $d = DBIx::Class::Optional::Dependencies->req_group_list;
delete $d->{$_} for grep /rdbms_generic/, keys %$d;
delete $_->{effective_modreqs_differ} for values %$d;
print Dumper $d
' | sha1sum
Read under -w
Peter Rabbitson [Sun, 26 Oct 2014 12:23:28 +0000]
Make OptDeps aware of envvars as first-class requirements (test_* groups only)
Peter Rabbitson [Thu, 8 Jan 2015 18:33:57 +0000]
Deprecate req_errorlist_for, proxy to to modreq_errorlist_for
Leaving the old method in place indefinitely - may be used in the wild
Peter Rabbitson [Fri, 17 Oct 2014 09:27:27 +0000]
Rewrite optdeps to accept a list of groups
Requires a one-time break of req_group_list()
Also tighten up the optdep spec and handle conflicting versions saner
As a side effect fix
34d2deae to be truly copy-paste-able (no , separator)
Changes to ::Opt::Deps and xt/optional_deps.t best viewed under -w
Peter Rabbitson [Sat, 17 Jan 2015 13:23:44 +0000]
OptDeps doc-mangling, no functional changes
Switch synopsis away from M::I (it still scares people, boggle)
Document volatility of req_group_list
Better language in places
Coarser granularity of heredocs (easier to read)
Read changes under -w
Peter Rabbitson [Sun, 18 Jan 2015 22:22:39 +0000]
Minor streamline of optdeps usage (no functional changes)
Peter Rabbitson [Sun, 25 Jan 2015 09:59:55 +0000]
Standardize the struct-cloning interface throughout the codebase
Peter Rabbitson [Sat, 24 Jan 2015 19:53:22 +0000]
Neither of these can fail now that we dep on 5.8.1 (missed in
55087b99)
Peter Rabbitson [Sat, 24 Jan 2015 17:23:23 +0000]
Delete useless test
Peter Rabbitson [Thu, 22 Jan 2015 22:27:22 +0000]
Fix intermittent failures in the LeakTracer on 5.18+, remove all workarounds
ARGH! In the end the issue turned out to be stupid-simple: when PadWalker is
called it always returns us a hashref. This hashref is traversed, with its
own address recorded as seen. Once processed, the hashref itself is GCed,
leaving a stale entry in $seen_refs.
Up until 5.18 this was not a problem, likely due to no address reuse. However
after 5.18 PadWalker is very happy to return different hashrefs under the same
address over and over again. The $seen_refs entry leads to a skip, and the
actual contents of the pad are never examined - we've seen this hash already.
ARGH!
The decisive change is the single line around PadWalker - everything else is
cleanup of workarounds for this problem.
Peter Rabbitson [Thu, 22 Jan 2015 15:19:57 +0000]
Keep T::B IO open in t/52leaks.t END - other ENDs pending (esp. since
961d79db)
Read under -w
Peter Rabbitson [Thu, 8 Jan 2015 00:42:52 +0000]
Fix Postgres test hang with DBD::Pg 3.5.0+ (work around RT#100648)
It is not 100% clear to me why this is happening, as the change discussed[1]
was not supposed to change behaviour outside of a transaction.
The test in question essentially does a blocking request with a timeout. The
way the default DBIC machinery works is to issue a ping() after the timeout
induced exception, in order to opossibly retry the operation IFF the link
went stale ($dbh thinks everything is in order, the RDBMS believes otherwise).
After the changes in 3.5.0 this ping() now hangs, even though the access
pattern did not change (there are no transactions involved on the handle in
question). Either way, the fix is simple: just shim-away the ping for the
duration of the test.
In general this should be resolved on a higher level by disabling retries for
the duration of the test (or maybe even disabling retries by default in the
distant feature - see RT#47005)
[1] https://rt.cpan.org/Ticket/Display.html?id=100648#txn-1438093
Peter Rabbitson [Sun, 11 Jan 2015 09:39:48 +0000]
(travis) Remove trial SQLA workaround (no longer on CPAN) - retry another time
Read diff under -w
Peter Rabbitson [Sun, 11 Jan 2015 09:45:08 +0000]
(travis) Revert
17b09b77 - Test::Deep 0.115 now available
https://github.com/rjbs/Test-Deep/issues/22
Dagfinn Ilmari Mannsåker [Tue, 6 Jan 2015 19:24:05 +0000]
Preserve space after <nick>: in the AUTHORS section of DBIx/Class.pod