improvements to get code size down for programs that don't use GC
This commit is contained in:
parent
c9d21164be
commit
e25384db89
18 changed files with 642 additions and 181 deletions
16
todo.txt
16
todo.txt
|
|
@ -2,7 +2,6 @@ High priority (version 0.8.12)
|
|||
==============================
|
||||
* test threads on windows; thread analysis needs to be even more restrictive!
|
||||
* implement message passing built-ins: channels/queues
|
||||
* outOfMem hook; 0 over-head for hello world program :-)
|
||||
* bug: {:}.toTable[int, string]()
|
||||
|
||||
|
||||
|
|
@ -20,18 +19,6 @@ version 0.9.0
|
|||
- fix overloading resolution
|
||||
- make ^ available as operator
|
||||
- implement closures; implement proper coroutines
|
||||
- rethink comment syntax; people want::
|
||||
|
||||
type
|
||||
TConfig* = object
|
||||
limit*: int
|
||||
outFile*: string
|
||||
# Regex patterns to ignore. TLogEntry field -> regex pattern
|
||||
# If this matches, the request will NOT be shown.
|
||||
ignoreRe*: TTable[string, string]
|
||||
|
||||
I think a good compromise is to define positions for the docgen in the
|
||||
grammar and ignore other comments everywhere.
|
||||
|
||||
Bugs
|
||||
----
|
||||
|
|
@ -45,8 +32,7 @@ version 0.9.XX
|
|||
==============
|
||||
|
||||
- distinct types for array/seq indexes
|
||||
- GC: marker procs for native Nimrod GC and Boehm GC
|
||||
- code concerning 'assert' is wasteful and unnecessarily complex
|
||||
- GC: marker procs for native Nimrod GC and Boehm GC; precise stack marking
|
||||
- implicit ref/ptr->var conversion; the compiler may store an object
|
||||
implicitly on the heap for write barrier efficiency
|
||||
- resizing of strings/sequences could take into account the memory that
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue