commit 0b01807f10760f40e4196b83763d45523b65b239
parent 33c4972b3d77d02d4865da6c9bd159eeefa523ab
Author: NunoSempere <nuno.sempere@protonmail.com>
Date: Sat, 9 Sep 2023 11:01:04 +0200
improvements based on other wc versions
Diffstat:
4 files changed, 37 insertions(+), 17 deletions(-)
diff --git a/README.md b/README.md
@@ -1,21 +1,38 @@
-Desiderata
+# ww: count words in 50 lines of C
-- Simple: Simple operation in terms of counting spaces and \n.
-- Avoid "off by one" errors; make sure an empty file is reported as such.
- - Words as space or enter, followed by nonspace, followed by space? Make sure two spaces aren't two words?
-- Keep Linux only.
-- No flags. Only count words, not lines.
+## Desiderata
+
+- Simplicity: Just count words, as delimited by: spaces, tabs, newlines.
+- No flags.
+- Avoid off-by-one errors.
- Allow piping, as well as reading files.
- - Wonder how normal utilities handle this.
-- Could use zig? => Not for now
+- Small.
+- Linux only.
+
+## Comparison with wc.
+
+The GNU utils version ([github](https://github.com/coreutils/coreutils/tree/master/src/wc), [savannah](http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=blob;f=src/wc.c;hb=HEAD)) is a bit over 1K lines of C. It does many things and checks many possible failure modes.
+
+The busybox version ([git.busybox.net](https://git.busybox.net/busybox/tree/coreutils/wc.c)) of wc is much shorter, at 257 lines, while striving to be [POSIX-compliant](https://pubs.opengroup.org/onlinepubs/9699919799/), meaning it has flags.
+
+The plan9port version of wc ([github](https://github.com/9fans/plan9port/blob/master/src/cmd/wc.c)) implements some sort of table method, in 352 lines. So does the [plan9](https://9p.io/sources/plan9/sys/src/cmd/wc.c) version, which is worse documented, but shorter.
+
+[Here](https://github.com/dspinellis/unix-history-repo/blob/Research-V7-Snapshot-Development/usr/src/cmd/wc.c) is a version of wc from UNIX V7, at 86 lines, and allowing for both word and line counts. I couldn't find a version in UNIX V6. Of all the versions, I think I understand this one best.
-Steps:
+## Steps:
- [x] Look into how C utilities both read from stdin and from files.
- [x] Program first version of the utility
-- [ ] Compare with other implementations, see how they do it, after I've read my own version
- - [ ] Compare with gnu utils,
- - Compare with musl/busybox implementations,
- - Maybe make some pull requests, if I'm doing something better?
+- [x] Compare with other implementations, see how they do it, after I've read my own version
+ - [x] Compare with gnu utils.
+
+ - [x] Compare with musl/busybox implementations,
+ - ~~Maybe make some pull requests, if I'm doing something better? => doesn't seem like it~~
- [ ] Install to ww, but check that ww is empty (installing to wc2 or smth would mean that you don't save that many keypresses vs wc -w)
-- [ ] ...
+- ~~[ ] Could use zig? => Not for now~~
+- [ ] Look specifically at how other versions
+ - [ ] Distinguish between reading from stdin and reading from a file
+ - If it doesn't have arguments, read from stdin.
+ - [ ] Open files, read characters.
+- [ ] Write version that counts lines
+- [ ]
diff --git a/makefile b/makefile
@@ -15,7 +15,7 @@ OUT=ww
DEBUG= #'-g'
STANDARD=-std=c99
WARNINGS=-Wall
-OPTIMIZED=-O0
+OPTIMIZED=-O3
# OPTIMIZED=-O3 #-Ofast
## Formatter
@@ -29,6 +29,9 @@ build: $(SRC)
format: $(SRC)
$(FORMATTER) $(SRC)
+install:
+ cp -n $(OUT) /bin/$(OUT)
+
test: $(OUT)
/bin/echo -e "123\n45 67" | ./$(OUT)
/bin/echo -n "" | ./ww
diff --git a/ww b/ww
Binary files differ.
diff --git a/ww.c b/ww.c
@@ -1,7 +1,6 @@
#include <stdio.h>
-#include <unistd.h> // read, isatty
+#include <unistd.h>
-// STDIN_FILENO
int process_fn(int fn)
{
char c[1];
@@ -41,6 +40,7 @@ int main(int argc, char** argv)
perror("Could not open file");
return 1;
}
+ fclose(fp);
return process_fn(fileno(fp));
} else {
printf("Usage: ww file.txt\n");
