hi,

v3 simplies some logic in the if statement

---
 Makefile |  4 +++-
 yes.ha   | 22 ++++++++++++++++++++++
 2 files changed, 25 insertions(+), 1 deletion(-)
 create mode 100644 yes.ha

diff --git a/Makefile b/Makefile
index c568580..298db4b 100644
--- a/Makefile
+++ b/Makefile
@@ -18,7 +18,8 @@ utils=\
 	true \
 	uname \
 	uniq \
-	wc
+	wc \
+	yes
 
 all: $(utils)
 
@@ -46,3 +47,4 @@ true: true.ha
 uname: uname.ha main/main.ha
 uniq: uniq.ha main/main.ha
 wc: wc.ha main/main.ha
+yes: yes.ha main/main.ha
diff --git a/yes.ha b/yes.ha
new file mode 100644
index 0000000..957bbc0
--- /dev/null
+++ b/yes.ha
@@ -0,0 +1,22 @@
+use fmt;
+use getopt;
+use os;
+use main;
+
+export fn utilmain() (main::error | void) = {
+	const help: []getopt::help = [
+		"output string repeatedly",
+		"<string>",
+	];
+	const cmd = getopt::parse(os::args, help...);
+	defer getopt::finish(&cmd);
+	if (len(cmd.args) > 1) {
+		getopt::printhelp(os::stderr, os::args[0], help);
+		os::exit(1);
+	};
+	const s = if (len(cmd.args) == 1) cmd.args[0] else "y";
+	for (true) {
+		fmt::printfln(s)!;
+	};
+	return void;
+};
-- 
2.39.1

On Tue Feb 21, 2023 at 8:32 AM EST, jgart wrote:
> hi,
>
> v3 simplies some logic in the if statement
>
> ---
> (snip)
> diff --git a/yes.ha b/yes.ha
> new file mode 100644
> index 0000000..957bbc0
> --- /dev/null
> +++ b/yes.ha
> @@ -0,0 +1,22 @@
> +use fmt;
> +use getopt;
> +use os;
> +use main;
> +
> +export fn utilmain() (main::error | void) = {
> +	const help: []getopt::help = [
> +		"output string repeatedly",
> +		"<string>",
> +	];
> +	const cmd = getopt::parse(os::args, help...);
> +	defer getopt::finish(&cmd);
> +	if (len(cmd.args) > 1) {
> +		getopt::printhelp(os::stderr, os::args[0], help);
> +		os::exit(1);
> +	};
> +	const s = if (len(cmd.args) == 1) cmd.args[0] else "y";
> +	for (true) {
> +		fmt::printfln(s)!;

fmt::printfln should not be used, because s is not a format string:

master$ ./yes '{}'
Abort: /usr/src/hare/stdlib/fmt/fmt.ha:220:31: Not enough parameters given
Aborted (core dumped)

You should use fmt::println instead.

> +	};
> +	return void;
> +};
> -- 
> 2.39.1

Hi Sebastian,

> You should use fmt::println instead.

See v4 where this is fixed. I just saw your message now but someone else had also pointed this out to me offlist.

Thank you for the review!

all best,

jgart

On 2023-02-21 14:32 UTC, jgart wrote:
> hi,
>
> v3 simplies some logic in the if statement
>
> ---
>  Makefile |  4 +++-
>  yes.ha   | 22 ++++++++++++++++++++++
>  2 files changed, 25 insertions(+), 1 deletion(-)
>  create mode 100644 yes.ha
>
> diff --git a/Makefile b/Makefile
> index c568580..298db4b 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -18,7 +18,8 @@ utils=\
>  	true \
>  	uname \
>  	uniq \
> -	wc
> +	wc \
> +	yes
>  
>  all: $(utils)
>  
> @@ -46,3 +47,4 @@ true: true.ha
>  uname: uname.ha main/main.ha
>  uniq: uniq.ha main/main.ha
>  wc: wc.ha main/main.ha
> +yes: yes.ha main/main.ha
> diff --git a/yes.ha b/yes.ha
> new file mode 100644
> index 0000000..957bbc0
> --- /dev/null
> +++ b/yes.ha
> @@ -0,0 +1,22 @@
> +use fmt;
> +use getopt;
> +use os;
> +use main;
> +
> +export fn utilmain() (main::error | void) = {
> +	const help: []getopt::help = [
> +		"output string repeatedly",
> +		"<string>",
> +	];
> +	const cmd = getopt::parse(os::args, help...);
> +	defer getopt::finish(&cmd);
> +	if (len(cmd.args) > 1) {
> +		getopt::printhelp(os::stderr, os::args[0], help);
> +		os::exit(1);
> +	};
> +	const s = if (len(cmd.args) == 1) cmd.args[0] else "y";
> +	for (true) {
> +		fmt::printfln(s)!;
> +	};
> +	return void;
> +};
> -- 
> 2.39.1

I don't know if performance parity with coreutils is a goal for hautils
in general or for `yes` is particular, but they are several orders of
magnitude apart:

  $ timeout 5 /usr/bin/yes | pv -ba >/dev/null
  22.8GiB [4.56GiB/s]

  $ timeout 5 ./yes | pv -ba >/dev/null
  11.2MiB [2.24MiB/s]

coreutils' `yes` does buffered IO:

  $ strace /usr/bin/yes >/dev/null
  ...
  write(1, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192
  write(1, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192
  write(1, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192
  write(1, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192
  ...

I think they pre fill a buffer and just call write(2) repeatedly with
it.  If performance is a consideration in this case, it shouldn't be
that hard to do something similar.

The performance of yes(1) is not of any conseqeunce. Simplicity is the
word.

On 2023-02-21 19:59 UTC, Drew DeVault wrote:
> The performance of yes(1) is not of any conseqeunce. Simplicity is the
> word.

Yes, it makes sense.  Just wondering if someone could have some use-case
expecting a performant yes(1).  Not my case, though.

It seems that hautils might lean more in the spirit/aesthetics of sbase but I may be wrong:

https://github.com/michaelforney/sbase

Thanks for the all the code reviews. Much appreciated!

--
jgart

Skipping this on the basis that it's not POSIX.