67c44bcf62
base64: Fix base64 for strings that don't end in equal signs.
...
Some uses of base64 in the wild do not have trailing ==, such as JWT (JSON Web Token) strings.
2017-06-08 21:03:40 -05:00
57de1f5ab0
math.functions: make -0.3 truncate return 0.0 instead of -0.0.
2017-06-08 11:27:48 -07:00
1dceb069ad
formatting: fix bug when using %e and the number rounds up in magnitude
...
99/10 0 format-scientific-simple was outputting 1.0e0 because
the mantissa was rounded up, so it had one extra character, and
the exponent was wrong.
2017-06-08 11:23:38 -07:00
4df21818f5
formatting: change %.0f and %.0e to remove the radix character. It's more standard.
2017-06-08 11:23:38 -07:00
756184607b
formatting: make these tests actually test the expected code path
2017-06-08 11:23:38 -07:00
552791d1cf
formatting: don't force the presence of ".0" for %f and %e when precision is 0
...
It's misleading because you can think you have more precision than you really
do when looking at the output. So instead of "1.0", we can format it as "1" or
"1.". And instead of "1.0e+00" we can format "1e+00" or "1.e+00". In C,
printf does not print the radix character, it's prettier. But since the factor
parser accepts both styles, and keeping the radix character gives a stronger sense
that the number is approximate, I prefer to keep the radix character..
2017-06-08 11:23:38 -07:00
762b22e1a8
formatting-docs, document ' ' as an alternative to '+' for positive numbers sign prefixing
2017-06-08 11:23:38 -07:00
6061da92be
formatting, allow uppercase exponent for bin floats in base 2 and 8
2017-06-08 11:23:37 -07:00
ccbb8116fc
formatting, make %d, %o, %b work for ratios and floats
2017-06-08 11:23:37 -07:00
438e22fceb
formatting, fix %f and %e for ratios and integers
2017-06-08 11:23:37 -07:00
e1df9df735
formatting: fix missing '+' for %+e with negative exponent
2017-06-08 11:23:37 -07:00
80e9d7c115
move math.extras:round-to-even to math.functions to use in basis:formatting
...
Also add round-to-odd in case it is needed.
Also change float rounding to be independent of the current rounding mode
2017-06-08 11:23:37 -07:00
bf852cea6a
add basis.math.functions.integer-logs: exact integer logarithms
2017-06-08 11:22:40 -07:00
34f36a529e
math.functions, speed up truncate for floats
2017-06-08 11:22:40 -07:00
c74cfe6c54
math.functions: fix truncate and round to output -0.0 for negative floats near 0
2017-06-08 11:22:40 -07:00
271c7d9519
math.functions, fix round for ratios between -1/2 and -1
...
it was outputting 1 instead of -1
2017-06-08 11:22:40 -07:00
acf14a7bd1
io.directories.windows: add size slot to windows-directory-entry tuple
2017-06-05 16:47:01 -07:00
75d8c1e050
Add .gitattributes to enforce EOL policy for factor sources
2017-06-05 13:58:08 -07:00
6d355c998b
help.lint.spaces: new vocab
2017-06-05 13:16:07 -07:00
2ffa446f5b
*-docs: replace double spaces with single spaces
...
Code examples and other places where multiple spaces were used to align
text were left intact.
2017-06-05 13:16:07 -07:00
445b55b3e5
test runner: non-zero exit code in case of failed tests
...
This makes it easier to detect unit test failures from the
command line.
2017-06-05 13:13:09 -07:00
27423b78aa
classes.intersection: Fix typo in commented-out unit-test.
2017-06-04 20:03:02 -05:00
9aacb29667
unmaintained: New repo here: https://github.com/factor/factor-unmaintained
...
It's confusing to ripgrep through unmaintained/ and this eliminates all
incompatible/outdated code in the Factor repository.
Please contribute the worthy code back to Factor and remove it from factor-unmaintained.
2017-06-04 15:45:56 -05:00
9e76899dbd
tools.test: Make the flag public. Finish porting tester changes to fuzzer.
2017-06-04 14:47:35 -05:00
1708c8011f
classes.singleton: Update another test.
2017-06-04 12:57:38 -05:00
50fbf0457a
tools.deploy: Change unit-test to long-unit-test for deploy tests.
...
long-unit-tests are enabled by default.
2017-06-04 12:55:58 -05:00
be1175b3c5
tools.test: Add long-unit-test and refactor a bit.
2017-06-04 12:55:52 -05:00
06fd324379
singleton: Move definitions to top level in tests.
2017-06-04 10:41:26 -05:00
b540feb95d
Nmakefile: add the /largeaddressaware linker option for 32-bit builds
...
This somewhat mitigates issue #1828 by allowing more memory to be allocated
for the 32-bit apps on some systems. Has no effect on 32-bit Windows XP.
2017-06-04 00:11:35 -05:00
b3ba829601
cuda.devices: Utility word to print all cuda devices.
2017-06-03 20:59:51 -05:00
acffb259f7
kernel: Fix the docs for curry/compose a bit.
2017-06-03 02:10:06 -05:00
72eaa9d466
compiler.tree.propagation.call-effect: I deleted an important method on compose on accident.
2017-06-02 23:31:35 -05:00
7a703658f1
sequences: Remove start for real in favor of subseq-start.
...
Update a usage of start that was masked by the ALIAS:
Compatibility/migration is a .99 feature.
2017-06-02 17:41:10 -05:00
48344ab55c
factor: last iota -> <iota> fix.
2017-06-02 17:39:40 -05:00
32710df620
core/basis: Rename tuples kernel:curry -> curried, kernel:compose -> composed.
...
Also rename the stack-checker curried -> curried-effect, composed -> composed-effect.
2017-06-02 17:39:20 -05:00
1dc65c13e7
words: Remove dummy ``word`` word and comment. Totally bogus.
2017-06-02 16:58:14 -05:00
195630f637
iota: Fix second iota -> <iota> in line several times.
2017-06-02 15:41:51 -05:00
e189954ecc
core: Rename iota to <iota> so we can have TUPLE: iota ... ; instead of TUPLE: iota-tuple ... ;
2017-06-02 11:57:05 -05:00
0ca1b4656f
ui.backend.cocoa.views: Use ?-> to avoid objc lookup errors if backingScaleFactor selector does not exist.
2017-06-01 16:53:10 -05:00
d838f95370
cocoa: Allow ?-> syntax for methods that might not exist.
...
If a method doesn't exist we need to provide a signature so the stack is balanced. This should also allow deploying from macOS versions that do not contain methods to ones that do. This is an alternative to asking a class if it provides a selector.
2017-06-01 16:53:05 -05:00
f7ce73b962
gml.runtime: Don't let gml double up on class/word names.
...
Related to #358 .
2017-06-01 15:47:49 -05:00
88bff3a034
sequences: Add start as an alias for subseq-start, start* aliasing subseq-start-from.
2017-06-01 14:55:15 -05:00
42a7a8e572
compiler: Rename start -> start-compilation, finish -> finish-compilation.
...
These start/finish are cute names, but ``finish`` is pretty vague even in the same file in the ``deoptimize-with`` word.
2017-06-01 14:51:18 -05:00
c5b563ed09
factor: Rename start -> subseq-start, start* -> subseq-start-from.
...
I also removed aliases for start/end from interval-maps and interval-sets and added a comment. I don't think it's any less clear what's going on.
2017-06-01 14:45:54 -05:00
c706bdae4e
factor: Fix misc/ for GENERIC# -> GENERIC#: rename.
2017-06-01 14:05:15 -05:00
7c7bb93c55
factor: Rename GENERIC# to GENERIC#:.
...
Fixes #1670 .
2017-06-01 13:58:58 -05:00
5aa89e6251
factor: Add newlines to every line.
2017-06-01 13:46:32 -05:00
df189eada5
io.files.info.windows: Fix using list.
...
We are using fry twice and it shows up in bootstrap.
2017-05-31 18:40:08 -05:00
e93d8f82bc
core: Add TH{ for making assoc tuples.
...
See #1449 .
https://github.com/factor/factor/issues/1449
2017-05-30 20:14:29 -05:00
f3bbc30c8a
namespaces: Way better implementation of with-variable-on.
2017-05-30 19:58:19 -05:00
Doug Coleman
John Benediktsson
Jon Harper
Alexander Iljin
Robert Vollmert