For when the length of the C string is already known, so it doesn't have
to be recalculated.

Signed-off-by: Sebastian <sebastian@sebsite.pw>
---
 types/c/strings.ha | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/types/c/strings.ha b/types/c/strings.ha
index 20dae411..dce99c3c 100644
--- a/types/c/strings.ha
+++ b/types/c/strings.ha
@@ -13,11 +13,16 @@ export fn strlen(cstr: *const char) size = {
 // Converts a C string to a Hare string in O(n), and does not check if it's
 // valid UTF-8.
 export fn tostr_unsafe(cstr: *const char) const str = {
-	const l = strlen(cstr);
+	return tostrn_unsafe(cstr, strlen(cstr));
+};
+
+// Converts a C string with a given length to a Hare string, and does not check
+// if it's valid UTF-8.
+export fn tostrn_unsafe(cstr: *const char, length: size) const str = {
 	const s = types::string {
 		data = cstr: *[*]u8,
-		length = l,
-		capacity = l + 1,
+		length = length,
+		capacity = length + 1,
 	};
 	return *(&s: *const str);
 };
@@ -25,7 +30,13 @@ export fn tostr_unsafe(cstr: *const char) const str = {
 // Converts a C string to a Hare string in O(n). If the string is not valid
 // UTF-8, return [[encoding::utf8::invalid]].
 export fn tostr(cstr: *const char) (const str | utf8::invalid) = {
-	let s = tostr_unsafe(cstr);
+	return tostrn(cstr, strlen(cstr));
+};
+
+// Converts a C string with a given length to a Hare string. If the string is
+// not valid UTF-8, return [[encoding::utf8::invalid]].
+export fn tostrn(cstr: *const char, length: size) (const str | utf8::invalid) = {
+	let s = tostrn_unsafe(cstr, length);
 	return if (utf8::valid(s)) s else utf8::invalid;
 };
 
-- 
2.38.5

hare/patches: SUCCESS in 1m44s

[types::c: add tostrn and tostrn_unsafe][0] from [Sebastian][1]

[0]: https://lists.sr.ht/~sircmpwn/hare-dev/patches/41099
[1]: sebastian@sebsite.pw

✓ #988959 SUCCESS hare/patches/alpine.yml  https://builds.sr.ht/~sircmpwn/job/988959
✓ #988960 SUCCESS hare/patches/freebsd.yml https://builds.sr.ht/~sircmpwn/job/988960

+1. I also think that it would be nice to have a function here which can
convert strings to C strings without allocation if the caller ensures
that the nul delimiter is there. Should be simple so that it can be
easily used for interop:

printf(c::string("hello world!\n\0"));

On Sat May 13, 2023 at 8:12 AM UTC, Drew DeVault wrote:
> +1. I also think that it would be nice to have a function here which can
> convert strings to C strings without allocation if the caller ensures
> that the nul delimiter is there. Should be simple so that it can be
> easily used for interop:
>
> printf(c::string("hello world!\n\0"));

c::tostr currently does that, maybe we should rename it? i'm also not a
huge fan of tostrn as a name, but i don't have ideas on how it could be
better

On Sat May 13, 2023 at 9:55 AM EDT, Ember Sawady wrote:
> On Sat May 13, 2023 at 8:12 AM UTC, Drew DeVault wrote:
> > +1. I also think that it would be nice to have a function here which can
> > convert strings to C strings without allocation if the caller ensures
> > that the nul delimiter is there. Should be simple so that it can be
> > easily used for interop:
> >
> > printf(c::string("hello world!\n\0"));
>
> c::tostr currently does that, maybe we should rename it? i'm also not a
> huge fan of tostrn as a name, but i don't have ideas on how it could be
> better

No, c::fromstr is what Drew is suggesting (converting from a Hare string
to a C string), but it allocates. I agree that an alternative that
doesn't allocate would be nice, but maybe fromstr should be renamed.

Agreed that tostrn isn't great, open to suggestions.