Andreas Rumpf
ef4b755183
allows a destructor to be attached to a tyString/tySequence
2018-07-27 18:20:13 +02:00
Andreas Rumpf
4ec91a30c4
allocators: add deallocAll proc pointer
2018-07-27 18:19:18 +02:00
Andreas Rumpf
f485ebe162
--gc:destructors: next steps; WIP
2018-07-21 13:16:53 +02:00
Andreas Rumpf
4389409e26
fix merge conflict
2018-07-18 09:46:30 +02:00
Andreas Rumpf
ae09879b3a
fixes #8223
2018-07-17 18:21:09 +02:00
Andreas Rumpf
59a5c0462f
fixes #8338
2018-07-17 15:49:55 +02:00
Andreas Rumpf
32afdc09c6
WIP: strings/seqs based on destructors
2018-07-17 13:19:42 +02:00
John Novak
692d1da95d
Fix setCursorPos and setCursorXPos on POSIX ( #8310 )
2018-07-17 11:35:26 +02:00
Jeff Ciesielski
f8723cd1e3
Add extended baud rates to posix/termios ( #8322 )
...
115200 is really common. I'd like to get support in nimserial, but
I'm guessing it would help to have support in the stdlib first.
2018-07-16 15:15:47 -04:00
Timothee Cour
2ac22b4cf0
fix #8273 times format regression, and fix inconsistent ordering in 1 format overload ( #8290 )
...
* Fix issue #8273 [regression] [times.format] Error: attempting to call
undeclared routine: 'format'
* rename format to f for consistency with other overloads and avoid
similar bugs as #8273
* breaking change since PR 8094: changed format*(f: TimeFormat, dt: DateTime) to format*(dt: DateTime, f: TimeFormat) for consistency w other overloads
* use consistent ordering for times.parse procs
2018-07-16 21:15:17 +02:00
Andreas Rumpf
74bf316619
more progress on destructor based strings
2018-07-13 21:15:47 +02:00
momf
dd47013017
add SSL_set_SSL_CTX for SNI(Server Name Indication) ( #8308 )
...
* add SSL_set_SSL_CTX for SNI
* fix SSL_set_SSL_CTX
2018-07-13 17:09:22 +01:00
Yuriy Glukhov
dfe3f16022
Don't depend on string.h in codegen ( #8299 )
2018-07-13 17:41:59 +02:00
Timothee Cour
40f44a0c16
fix issue #8251 ospaths.isAbsolute: out of bound errors ( #8291 )
...
* fix issue #8251 ospaths.isAbsolute: out of bound errors
* address comments
* add reference to a spec for quirky macos paths
2018-07-13 03:54:48 -04:00
LemonBoy
3163a0f466
Do not consider enums with holes as ordinals ( #8264 )
...
Make the compiler behave consistently with respect to what's written in
the manual.
Fixes #1239
2018-07-12 18:08:45 +02:00
Eugene Kabanov
3fec2ba5e5
Fix terminal.nim to be GCSAFE. ( #8296 )
...
* Also fix deprecated callsite().
2018-07-12 18:06:15 +02:00
Joey Yakimowich-Payne
9aad193d27
Fix undefined PObject reference
2018-07-12 11:54:25 +01:00
Kaushal Modi
1102f9aaf1
Make the Style enum a proper Ordinal (no holes) ( #8282 )
...
The Style enum needs to be an Ordinal so that it can be used in sets.
2018-07-12 11:30:51 +02:00
Quelklef
ac3c4a94ad
Fixed $ on None[T] for T with .name ( #8293 )
2018-07-12 11:01:48 +02:00
Andreas Rumpf
5b59852406
system.substr is not implemented with compilerProcs anymore
2018-07-11 16:39:16 +02:00
Andreas Rumpf
16d8fab310
mmdisp: code cleanups
2018-07-11 16:38:13 +02:00
Timothee Cour
25bf0d1683
add os.absolutePath; fixes #8174 ( #8213 )
...
* add os.absolutePath
* fixup
* fixup
* Fixes absolutePath error message.
2018-07-10 20:31:13 +01:00
Andreas Rumpf
8f0c9b2fdd
string comparisons don't have to know the strings representation
2018-07-10 11:37:30 +02:00
Andreas Rumpf
ce01472ff3
refactorings in preparations for the new runtime
2018-07-10 10:39:53 +02:00
Dominik Picheta
f805018461
Fixes #5880 . ( #7229 )
2018-07-10 00:33:31 +02:00
Oscar Nihlgård
3b310e91cd
New implementations of times.parse & times.format ( #8094 )
2018-07-09 20:04:25 +02:00
cooldome
dec97924a4
Custom pragmas in proc types ( #8205 )
2018-07-09 12:11:03 +02:00
genotrance
c115090f6e
Testament should run tests with no action ( #8232 )
...
* Testament should run tests with no action
* Fix strutils tests for JS
2018-07-08 20:38:46 -04:00
Timothee Cour
befca425c4
workaround issue #8223 nim doc fails with doAssertRaises ( #8245 )
2018-07-08 08:16:13 +02:00
Dmitry Atamanov
a6c3bbf01a
Add proc toOpenArray[byte] for strings ( #7820 )
2018-07-07 21:03:22 +02:00
Mamy Ratsimbazafy
75ef90da82
Put back deprecation path for SomeReal
...
was wrongly deleted in c7298561c1
2018-07-07 15:04:41 +02:00
Dominik Picheta
53ce58f050
Merge pull request #6587 from FedericoCeratto/normalizePath
...
Add normalizePath and tests
2018-07-07 10:12:07 +01:00
Araq
8667899d53
hotfix: fixes possible out of bounds access in ospaths.joinPath
2018-07-07 02:31:43 +02:00
Federico Ceratto
d65429d857
Add normalizePath and tests
2018-07-06 20:21:39 +01:00
Araq
531013727d
system.nim: minor code style change
2018-07-06 19:13:54 +02:00
Araq
5384de685f
Merge branch 'devel' of github.com:nim-lang/Nim into araq-devel
2018-07-06 18:41:00 +02:00
Dominik Picheta
5e338c371e
Use '\L' instead of '\n' in system for backwards compatibility.
...
See https://github.com/nim-lang/packages/pull/785 .
2018-07-06 15:48:17 +01:00
Araq
8bcaee1fdf
make tmemfile2 work again
2018-07-06 15:53:32 +02:00
Joey
a1457bfa9f
Rewrite the memory management code for Nintendo Switch ( #8169 )
...
Rewrite the memory management code for Nintendo Switch
The first implementation was naive and did not account for multiple
memory allocations. However, this implementation may still be
incomplete. Currently, when running applications, the code runs fine.
When the application is exited via code (the end of the program is
reached or quit() is called), the Switch will crash. Not sure why this
happens, but I suspect it is from Nim memory allocations.
I suspect the memory allocations because when I compile the helloworld
application without any Nim allocations (just C function calls) and use
`--gc:none` as a compile option, the application exits fine.
2018-07-05 16:33:15 +02:00
Andreas Rumpf
683c60e2fd
restore deprecated PNimrodNode in order to make the CIs happy
2018-07-05 09:17:44 +02:00
Andreas Rumpf
86a7054c88
Merge branch 'devel' into araq-devel
2018-07-04 19:25:53 +02:00
Timothee Cour
426e5c2d1f
simplify typetraits.nim test ( #8193 )
2018-07-03 12:19:46 +02:00
Vindaar
681d8e0749
Deprecate 'c', 'C' prefix for octal literals, fixes #8082 ( #8178 )
...
* deprecate `0c`, `0C` prefix for octal literals
Deprecates the previously allowed syntax of `0c` and `0C` as a prefix for
octal literals to bring the literals in line with the behavior of
`parseOct` from parseutils.
* add `msgKind` arg to `lexMessageLitNum` for deprecation messages
* change literal tests to check all valid integer literals
Also adds the `tinvaligintegerlit3` test to test for the (still)
invalid `0O` prefix.
2018-07-03 01:56:36 +02:00
zargot
e63369f2b9
asyncftpclient: add missing file.close to retrFile
2018-07-02 18:07:24 +02:00
Tomohiro
c79f488027
Fix os.unixToNativePath proc returns wrong result( #8179 ) ( #8181 )
...
* Fix os.unixToNativePath proc returns wrong result(#8179 )
* Add tests for unixToNativePath
2018-07-02 17:14:26 +02:00
Tomohiro
a2b2bc1a36
Fixed #8173 ( #8175 )
2018-07-02 00:46:56 +02:00
Andreas Rumpf
e455023902
Merge branch 'devel' into araq-devel
2018-07-01 16:24:56 +02:00
Andreas Rumpf
67c7e49e2a
gc_common: minor reformating
2018-07-01 15:10:18 +02:00
Oscar Nihlgård
b61e69202b
Unexport times.stringifyUnit ( #8168 )
2018-07-01 12:54:13 +02:00
Timothee Cour
fbd9f9da95
execCmdEx stderr redirection ( #8162 )
2018-07-01 04:38:15 -04:00