std: Implement CString-related RFCs

This commit is an implementation of [RFC 592][r592] and [RFC 840][r840]. These
two RFCs tweak the behavior of `CString` and add a new `CStr` unsized slice type
to the module.

[r592]: https://github.com/rust-lang/rfcs/blob/master/text/0592-c-str-deref.md
[r840]: https://github.com/rust-lang/rfcs/blob/master/text/0840-no-panic-in-c-string.md

The new `CStr` type is only constructable via two methods:

1. By `deref`'ing from a `CString`
2. Unsafely via `CStr::from_ptr`

The purpose of `CStr` is to be an unsized type which is a thin pointer to a
`libc::c_char` (currently it is a fat pointer slice due to implementation
limitations). Strings from C can be safely represented with a `CStr` and an
appropriate lifetime as well. Consumers of `&CString` should now consume `&CStr`
instead to allow producers to pass in C-originating strings instead of just
Rust-allocated strings.

A new constructor was added to `CString`, `new`, which takes `T: IntoBytes`
instead of separate `from_slice` and `from_vec` methods (both have been
deprecated in favor of `new`). The `new` method returns a `Result` instead of
panicking.  The error variant contains the relevant information about where the
error happened and bytes (if present). Conversions are provided to the
`io::Error` and `old_io::IoError` types via the `FromError` trait which
translate to `InvalidInput`.

This is a breaking change due to the modification of existing `#[unstable]` APIs
and new deprecation, and more detailed information can be found in the two RFCs.
Notable breakage includes:

* All construction of `CString` now needs to use `new` and handle the outgoing
  `Result`.
* Usage of `CString` as a byte slice now explicitly needs a `.as_bytes()` call.
* The `as_slice*` methods have been removed in favor of just having the
  `as_bytes*` methods.

Closes rust-lang#22469
Closes rust-lang#22470
[breaking-change]

Author: alexcrichton

src/doc/trpl/ffi.md

@@ -435,8 +435,8 @@ extern {
 }
 
 fn main() {
-    let prompt = CString::from_slice(b"[my-awesome-shell] $");
-    unsafe { 
+    let prompt = CString::new("[my-awesome-shell] $").unwrap();
+    unsafe {
         rl_prompt = prompt.as_ptr();
 
         println!("{:?}", rl_prompt);

src/librustc/metadata/loader.rs

@@ -744,7 +744,7 @@ fn get_metadata_section_imp(is_osx: bool, filename: &Path) -> Result<MetadataBlo
         }
     }
     unsafe {
-        let buf = CString::from_slice(filename.as_vec());
+        let buf = CString::new(filename.as_vec()).unwrap();
         let mb = llvm::LLVMRustCreateMemoryBufferWithContentsOfFile(buf.as_ptr());
         if mb as int == 0 {
             return Err(format!("error reading library: '{}'",

src/librustc_llvm/archive_ro.rs

@@ -30,7 +30,7 @@ impl ArchiveRO {
     /// raised.
     pub fn open(dst: &Path) -> Option<ArchiveRO> {
         unsafe {
-            let s = CString::from_slice(dst.as_vec());
+            let s = CString::new(dst.as_vec()).unwrap();
             let ar = ::LLVMRustOpenArchive(s.as_ptr());
             if ar.is_null() {
                 None
@@ -44,7 +44,7 @@ impl ArchiveRO {
     pub fn read<'a>(&'a self, file: &str) -> Option<&'a [u8]> {
         unsafe {
             let mut size = 0 as libc::size_t;
-            let file = CString::from_slice(file.as_bytes());
+            let file = CString::new(file).unwrap();
             let ptr = ::LLVMRustArchiveReadSection(self.ptr, file.as_ptr(),
                                                    &mut size);
             if ptr.is_null() {

src/librustc_llvm/lib.rs

@@ -2149,7 +2149,7 @@ impl Drop for TargetData {
 }
 
 pub fn mk_target_data(string_rep: &str) -> TargetData {
-    let string_rep = CString::from_slice(string_rep.as_bytes());
+    let string_rep = CString::new(string_rep).unwrap();
     TargetData {
         lltd: unsafe { LLVMCreateTargetData(string_rep.as_ptr()) }
     }

src/librustc_trans/back/lto.rs

@@ -140,7 +140,7 @@ pub fn run(sess: &session::Session, llmod: ModuleRef,
 
     // Internalize everything but the reachable symbols of the current module
     let cstrs: Vec<CString> = reachable.iter().map(|s| {
-        CString::from_slice(s.as_bytes())
+        CString::new(s.clone()).unwrap()
     }).collect();
     let arr: Vec<*const libc::c_char> = cstrs.iter().map(|c| c.as_ptr()).collect();
     let ptr = arr.as_ptr();

src/librustc_trans/back/write.rs

@@ -22,7 +22,7 @@ use syntax::codemap;
 use syntax::diagnostic;
 use syntax::diagnostic::{Emitter, Handler, Level, mk_handler};
 
-use std::ffi::{self, CString};
+use std::ffi::{CStr, CString};
 use std::old_io::Command;
 use std::old_io::fs;
 use std::iter::Unfold;
@@ -49,7 +49,7 @@ pub fn llvm_err(handler: &diagnostic::Handler, msg: String) -> ! {
         if cstr == ptr::null() {
             handler.fatal(&msg[]);
         } else {
-            let err = ffi::c_str_to_bytes(&cstr);
+            let err = CStr::from_ptr(cstr).to_bytes();
             let err = String::from_utf8_lossy(err).to_string();
             libc::free(cstr as *mut _);
             handler.fatal(&format!("{}: {}",
@@ -67,7 +67,7 @@ pub fn write_output_file(
         output: &Path,
         file_type: llvm::FileType) {
     unsafe {
-        let output_c = CString::from_slice(output.as_vec());
+        let output_c = CString::new(output.as_vec()).unwrap();
         let result = llvm::LLVMRustWriteOutputFile(
                 target, pm, m, output_c.as_ptr(), file_type);
         if !result {
@@ -221,13 +221,13 @@ fn create_target_machine(sess: &Session) -> TargetMachineRef {
     let triple = &sess.target.target.llvm_target[];
 
     let tm = unsafe {
-        let triple = CString::from_slice(triple.as_bytes());
+        let triple = CString::new(triple.as_bytes()).unwrap();
         let cpu = match sess.opts.cg.target_cpu {
             Some(ref s) => &**s,
             None => &*sess.target.target.options.cpu
         };
-        let cpu = CString::from_slice(cpu.as_bytes());
-        let features = CString::from_slice(target_feature(sess).as_bytes());
+        let cpu = CString::new(cpu.as_bytes()).unwrap();
+        let features = CString::new(target_feature(sess).as_bytes()).unwrap();
         llvm::LLVMRustCreateTargetMachine(
             triple.as_ptr(), cpu.as_ptr(), features.as_ptr(),
             code_model,
@@ -380,7 +380,7 @@ unsafe extern "C" fn diagnostic_handler(info: DiagnosticInfoRef, user: *mut c_vo
         }
 
         llvm::diagnostic::Optimization(opt) => {
-            let pass_name = str::from_utf8(ffi::c_str_to_bytes(&opt.pass_name))
+            let pass_name = str::from_utf8(CStr::from_ptr(opt.pass_name).to_bytes())
                                 .ok()
                                 .expect("got a non-UTF8 pass name from LLVM");
             let enabled = match cgcx.remark {
@@ -424,7 +424,7 @@ unsafe fn optimize_and_codegen(cgcx: &CodegenContext,
     if config.emit_no_opt_bc {
         let ext = format!("{}.no-opt.bc", name_extra);
         let out = output_names.with_extension(&ext);
-        let out = CString::from_slice(out.as_vec());
+        let out = CString::new(out.as_vec()).unwrap();
         llvm::LLVMWriteBitcodeToFile(llmod, out.as_ptr());
     }
 
@@ -440,7 +440,7 @@ unsafe fn optimize_and_codegen(cgcx: &CodegenContext,
             // If we're verifying or linting, add them to the function pass
             // manager.
             let addpass = |pass: &str| {
-                let pass = CString::from_slice(pass.as_bytes());
+                let pass = CString::new(pass).unwrap();
                 llvm::LLVMRustAddPass(fpm, pass.as_ptr())
             };
             if !config.no_verify { assert!(addpass("verify")); }
@@ -453,7 +453,7 @@ unsafe fn optimize_and_codegen(cgcx: &CodegenContext,
             }
 
             for pass in &config.passes {
-                let pass = CString::from_slice(pass.as_bytes());
+                let pass = CString::new(pass.clone()).unwrap();
                 if !llvm::LLVMRustAddPass(mpm, pass.as_ptr()) {
                     cgcx.handler.warn(&format!("unknown pass {:?}, ignoring", pass));
                 }
@@ -477,7 +477,7 @@ unsafe fn optimize_and_codegen(cgcx: &CodegenContext,
                     if config.emit_lto_bc {
                         let name = format!("{}.lto.bc", name_extra);
                         let out = output_names.with_extension(&name);
-                        let out = CString::from_slice(out.as_vec());
+                        let out = CString::new(out.as_vec()).unwrap();
                         llvm::LLVMWriteBitcodeToFile(llmod, out.as_ptr());
                     }
                 },
@@ -511,15 +511,15 @@ unsafe fn optimize_and_codegen(cgcx: &CodegenContext,
     if config.emit_bc {
         let ext = format!("{}.bc", name_extra);
         let out = output_names.with_extension(&ext);
-        let out = CString::from_slice(out.as_vec());
+        let out = CString::new(out.as_vec()).unwrap();
         llvm::LLVMWriteBitcodeToFile(llmod, out.as_ptr());
     }
 
     time(config.time_passes, "codegen passes", (), |()| {
         if config.emit_ir {
             let ext = format!("{}.ll", name_extra);
             let out = output_names.with_extension(&ext);
-            let out = CString::from_slice(out.as_vec());
+            let out = CString::new(out.as_vec()).unwrap();
             with_codegen(tm, llmod, config.no_builtins, |cpm| {
                 llvm::LLVMRustPrintModule(cpm, llmod, out.as_ptr());
             })
@@ -1004,7 +1004,7 @@ unsafe fn configure_llvm(sess: &Session) {
     let mut llvm_args = Vec::new();
     {
         let mut add = |arg: &str| {
-            let s = CString::from_slice(arg.as_bytes());
+            let s = CString::new(arg).unwrap();
             llvm_args.push(s.as_ptr());
             llvm_c_strs.push(s);
         };

src/librustc_trans/trans/asm.rs

@@ -120,8 +120,8 @@ pub fn trans_inline_asm<'blk, 'tcx>(bcx: Block<'blk, 'tcx>, ia: &ast::InlineAsm)
         ast::AsmIntel => llvm::AD_Intel
     };
 
-    let asm = CString::from_slice(ia.asm.as_bytes());
-    let constraints = CString::from_slice(constraints.as_bytes());
+    let asm = CString::new(ia.asm.as_bytes()).unwrap();
+    let constraints = CString::new(constraints).unwrap();
     let r = InlineAsmCall(bcx,
                           asm.as_ptr(),
                           constraints.as_ptr(),

src/librustc_trans/trans/base.rs

@@ -86,7 +86,7 @@ use util::nodemap::NodeMap;
 
 use arena::TypedArena;
 use libc::{c_uint, uint64_t};
-use std::ffi::{self, CString};
+use std::ffi::{CStr, CString};
 use std::cell::{Cell, RefCell};
 use std::collections::HashSet;
 use std::mem;
@@ -186,7 +186,7 @@ impl<'a, 'tcx> Drop for StatRecorder<'a, 'tcx> {
 pub fn decl_fn(ccx: &CrateContext, name: &str, cc: llvm::CallConv,
                ty: Type, output: ty::FnOutput) -> ValueRef {
 
-    let buf = CString::from_slice(name.as_bytes());
+    let buf = CString::new(name).unwrap();
     let llfn: ValueRef = unsafe {
         llvm::LLVMGetOrInsertFunction(ccx.llmod(), buf.as_ptr(), ty.to_ref())
     };
@@ -340,7 +340,7 @@ pub fn get_extern_const<'a, 'tcx>(ccx: &CrateContext<'a, 'tcx>, did: ast::DefId,
         None => ()
     }
     unsafe {
-        let buf = CString::from_slice(name.as_bytes());
+        let buf = CString::new(name.clone()).unwrap();
         let c = llvm::LLVMAddGlobal(ccx.llmod(), ty.to_ref(), buf.as_ptr());
         // Thread-local statics in some other crate need to *always* be linked
         // against in a thread-local fashion, so we need to be sure to apply the
@@ -2788,7 +2788,7 @@ pub fn get_item_val(ccx: &CrateContext, id: ast::NodeId) -> ValueRef {
                                 &format!("Illegal null byte in export_name \
                                          value: `{}`", sym)[]);
                         }
-                        let buf = CString::from_slice(sym.as_bytes());
+                        let buf = CString::new(sym.clone()).unwrap();
                         let g = llvm::LLVMAddGlobal(ccx.llmod(), llty,
                                                     buf.as_ptr());
 
@@ -2826,7 +2826,7 @@ pub fn get_item_val(ccx: &CrateContext, id: ast::NodeId) -> ValueRef {
                                                  &sect)[]);
                     }
                     unsafe {
-                        let buf = CString::from_slice(sect.as_bytes());
+                        let buf = CString::new(sect.as_bytes()).unwrap();
                         llvm::LLVMSetSection(v, buf.as_ptr());
                     }
                 },
@@ -2993,15 +2993,15 @@ pub fn write_metadata(cx: &SharedCrateContext, krate: &ast::Crate) -> Vec<u8> {
     let name = format!("rust_metadata_{}_{}",
                        cx.link_meta().crate_name,
                        cx.link_meta().crate_hash);
-    let buf = CString::from_vec(name.into_bytes());
+    let buf = CString::new(name).unwrap();
     let llglobal = unsafe {
         llvm::LLVMAddGlobal(cx.metadata_llmod(), val_ty(llconst).to_ref(),
                             buf.as_ptr())
     };
     unsafe {
         llvm::LLVMSetInitializer(llglobal, llconst);
         let name = loader::meta_section_name(cx.sess().target.target.options.is_like_osx);
-        let name = CString::from_slice(name.as_bytes());
+        let name = CString::new(name).unwrap();
         llvm::LLVMSetSection(llglobal, name.as_ptr())
     }
     return metadata;
@@ -3039,8 +3039,8 @@ fn internalize_symbols(cx: &SharedCrateContext, reachable: &HashSet<String>) {
                     continue
                 }
 
-                let name = ffi::c_str_to_bytes(&llvm::LLVMGetValueName(val))
-                               .to_vec();
+                let name = CStr::from_ptr(llvm::LLVMGetValueName(val))
+                                .to_bytes().to_vec();
                 declared.insert(name);
             }
         }
@@ -3056,8 +3056,8 @@ fn internalize_symbols(cx: &SharedCrateContext, reachable: &HashSet<String>) {
                     continue
                 }
 
-                let name = ffi::c_str_to_bytes(&llvm::LLVMGetValueName(val))
-                               .to_vec();
+                let name = CStr::from_ptr(llvm::LLVMGetValueName(val))
+                                .to_bytes().to_vec();
                 if !declared.contains(&name) &&
                    !reachable.contains(str::from_utf8(&name).unwrap()) {
                     llvm::SetLinkage(val, llvm::InternalLinkage);

src/librustc_trans/trans/builder.rs

@@ -431,7 +431,7 @@ impl<'a, 'tcx> Builder<'a, 'tcx> {
             if name.is_empty() {
                 llvm::LLVMBuildAlloca(self.llbuilder, ty.to_ref(), noname())
             } else {
-                let name = CString::from_slice(name.as_bytes());
+                let name = CString::new(name).unwrap();
                 llvm::LLVMBuildAlloca(self.llbuilder, ty.to_ref(),
                                       name.as_ptr())
             }
@@ -786,7 +786,7 @@ impl<'a, 'tcx> Builder<'a, 'tcx> {
             let comment_text = format!("{} {}", "#",
                                        sanitized.replace("\n", "\n\t# "));
             self.count_insn("inlineasm");
-            let comment_text = CString::from_vec(comment_text.into_bytes());
+            let comment_text = CString::new(comment_text).unwrap();
             let asm = unsafe {
                 llvm::LLVMConstInlineAsm(Type::func(&[], &Type::void(self.ccx)).to_ref(),
                                          comment_text.as_ptr(), noname(), False,

src/librustc_trans/trans/common.rs

@@ -488,7 +488,7 @@ impl<'a, 'tcx> FunctionContext<'a, 'tcx> {
                      opt_node_id: Option<ast::NodeId>)
                      -> Block<'a, 'tcx> {
         unsafe {
-            let name = CString::from_slice(name.as_bytes());
+            let name = CString::new(name).unwrap();
             let llbb = llvm::LLVMAppendBasicBlockInContext(self.ccx.llcx(),
                                                            self.llfn,
                                                            name.as_ptr());
@@ -761,7 +761,7 @@ pub fn C_integral(t: Type, u: u64, sign_extend: bool) -> ValueRef {
 
 pub fn C_floating(s: &str, t: Type) -> ValueRef {
     unsafe {
-        let s = CString::from_slice(s.as_bytes());
+        let s = CString::new(s).unwrap();
         llvm::LLVMConstRealOfString(t.to_ref(), s.as_ptr())
     }
 }
@@ -839,7 +839,8 @@ pub fn C_cstr(cx: &CrateContext, s: InternedString, null_terminated: bool) -> Va
                                                 !null_terminated as Bool);
 
         let gsym = token::gensym("str");
-        let buf = CString::from_vec(format!("str{}", gsym.usize()).into_bytes());
+        let buf = CString::new(format!("str{}", gsym.usize()));
+        let buf = buf.unwrap();
         let g = llvm::LLVMAddGlobal(cx.llmod(), val_ty(sc).to_ref(), buf.as_ptr());
         llvm::LLVMSetInitializer(g, sc);
         llvm::LLVMSetGlobalConstant(g, True);

src/librustc_trans/trans/context.rs

@@ -225,15 +225,15 @@ impl<'a, 'tcx> Iterator for CrateContextMaybeIterator<'a, 'tcx> {
 
 unsafe fn create_context_and_module(sess: &Session, mod_name: &str) -> (ContextRef, ModuleRef) {
     let llcx = llvm::LLVMContextCreate();
-    let mod_name = CString::from_slice(mod_name.as_bytes());
+    let mod_name = CString::new(mod_name).unwrap();
     let llmod = llvm::LLVMModuleCreateWithNameInContext(mod_name.as_ptr(), llcx);
 
-    let data_layout = &*sess.target.target.data_layout;
-    let data_layout = CString::from_slice(data_layout.as_bytes());
+    let data_layout = sess.target.target.data_layout.as_bytes();
+    let data_layout = CString::new(data_layout).unwrap();
     llvm::LLVMSetDataLayout(llmod, data_layout.as_ptr());
 
-    let llvm_target = &*sess.target.target.llvm_target;
-    let llvm_target = CString::from_slice(llvm_target.as_bytes());
+    let llvm_target = sess.target.target.llvm_target.as_bytes();
+    let llvm_target = CString::new(llvm_target).unwrap();
     llvm::LLVMRustSetNormalizedTarget(llmod, llvm_target.as_ptr());
     (llcx, llmod)
 }