Add fromPtr variant that is NULL-terminated

[sjshuck]

For my case, I need to copy a region of UTF-16 text from the heap into a Text. It is NULL (U+0000) terminated.

Currently I'm scanning each 16-bit code unit in a loop to detemine the length and then calling Data.Text.Foreign.fromPtr, but I wonder if an easier method can be provided.

[sjshuck]

Finally got around to reading https://en.wikipedia.org/wiki/UTF-16#Description. I believe Foreign.lengthArray0 (0 :: CUShort) will work to get the length of the raw string. Anyway text is UTF-8 now, and so is my case, and the logic is the same per https://en.wikipedia.org/wiki/UTF-8#Encoding, ergo Foreign.lengthArray0 (0 :: CUChar).

Note that Foreign.lengthArray0 works via a loop, and is only "easier" in that I don't have to implement the loop myself. But a more efficient solution would use strlen(). I see that that C function is already being foreign-imported, but not exposed to the user in any way.

Respectfully I'd like the "question" label removed from this issue and a fromPtr variant added that uses strlen() to slurp a region of the heap, and/or a function that performs UTF-8 validation while also scanning for NULL, but that implements this issue's title.

[Bodigrim]

@sjshuck isn't cstringLength# enough for your purposes?

[sjshuck]

Sure. I didn't look at how it's implemented; the docs don't discuss; I'd expect it to be something like strlen(). In any case I'd like it if text exposed a higher-level function that didn't require me to dip into GHC.Exts and MagicHash.

Asking directly: Would you be opposed to

fromPtr0 :: Ptr Word8 -> IO Text
fromPtr0 ptr = do
    len <- c_strlen $ castPtr ptr
    fromPtr ptr (fromIntegral len)

?

[Bodigrim]

Looks good to me, feel free to raise a PR.