libdelimit: fixes tabs

Makefile

@@ -30,7 +30,8 @@ SYSEXITS != printf '\043include <sysexits.h>\n' | cpp -M - | tr ' ' '\n' \
 
 CC ?= cc
 RUSTC ?= rustc
-RUSTFLAGS += --extern getopt=build/o/libgetopt.rlib \
+RUSTFLAGS += --extern delimit=build/o/libdelimit.rlib \
+	--extern getopt=build/o/libgetopt.rlib \
 	--extern strerror=build/o/libstrerror.rlib \
 	--extern sysexits=build/o/libsysexits.rlib
 CFLAGS += -I$(SYSEXITS)
@@ -70,12 +71,16 @@ TESTS != printf '%s\n' "$(TESTFILES)" | xargs -n1 basename \
 include $(TESTFILES)
 
 .PHONY: test
-test: all $(TESTS) /tmp/getopt
+test: all $(TESTS) /tmp/delimit /tmp/getopt
 	@echo $(TESTS)
+	/tmp/delimit
 	/tmp/getopt
 
+/tmp/delimit: src/libdelimit.rs
+	$(RUSTC) --test -o $@ src/libdelimit.rs
+
 /tmp/getopt: src/libgetopt.rs
-	$(RUSTC) --test -o /tmp/getopt src/libgetopt.rs
+	$(RUSTC) --test -o $@ src/libgetopt.rs
 
 .PHONY: docs
 docs: docs/ build
@@ -88,8 +93,12 @@ docs: docs/ build
 include $(OS_INCLUDE)
 
 .PHONY: rustlibs
-rustlibs: build/o/libgetopt.rlib build/o/libstrerror.rlib \
-	 build/o/libsysexits.rlib $(OSLIB)
+rustlibs: build/o/libdelimit.rlib build/o/libgetopt.rlib \
+	 build/o/libstrerror.rlib build/o/libsysexits.rlib $(OSLIB)
+
+build/o/libdelimit.rlib: build src/libdelimit.rs
+	$(RUSTC) $(RUSTFLAGS) --crate-type=lib --crate-name=delimit \
+		-o $@ src/libdelimit.rs
 
 build/o/libgetopt.rlib: build src/libgetopt.rs
 	$(RUSTC) $(RUSTFLAGS) --crate-type=lib --crate-name=getopt \

docs/fop.1

@@ -11,10 +11,9 @@ fop \(en field operator
 .SH SYNOPSIS
 
 fop
-.RB ( -d )
-.RB [ delimiter ]
-.RB index 
-.RB program...
+.RB [ -d\ delimiter ]
+index program
+.RB [ arguments... ]
 .\"
 .SH DESCRIPTION
 
@@ -26,8 +25,17 @@ Performs operations on specified fields in data read from the standard input.
 Sets a delimiter by which the input data will be split into fields. The default
 is an ASCII record separator.
 .\"
+.SH DIAGNOSTICS
+in the event of an error, a debug message will be printed and the program will
+exit with the appropriate sysexits.h(3) error code.
+.\"
 .SH CAVEATS
 
+If the specified index does not exist in the data, the program
+will print all data to the standard output before exiting with an error. If
+input data is not delimited by the specified delimiter, the program will fill
+memory with the contents of the stream before it is output.
+
 Field indices are zero-indexed, which may be unexpected behavior for some users.
 .\"
 .SH RATIONALE

src/fop.rs

@@ -18,14 +18,16 @@
 
 use std::{
 	env::args,
-	io::{ Error, Read, Write, stdin, stdout },
+	io::{ Error, Write, stdin, stdout },
 	process::{ Command, ExitCode, Stdio, exit },
 };
 
+extern crate delimit;
 extern crate getopt;
 extern crate strerror;
 extern crate sysexits;
 
+use delimit::Delimited;
 use getopt::GetOpt;
 use strerror::StrError;
 use sysexits::{ EX_DATAERR, EX_IOERR, EX_UNAVAILABLE, EX_USAGE };
@@ -90,78 +92,97 @@ fn main() -> ExitCode {
 		exit(usage(&argv[0]).into());
 	});
 
-	/* read entire standard input into memory */
-	let mut buf = String::new();
-	if let Err(e) = stdin().read_to_string(&mut buf) {
-		err(&argv[0], e);
-		exit(EX_IOERR.into());
-	};
+	let stdin = stdin().lock();
 
-	/* split the buffer by the delimiter (by default, '\u{1E}') */
-	let mut fields = buf.split(&d).collect::<Vec<&str>>();
+	let mut input = Delimited::new(stdin, d.clone().as_bytes());
+	let mut n = 0;
 
-	/* collect arguments for the operator command */
-	let command_args = argv
-		.iter()
-		.clone()
-		.skip(command_arg + 1) /* skip the command name */
-		.collect::<Vec<&String>>();
+	let mut fopped = false;
 
-	/* spawn the command to operate on the field */
-	let mut spawned = Command::new(operator)
-		.args(command_args) /* spawn with the specified arguments */
-		.stdin(Stdio::piped())
-		.stdout(Stdio::piped()) /* piped stdout to handle output ourselves */
-		.spawn()
-		.unwrap_or_else( |e| {
+	while let Some(i) = input.next() {
+		let v = match i {
+			Ok(v) => v,
+			Err(e) => {
+				err(&argv[0], e);
+				exit(EX_IOERR.into());
+			},
+		};
+
+		let mut out = Vec::new();
+
+		if n == index { /* fop it */
+			/* collect arguments for the operator command */
+			let command_args = argv
+				.iter()
+				.clone()
+				.skip(command_arg + 1) /* skip the command name */
+				.collect::<Vec<&String>>();
+
+			/* spawn the command to operate on the field */
+			let mut spawned = Command::new(operator)
+				.args(command_args) /* spawn with the specified arguments */
+				.stdin(Stdio::piped())
+				/* piped stdout to handle output ourselves */
+				.stdout(Stdio::piped()) 
+				.spawn()
+				.unwrap_or_else( |e| {
+					err(&argv[0], e);
+					exit(EX_UNAVAILABLE.into());
+				});
+
+			/* feed the spawned program’s stdin the field value */
+			if let Some(mut child_stdin) = spawned.stdin.take() {
+				let _ = child_stdin.write_all(&v);
+				drop(child_stdin); /* stay safe! drop your children! */
+
+				let output = spawned.wait_with_output().unwrap_or_else(|e| {
+					err(&argv[0], e);
+					exit(EX_IOERR.into());
+			});
+
+				/* get the output with which the original field will
+				 * be replaced */
+				let mut replace = output.stdout.clone();
+
+				/* pop trailing newline out if the input did not contain it */
+				if v.iter().last() != Some(&b'\n')
+					&& replace.pop() != Some(b'\n')
+				{
+					out = output.stdout;
+				} else {
+					out = replace;
+				}
+			}
+
+			fopped = true;
+		} else {
+			out = v;
+		}
+
+		/* since we cannot know when we’re done, place a new delimiter before
+		 * each index unless it is the 0th */
+		if n != 0 {
+			stdout().write_all(d.as_bytes()).unwrap_or_else(|e| {
+				err(&argv[0], e);
+				exit(EX_IOERR.into());
+			});
+		}
+
+		stdout().write_all(&out).unwrap_or_else(|e| {
 			err(&argv[0], e);
-			exit(EX_UNAVAILABLE.into());
+			exit(EX_IOERR.into());
 		});
 
-	/* get field we want to pipe into spawned program */
-	let field = fields.get(index).unwrap_or_else(|| {
+
+		n += 1;
+	}
+
+	if fopped {
+		return ExitCode::SUCCESS;
+	} else {
 		eprintln!("{}: {}: no such index in input", argv[0], index);
-		exit(EX_DATAERR.into());
-	});
-
-	/* get the stdin of the newly spawned program and feed it the field val */
-	if let Some(mut child_stdin) = spawned.stdin.take() {
-		let _ = child_stdin.write_all(field.as_bytes());
-		drop(child_stdin); /* stay safe! drop your children! */
+		return EX_DATAERR.into();
 	}
 
-	let output = spawned.wait_with_output().unwrap_or_else(|e| {
-		err(&argv[0], e);
-		exit(EX_IOERR.into());
-	});
 
-	/* get the output with which the original field will be replaced */
-	let mut replace = output.stdout.clone();
-
-	/* pop trailing newline out if the input did not contain it */
-	if fields[index].chars().last() != Some('\n') /* no newline */
-		&& replace.pop() != Some(b'\n') { /* pop last char of replacement */
-		/* restore replacement to original command output if popped char was not
-		 * a newline */
-		replace = output.stdout;
-	}
-
-	/* convert the output of the program to UTF-8 */
-	let new_field = String::from_utf8(replace).unwrap_or_else(|e| {
-		eprintln!("{}: {}", argv[0], e);
-		exit(EX_IOERR.into());
-	});
-
-	/* store the new field in the old fields vector */
-	fields[index] = &new_field;
-
-	/* fop it */ 
-	stdout().write_all(
-		fields.join(&d.to_string()).as_bytes()
-	).unwrap_or_else(|e| {
-		err(&argv[0], e);
-		exit(EX_IOERR.into());
-	});
-
-	ExitCode::SUCCESS
 }

src/libdelimit.rs

@@ -0,0 +1,101 @@
+/*
+ * Copyright (c) 2025 Emma Tebibyte <emma@tebibyte.media>
+ * Copyright (c) 2025 silty silt <silt@tebibyte.media>
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ *
+ * This program is free software: you can redistribute it and/or modify it
+ * under the terms of the GNU Affero General Public License as published by the
+ * Free Software Foundation, either version 3 of the License, or (at your
+ * option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License
+ * for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see https://www.gnu.org/licenses/.
+ */
+
+use std::{
+	io::{ Read, Result },
+	mem,
+};
+
+const BUFFER_SIZE: usize = 4096;
+
+pub struct Delimited<T: Read> {
+	delimiter: Vec<u8>,
+	buffer: Vec<u8>,
+	stream: T,
+}
+
+impl<T> Delimited<T> where T: Read {
+	pub fn new(stream: T, delimiter: &[u8]) -> Self {
+		Delimited {
+			stream,
+			delimiter: delimiter.to_vec(),
+			buffer: Vec::with_capacity(BUFFER_SIZE),
+		}
+	}
+}
+
+impl<T> Iterator for Delimited<T> where T: Read {
+	type Item = Result<Vec<u8>>;
+
+	fn next(&mut self) -> Option<Self::Item> {
+		let mut buf = [0; BUFFER_SIZE];
+
+		loop {
+			if let Some(p) = find_subslice(&self.buffer, &self.delimiter) {
+				let chunk = self.buffer.drain(..p).collect::<Vec<_>>();
+				
+				let _ = self.buffer.drain(..self.delimiter.len());
+
+				return Some(Ok(chunk));
+			}
+
+			match self.stream.read(&mut buf) {
+				Ok(0) => {
+					let _ = self.buffer.is_empty() && return None;
+
+					return Some(Ok(mem::take(&mut self.buffer)));
+				},
+				Ok(n) => {
+					self.buffer.extend_from_slice(&buf[..n]);
+				},
+				Err(e) => {
+					return Some(Err(e));
+				},
+			}
+		}
+	}
+}
+
+fn find_subslice(stack: &[u8], key: &[u8]) -> Option<usize> {
+	match key.len() {
+		/* TODO: is this optimization necessary? */
+		1 => stack.iter().position(|&b| b == key[0]),
+		_ => stack.windows(key.len()).position(|w| w == key),
+	}
+}
+
+#[cfg(test)]
+mod tests {
+	use Delimited;
+
+	#[test]
+	fn testing() {
+		let d = '\u{1E}'.to_string();
+		let input = vec!["meow", "woof", "ribbit"];
+		let r = input.join(&d);
+
+		let mut output = Delimited::new(r.as_bytes(), d.as_bytes());
+
+		let mut i = 0;
+		while let Some(item) = output.next() {
+			assert_eq!(input[i].as_bytes(), item.unwrap());
+			i += 1;
+		}
+	}
+}