about summary refs log tree commit diff stats
path: root/026call.cc
Commit message (Collapse)AuthorAgeFilesLines
* 4148Kartik K. Agaram2017-12-071-2/+6
|
* 4136Kartik K. Agaram2017-12-041-2/+4
|
* 4125Kartik K. Agaram2017-11-191-0/+15
|
* 4104Kartik K. Agaram2017-11-031-3/+3
| | | | | Stop hardcoding Max_depth everywhere; we had a default value for a reason but then we forgot all about it.
* 4099Kartik K. Agaram2017-11-011-0/+47
| | | | | | | | | | | | | | Generalize commit 4089 to arbitrary closures, and not just the current 'space' or call frame. Now we should be treating spaces just like any other data structure, and reclaiming all addresses inside them when we need to. The cost: all spaces must now specify what recipe generated them (so they know how to interpret the array of locations) using the /names property. We can probably make this ergonomic with a little 'type inference'. But at least things are safe now.
* 3898Kartik K. Agaram2017-05-311-34/+14
| | | | | | | | | | | There seems to be some chance of speed-up when I inline these functions. However, it's not a clear enough signal to justify improving the `build` script to handle the `inline` keyword. Current timing experiments: Before After ubuntu 1GB 9:22,8:48,8:51,9:16,9:17,8:36,9:05 8:55,8:41,8:15,8:27,8:29,8:54,9:29 OS X 8GB 4:05,4:00,4:18,4:09,3:40,3:51,3:56 3:58,3:52,4:01,4:13,4:16,4:31,4:13
* 3889 - profile refcount updates by recipeKartik K. Agaram2017-05-291-3/+4
| | | | | We're currently spending 50% of our time in `mu edit` performing refcount updates.
* 3888 - beginnings of a profilerKartik K. Agaram2017-05-281-0/+5
| | | | Time to make my ad hoc commented out code fragments a first-class feature.
* 3887 - clean up early exits in interpreter loopKartik K. Agaram2017-05-281-14/+13
| | | | | | | | | | It's always confusing when `break` refers to a `switch` but `continue` refers to the loop around the `switch`. But we've done ugly things like this and `goto` for expedience. However, we're starting to run into cases where we now need to insert code at every `continue` or `continue`-mimicking `goto` inside the core interpreter loop. Better to make the loop single-entry-single-exit. Common things to run after every instruction will now happen inside the `finish_instruction` function rather than at the `finish_instruction` label.
* 3877Kartik K. Agaram2017-05-261-1/+1
|
* 3809Kartik K. Agaram2017-04-041-2/+2
|
* 3803Kartik K. Agaram2017-03-201-1/+1
|
* 3380Kartik K. Agaram2016-09-171-8/+8
| | | | | One more place we were missing expanding type abbreviations: inside container definitions.
* 3358Kartik K. Agaram2016-09-151-0/+5
|
* 3279Kartik K. Agaram2016-08-291-12/+12
| | | | | Stop inlining functions because that will complicate separate compilation. It also simplifies the code without impacting performance.
* 3120Kartik K. Agaram2016-07-211-1/+1
| | | | | | | | Always show instruction before any transforms in error messages. This is likely going to make some errors unclear because they *need* to show the original instruction. But if we don't have tests for those situations did they ever really work?
* 3105Stephen Malina2016-07-061-1/+1
| | | | Add number-of-instructions recipe to Mu
* 2862Kartik K. Agaram2016-04-241-0/+168
Layers 0-29 are now a complete rudimentary platform except for pointers and indirection.