[libstatistics-basic-perl] annotated tag upstream/1.6611 created (now 00c0109)
Salvatore Bonaccorso
carnil at debian.org
Sun Jul 26 13:42:09 UTC 2015
This is an automated email from the git hooks/post-receive script.
carnil pushed a change to annotated tag upstream/1.6611
in repository libstatistics-basic-perl.
at 00c0109 (tag)
tagging 95cec17ffc380b85589d3cf97fac91ac1c5ae0fe (commit)
replaces upstream/1.6607
tagged by Salvatore Bonaccorso
on Sun Jul 26 15:24:53 2015 +0200
- Log -----------------------------------------------------------------
Upstream version 1.6611
Anon (1):
https://rt.cpan.org/Ticket/Display.html?id=74155
Paul Miller (360):
mmm git
making this a bit more modern
more updates
started adding overloaders and cleaning house a bit
I added a todo to cover the things I just started ...
added many more s/no/use/ warnings
stripped off a bunch of shitty nonsense docs
more shit
whitespace
whitespace
I couldn't get the overloader class to work ...
ipres
more todo and some re-writes/updates, etc
we won't be doing this because the new(vector,size) style makes more sense
ok, I think these tests and overloads are finalized
gar, spellcheck
now split-jumps will work
mode seems all done, curious choice for first to complete
makefile.old, scalar::util
did mean and median too
orien wrote part of this?
mmm tastey
overloads for stddev
vector is all set I think
order
the modifications to covariance caused test failures
I rewrote the tests a little, thinking about the new vector attribute support
I made huge changes to the vector object
I tested much of the new Vector stuff
so many changes...
moving toward that beautiful-er arrangement
k, mean and variance are both awesome
did the same cool things to stddev and variance
tested the new cool things also
fixed the weaken() code
did the first-col stuff to median and mode
we need the computers in covariance to use our new cool style, but clearly the computer for variance depends on which other vector it is paired with...
described the last commit
renamed CoVariance to Covariance
typo
the tests were wrong since we fix_size on the unshift end now
forgot to add this get_linked ...
taught covariance to return both sizes
started adapting Correlation to the new layout
typo
I lost a couple lines from 0.42 that really shouldn't have been lost. Perhaps the git import from CVS didn't go so well? Odd.
I updated LSF with the latest layout stuff
better and more shortcuts
fixed shortcut exports
this is better I think
tested a bunch more stuff in t/07
this has come quite a ways, but there's a bunch of test problems in t/53 co
more tests (that fail)
more debugging messages
more verbose debugging messages
whitespace
more debugging and changes that make no difference at all
I added all kinds of cool debugging before I found the really stupid bug (next commit)
grrz, what a stupid bug
completed the tests
fixed a bunch of problems with testing while debugging is on
completed many todos and added a couple others
new version and merged changelogs
I think the basic/Basic pod is complete-ish
documented those latest changes
couple basic changes
minor doc changes
this documentation is barely acceptable, but better, somehow, than the original
typo
did the doc for correlation
topos
typo
wrote up a nice Cov pod
these are empty
while documenting variance (done) I noticed typos in these others
did stddev
typos
did mode
typos and docs
documented vectors
this is fine... my bad
paranoid I guess
forgot some stuff
from the docs
cranked out the insert/set stuff
hammered out a couple more todos
noted the docs
I found a bunch of typos already, starting the proofreading over from the top.
more doc fixes and the new x_given_y and y_given_x functions
more whitespace
little doc fixes
moved a bunch of shit around
grrz
doc fix
typo
doc fixes
added a readme
isa() is smarter than ref() because ref() may not match a correct isa()
cvec is happening!!
this will work shortly I hope
new version
documented the 1.5 version
self defined computers is better than a new class
awesome, the filtered vectors rock
lawl, the new filters rock
missed the new pod
whitepsace
I did some cleanup on the recalc code, but still can't account for the vile test failure in t/08_filter*
fixed median bug
forgot to version up, as usual
n::f and abs(-1) patches
this is pretty hot
this will work more reliably now that the tolerances are installed I imagine, but we need to actually test it.
when there are math::bigthings in the vectors, overload freaks out with the lhs in s::b and the rhs in m::bt
talked about bigint in the changelog
failed to remove debugging note
these dot ohs were for bignum (didn't help)
this test is hard to get working
moved the exists() to compile time
installed a typo
I put the 'eq' function back
wow, cut and paste accidents
moved tollerance before compile of use S::B
described this new compile time requirement
whitespace, non math::, etc
insert notification got promoted to debug=1
I was really thinking about the blessed() check incorrectly
this is releasable... I'm going to wait to see if I hear back from William R. Ward first though
forgot to manifest the t/60 M::BF test
reply to rating
better note
fix the t/60 failures?
require NF 1.61
version mismatch
this didn't help, but I didn't want to throw out the code...
bye bye
1.6003
bug found by Raj Chandran <rchandran at nextrials.com>
little cleanup
meta yml stuff
fixed manifest bugs
typo
meta.yml patches
fixed copyright information by request
docbug
never finished my thought...
documented docbugs
5.8.1 is actually required for this module... might work in earlier perls, but *NOT* 5.8.0
this test is going to be awesome
instead of freaking out when there are undefs, print them with sanity
ok, so these are highly correlated even though there's missin data...
new help function named
document new feature(s)
call it what people actually call it
version bump
add best practices tests to the dist
note additional tests
bump version for new shit
renamed all the recalcs
lose the warnings for early stringifications
made ginsert grow instead of setting the size
Made a ton of doc changes to satisfy pod_coverage.t (plus some
if I can DRY Correlation, then Covariance is next.
rename twovector package
use eval so errors burp up in the caller
it was harder than expected to make these truly DRY by using a baseclass, but they seem to work now
renamed base class
DRY progress notes
pod for two vector base
manifest the new pod
move this stuff to the base m odule
decided to include this afterall
moved query to the _TVB
LSF using _TVB
doc bugs
doc bugs
built the better pod for cov
the Mean conversion to a base package went flawlessly
ws
too correlation-centric
I began working on the conversion of Mean to _OVB, and noticed tons of POD issues
add norm
change this language, which is wrong... combe back in a bit
renaming normalize ... nofil
query_filled
query_filled test
document the fill disabler (changes)
facilitate using vectors that can be less filled than their max size and respond with undef when used
more query_filled stuff
had a lotta stuff tracking that I didn't really need to track
note the removal of the functions from LSF
FILL and NOFILL
return objects from accessors
return objects from accessors
this test doesn't work at all
manifest the moving average test
moving average test
eh
_OVBed Mode
converted median to _ovb
fixed some debugging bugs
_OVBed: StdDev, Variance
bugs introduced by base-classing
(full details)--
worked on the pods quite a lot
a lot of these shall not be documented and threfore, should be marked private
note new changes
started fixing the privates
for consistancy, size() should be query_size(), jesus
size -> query_size
note changes
epic fail at renaming these previously
finish fixing the privates (?)
working on the pods a bit
these pods are getting to be a nightmare, what can ya do
more pod stuff
more pods
more renamings
made another accessor return the object
changed ginsert to append in the docs, and changed a couple things
note changes
failed to bring my critic config over from Net::IMAP::Simple
failed to mainfest the rc
altered whitespace
ginsert vs append
trying to make things really consistent
completed the docs for vector
I *just* noticed leftover pod_coverage trustmes from PJF, hehe
typo
might be re-designing computed vectors very slightly so = computed(1,2,3)->insert(4) will work
return the object for these accessors
I screwed this up I guess
append
ws
I actually made all these do something though, so this was academic
altered whitespace
Superfluous
fixed copy
note changes
note to self
taught this to understand set_vector from computed vectors
taught this that it's wrong to do what I taught it last
noted autofill stuff
did some of the computed() docs
typo
make the dates a lot less specific... I've been ignoring the times anyway
this stuff is hard to keep track of
extra space
note that you can affect the contents, but shouldn't
use stronger language to discourage changing the ref
copy vectors instead of linking them. Computed vectors can serve this role, and it makes little sense to do this with regular vectors
note changes
this should always work now, or it's probably a bug
make accessors return the object
this stuff all checks out
note changes
Basic.pm [critic.t]
ComputedVector.pm [critic.t]
Covariance.pm [critic.t]
tons more [critic.t] changes
end of [critic.t] changes
continue making sure query refs stay the same under various transforms
since we copy instead of link, I think it *will* do what we want now
test the set_vector vs query_vector (ref) stuff on computed vectors
I think these docs are correct and complete
todo
note
note append vs ginsert
allow set_vector(1,2,3,4)
all these objects should be true if asked
take nonscalar arguments that look like numbers
use the nonscalar set_vector and the new bool overloads
note changes
more places to use the new bool context overload
made this pass tests using the new set_vector
strip useless spam
major bug, introduced in the set_vector(@_) epoch
no uninitizlied warnings. undefs are allowed.
there's no real reason the elements have to be numbers
last night, I re-did this test a little so I could see what was coming back instead of a boring fail
a new test to find warnings and size failures on construction
note changes
make these more descriptive so its easier to see which object is fail (all so far)
fix initialization bugs
remove debug msg
fixed up the LSF docs to pass pod_coverage
k, these docs are ready for public consumption I think.
note about new
new test that shows oddball warnings
less vertical when applicable
less vertical when applicable
add tags to computed vectors so they work in the various objects that get linked computers for two vector computations
add query_vector
I think the CV is ready
use B<> instead of C<> for =items
made the overload txt really consistent
note chaining
TODO REFORMAT
debug typo
the pods are coming along, so tedious.
introducing rand() to tests makes the tests fail sometimes
use chaining
provided an interface to the ENV vars without having to interact
critic bug
trying to eliminate the ENV vars alltogether
note changes
this is annoyingly difficult
removed ENV crap
must load base modules later
use qw(toler) rather than ENV{TOLER}
leftover debug
rework tests to avoid ENV (mostly)
new critic problems ... shhhhh
that was a lot of documentation to change, but I think it's all set
fixed the rest of the docs too
pod todo updates
Mean seems ok
couple little things
did a few more pods...
did a couple more pods... tedious
pods, tedious
I'm tired of updating this
this should be all of it
pod bugs
note changes
English bugs
little doc changes for readability
typo
additional notes
lists, not arrays ... probably
doc change
make the line of fit way more easy to understand and deal with
handle_missing alias
one more test
allow 5.6.x, not that it helps
ignore this stuff (updates)
don't set TEST_AUTHOR, when you're not the author, and then complain to me about the problems you have.
grrz
I happen to like it. Until perl itself removes this, why should I?
stfu
fixed actual problem
resurrect 5.6.1
https://rt.cpan.org/Ticket/Display.html?id=65065
https://rt.cpan.org/Ticket/Display.html?id=65065 (continued)
version bump
updated this so people know which version.
ignore this stuff (updates)
minor tweak
it's 2012 now?
update copyright and add note about Statistics::Descriptive
version up for re-release
bug
version
LC_ALL=C
Revert "LC_ALL=C"
locale hack to pass tests that assume 22/7 is 3.142 and not 3,142
ignore this stuff (updates)
ahh, I think this is what we want
Merge branch 'RT100943' of https://github.com/eserte/statistics--basic
oops
Salvatore Bonaccorso (1):
Imported Upstream version 1.6611
Slaven Rezic (1):
a workaround for RT #100943
jettero (4):
Initial revision
ok
ok
ok
-----------------------------------------------------------------------
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/libstatistics-basic-perl.git
More information about the Pkg-perl-cvs-commits
mailing list