Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Add warning about recvLineInto being bound to change | def | 2015-04-26 | 1 | -0/+3 |
| | |||||
* | Revert "Introduce FutureVar[T] to make recvLineInto safer." | def | 2015-04-26 | 3 | -74/+20 |
| | | | | This reverts commit 72b4912c84b16644657f94e54105739cba4b2457. | ||||
* | fixes regression: overloading by 'var' | Araq | 2015-04-25 | 1 | -128/+128 |
| | |||||
* | marshalling can be done at compile-time | Araq | 2015-04-25 | 2 | -21/+57 |
| | |||||
* | fixes #2517 | Araq | 2015-04-25 | 5 | -1826/+59 |
| | |||||
* | Merge branch 'def--clean-speedup-2' into devel | Dominik Picheta | 2015-04-24 | 5 | -96/+317 |
|\ | |||||
| * | Fix asyncdispatch on Windows. | Dominik Picheta | 2015-04-24 | 2 | -112/+111 |
| | | |||||
| * | Introduce FutureVar[T] to make recvLineInto safer. | Dominik Picheta | 2015-04-24 | 3 | -20/+74 |
| | | | | | | | | | | | | | | | | | | | | | | FutureVar[T] is a new distinct Future type which is designed to be used for situations where the highest performance is needed. It reduces the number of Future allocations needed. It acts as a replacement for 'var' params in async procs. This commit modifies @def-'s PR in order to make it safer. The recvLineInto procedure has been modified to take a ``FutureVar[string]`` param instead of a ``ptr string`` param. | ||||
| * | Some small cleanup. | Dominik Picheta | 2015-04-23 | 2 | -8/+7 |
| | | |||||
| * | Merge branch 'clean-speedup-2' of https://github.com/def-/Nim into ↵ | Dominik Picheta | 2015-04-23 | 4 | -99/+268 |
| |\ | | | | | | | | | | def--clean-speedup-2 | ||||
| | * | Rename clearStringTable to clear | def | 2015-03-17 | 2 | -3/+3 |
| | | | |||||
| | * | Some style cleanup | def | 2015-03-17 | 1 | -3/+3 |
| | | | |||||
| | * | Make respond a template again | def | 2015-03-17 | 1 | -11/+11 |
| | | | |||||
| | * | Move recvInto to asyncnet and don't export it | def | 2015-03-17 | 2 | -120/+120 |
| | | | |||||
| | * | Don't export readInto* templates | def | 2015-03-17 | 1 | -2/+2 |
| | | | |||||
| | * | Make recvLineInto a proc instead of template | def | 2015-03-17 | 2 | -52/+50 |
| | | | |||||
| | * | Rename resetStringTable to clearStringTable | def | 2015-03-17 | 2 | -2/+2 |
| | | | |||||
| | * | Remove unused newRequest proc | def | 2015-03-17 | 1 | -3/+0 |
| | | | |||||
| | * | Document asynchttpserver's respond template | def | 2015-03-17 | 1 | -1/+1 |
| | | | |||||
| | * | PNimrodNode -> NimNode in asyncdispatch | def | 2015-03-17 | 1 | -3/+3 |
| | | | |||||
| | * | Speed up asynchttpserver significantly using all the previous changes | def | 2015-03-17 | 2 | -47/+53 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Export socket field of AsyncHttpServer and addHeaders proc for templates - Make respond a template instead of proc because of how often it's called. This means no more "await" when invoking it. - Optimize respond template with special case for empty headers and Content-Length entry - newRequest doesn't allocate a hostname and body anymore because they're copied in later - Major changes to processClient to prevent allocations and copies | ||||
| | * | Make asyncnet usable when avoiding allocations. | def | 2015-03-17 | 1 | -61/+98 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - readInto, readIntoBuf, are templates instead of procs now - New recvLineInto template that reads directly into a string instead of creating a new one. Used by recvLine proc now - Need fd and bufLen fields of AsyncSocketDesc exported because of the templates - recv returns a shallow string to prevent copying - This gives significant speedups, mostly by using templates instead of creating new Futures and waiting for them all the time. | ||||
| | * | Make asyncdispatch usable when preventing allocations | def | 2015-03-17 | 1 | -0/+112 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Added a recvInto proc that takes a cstring as argument and reads into it instead of returning a newly allocated string. This is pretty unnice because of code duplication with recv. Calling recvInto from recv is not a good solution because of the additional future that gets created. - Windows version is totally untested | ||||
| | * | Make strtabs module usable when avoiding allocations | def | 2015-03-17 | 1 | -0/+6 |
| | | | | | | | | | | | | - resetStringTable proc | ||||
| | * | Make uri module usable for faster URI parsing: | def | 2015-03-17 | 1 | -10/+23 |
| | | | | | | | | | | | | | | | | | | - A version of parseUri that takes a uri as parameter and modifies it - Export initUri so you can use the new parseUri better - Avoid creating new strings | ||||
* | | | Fix Termios wrapper | def | 2015-04-24 | 1 | -6/+7 |
|/ / | |||||
* | | Turn some test outputs into actual tests | Oleh Prypin | 2015-04-21 | 12 | -49/+80 |
| | | |||||
* | | Fix romans module | Oleh Prypin | 2015-04-21 | 1 | -6/+3 |
| | | |||||
* | | Fix a test in xmltree | Oleh Prypin | 2015-04-21 | 1 | -1/+2 |
| | | |||||
* | | Don't run non-test code when defined(testing) | Oleh Prypin | 2015-04-21 | 41 | -77/+99 |
| | | |||||
* | | Merge pull request #2411 from reactormonk/merge-counttables | Andreas Rumpf | 2015-04-21 | 1 | -0/+61 |
|\ \ | | | | | | | merge for CountTable, see #1680 | ||||
| * | | doc comments for merge | Simon Hafner | 2015-03-26 | 1 | -0/+3 |
| | | | |||||
| * | | merge for CountTable, see #1680 | Simon Hafner | 2015-03-26 | 1 | -0/+58 |
| | | | |||||
* | | | Merge pull request #2572 from maxgrenderjones/fix-2431-against-devel | Andreas Rumpf | 2015-04-21 | 1 | -16/+27 |
|\ \ \ | | | | | | | | | {} returns nested results in json & standardise on 'key' not 'name' | ||||
| * | | | Allow {} to return nested results in json & standardise on 'key' not 'name' | Max Grender-Jones | 2015-04-18 | 1 | -16/+27 |
| | | | | | | | | | | | | | | | | | | | | | | | | - Note now returns nil if users try to index into a non-object; old behaviour was to throw a runtime error - Also s/transverse/traverse/ | ||||
* | | | | Merge branch 'devel' of https://github.com/Araq/Nim into devel | Araq | 2015-04-20 | 1 | -7/+7 |
|\| | | | |||||
| * | | | Make peekExitCode on Posix use WIFEXITED to determine exit | def | 2015-04-14 | 1 | -7/+7 |
| | | | | | | | | | | | | | | | | This fixes #2552 | ||||
* | | | | fixes #2568 | Araq | 2015-04-17 | 1 | -2/+2 |
|/ / / | |||||
* | | | Merge pull request #2493 from arnetheduck/more-comp-ropes | Andreas Rumpf | 2015-04-10 | 1 | -63/+80 |
|\ \ \ | | | | | | | | | More comp ropes | ||||
| * | | | lib_ropes: better buffering in equalsFile, cleanups | Jacek Sieka | 2015-04-08 | 1 | -63/+80 |
| | | | | |||||
* | | | | Remove unneeded import os from intsets | Oleh Prypin | 2015-04-09 | 1 | -1/+1 |
| | | | | |||||
* | | | | Merge pull request #2484 from yglukhov/macos-sigpipe | Andreas Rumpf | 2015-04-08 | 2 | -5/+7 |
|\ \ \ \ | |/ / / |/| | | | Fixed SIGPIPE on MacOS | ||||
| * | | | Fixed SIGPIPE on MacOS | yglukhov | 2015-04-08 | 2 | -5/+7 |
| | | | | |||||
* | | | | Merge pull request #2486 from endragor/rational-hash | Andreas Rumpf | 2015-04-08 | 1 | -0/+14 |
|\ \ \ \ | | | | | | | | | | | Add hash proc for Rational | ||||
| * | | | | Added hash proc for Rational | Ruslan Mustakov | 2015-04-08 | 1 | -0/+14 |
| | | | | | |||||
* | | | | | Added procvar pragma to rationals.cmp | Ruslan Mustakov | 2015-04-08 | 1 | -1/+1 |
|/ / / / | |||||
* | | | | Merge pull request #2472 from def-/natural | Andreas Rumpf | 2015-04-07 | 6 | -1022/+1022 |
|\ \ \ \ | |/ / / |/| | | | Use more Natural and Positive numbers in proc parameters | ||||
| * | | | Use more Natural and Positive numbers in proc parameters | def | 2015-04-06 | 6 | -1022/+1022 |
| | | | | | | | | | | | | | | | | | | | | - Didn't go through all modules, only the main ones I thought of - Building the compiler and tests still work | ||||
* | | | | Fix toRational | def | 2015-04-07 | 1 | -2/+2 |
|/ / / | |||||
* | | | Merge pull request #2428 from arnetheduck/comp-lib-ropes | Andreas Rumpf | 2015-04-06 | 1 | -12/+13 |
|\ \ \ | | | | | | | | | Comp lib ropes |