Age | Commit message (Collapse) | Author |
|
Also updated libraries in makefiles that were deprecated.
|
|
|
|
As we no longer have the FILE type, we need to buffer our reading and
writing so that we don't have to make so many syscalls. The API is nice
so that we can buffer other readers. We will update it so that it eats
io·Readers/io·Writers.
|
|
We use weak linking to ensure we clean up at exit time correctly. If
libc is linked, then we call our cleanup function by registering an
atexit callback with the library. If libc is not linked, we have a weak
symbol that results in a noop. Similarly, if we call rt·exit while
linked with libc, this immediately calls libc's exit (which will call
our cleanup as we registered it). If we are not linked to libc, exit()
is given as a weak link to a noop function.
|
|
|
|
I found the split to be arbitrary. Better to include the functionality
in the standard library. I also split the headers to allow for more
granular inclusion (but the library is still monolithic). The only
ugliness is the circular dependency introduced with libutf's generated
functions. We put explicit prereqs with the necessary object files
instead.
|
|
This is a large change. In order to remove myself from libc's arcane
interface, I implemented an independent runtime layer. It is based on
musl's wonderful implementation mostly. Critically, if libc is linked to
the program, then we cooperate. Namely, we call start main and let libc
do all initialization. If not, then we have a noop defined in rt3.a.
The general structure of the file is:
1. sys/$os/$arch contains all architecture dependent code
2. sys/$os/port contains all code that depends on the os, but is
portable
3. rt/$arch contains all the runtime architecture dependent code
4. rt/* contains the portable runtime code.
Obviously testing is needed. Specifically, while code is checked in for
the most popular architectures, it only has been tested on one computer!
Overall this is exciting and as been educational.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Heredocs are simply strings written to tmp files.
There was minimal bug testing here.
Also, various bug fixes are included
|
|
|
|
However, it is not tested in the slightest.
|
|
Introduced shell globbing
|
|
|
|
This is very much a work in progress. Still ruminating on the structure
of the library. It feels right but I want a more "social" presence -
namely the ability to link to a libc seemlessly. The solution is most
likely weak aliasing that musl uses - but musl itself weak aliases
global symbols, e.g malloc. We would have to define weak internal
symbols that musl defines as strong links but this knows too much about
the internals of musl...
|
|
Unsure about my modification to the language. I found the parsing of the
case body within switches to be odd - specifically that it parses
liberally and then checks that it has case -> cmd structuring while it
walks the code. This means the language is more permissive than the
semantics. I modified it to be more explicit, but at the cost of having
to end each case statement with a semicolon. I wanted a colon, but this
is a valid word character and thus will be lexed as part of the word.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Updated our assumptions of readline to handle valid unicode input.
This required integrating with an explicit library to handle unicode
knowledge.
|
|
|
|
|
|
|
|
|
|
The readline functionality operated on the assumption that 1 byte = 1
character. This is obviously wrong if you input a non-ascii character.
This commit temporarily removes a lot of functionality but parses input
bytes in a unicode-aware manner.
The outstanding problem now is 1 unicode rune != 1 column. There are
double wide characters, as well as zero width runes, that further break
our assumption that 1 rune = 1 character = 1 column. This is the next
iteration.
|
|
I was hiding too many important constants. This commit moves them to the
main exported header.
|
|
Additionally, decode can now apply backwards on a byte string.
|
|
|