5950717515

Committed 23 Aug 2023 11:30AM UTC coverage: 23.255% (-0.8%) from 24.028%

Build # 5950717515

Build Type

push

github

Committed by

neon-sunset

Commit Message

feat: implement .Replace() variants, clean up searching overloads

Run Details

154 of 856 branches covered (17.99%)

Branch coverage included in aggregate %.

86 of 86 new or added lines in 4 files covered. (100.0%)

499 of 1952 relevant lines covered (25.56%)

42232.97 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

0.0

/src/U8String.Manipulation.cs

using System.Buffers;
using System.Diagnostics.CodeAnalysis;
using System.Runtime.InteropServices;
using System.Text;
using U8Primitives.InteropServices;

namespace U8Primitives;

#pragma warning disable IDE0046, IDE0057 // Why: range slicing and ternary expressions do not produce desired codegen
public readonly partial struct U8String
{
    // TODO: Optimize/deduplicate Concat variants
    // TODO: Investigate if it is possible fold validation for u8 literals
    public static U8String Concat(U8String left, U8String right)
    {
        if (!left.IsEmpty)
        {
            if (!right.IsEmpty)
            {
                return U8Manipulation.ConcatUnchecked(
                    left.UnsafeSpan,
                    right.UnsafeSpan);
            }

            return left;
        }

        return right;
    }

    public static U8String Concat(U8String left, ReadOnlySpan<byte> right)
    {
        if (!right.IsEmpty)
        {
            Validate(right);
            if (!left.IsEmpty)
            {
                return U8Manipulation.ConcatUnchecked(left.UnsafeSpan, right);
            }

            return new U8String(right, skipValidation: true);
        }

        return left;
    }

    public static U8String Concat(ReadOnlySpan<byte> left, U8String right)
    {
        if (!left.IsEmpty)
        {
            Validate(left);
            if (!right.IsEmpty)
            {
                return U8Manipulation.ConcatUnchecked(left, right.UnsafeSpan);
            }

            return new U8String(left, skipValidation: true);
        }

        return right;
    }

    public static U8String Concat(ReadOnlySpan<byte> left, ReadOnlySpan<byte> right)
    {
        var length = left.Length + right.Length;
        if (length != 0)
        {
            var value = new byte[length];

            left.CopyTo(value);
            right.CopyTo(value.SliceUnsafe(left.Length, right.Length));

            Validate(value);
            return new U8String(value, 0, length);
        }

        return default;
    }

    /// <inheritdoc />
    public void CopyTo(byte[] destination, int index)
    {
        var src = this;
        var dst = destination.AsSpan()[index..];
        if (src.Length > dst.Length)
        {
            ThrowHelpers.ArgumentOutOfRange(nameof(index));
        }

        src.UnsafeSpan.CopyTo(dst);
    }

    /// <summary>
    /// Normalizes current <see cref="U8String"/> to the specified Unicode normalization form (default: <see cref="NormalizationForm.FormC"/>).
    /// </summary>
    /// <returns>A new <see cref="U8String"/> normalized to the specified form.</returns>
    public U8String Normalize(NormalizationForm form = NormalizationForm.FormC)
    {
        throw new NotImplementedException();
    }

    [MethodImpl(MethodImplOptions.AggressiveInlining)]
    public U8String Replace(byte oldValue, byte newValue)
    {
        return U8Manipulation.Replace(this, oldValue, newValue);
    }

    [MethodImpl(MethodImplOptions.AggressiveInlining)]
    public U8String Replace(char oldValue, char newValue)
    {
        return U8Manipulation.Replace(this, oldValue, newValue);
    }

    [MethodImpl(MethodImplOptions.AggressiveInlining)]
    public U8String Replace(Rune oldValue, Rune newValue)
    {
        return U8Manipulation.Replace(this, oldValue, newValue);
    }

    [MethodImpl(MethodImplOptions.AggressiveInlining)]
    public U8String Replace(ReadOnlySpan<byte> oldValue, ReadOnlySpan<byte> newValue)
    {
        return U8Manipulation.Replace(this, oldValue, newValue);
    }

    [MethodImpl(MethodImplOptions.AggressiveInlining)]
    public U8String Replace(U8String oldValue, U8String newValue)
    {
        return U8Manipulation.Replace(this, oldValue, newValue);
    }

    public U8String ReplaceLineEndings()
    {
        var source = this;
        if (!source.IsEmpty)
        {
            if (!OperatingSystem.IsWindows())
            {
                return U8Manipulation.ReplaceCore(
                    source, "\r\n"u8, "\n"u8, validate: false);
            }

            throw new NotImplementedException();
        }

        return source;
    }

    /// <summary>
    /// Retrieves a substring from this instance. The substring starts at a specified
    /// character position and continues to the end of the string.
    /// </summary>
    /// <param name="start">The zero-based starting character position of a substring in this instance.</param>
    /// <returns>A substring view that begins at <paramref name="start"/>.</returns>
    /// <exception cref="ArgumentOutOfRangeException">
    /// <paramref name="start"/> is less than zero or greater than the length of this instance.
    /// </exception>
    /// <exception cref="ArgumentException">
    /// The resulting substring splits at a UTF-8 code point boundary and would result in an invalid UTF-8 string.
    /// </exception>
    public U8String Slice(int start)
    {
        var source = this;
        // From ReadOnly/Span<T> Slice(int) implementation
        if ((ulong)(uint)start > (ulong)(uint)source.Length)
        {
            ThrowHelpers.ArgumentOutOfRange();
        }

        var length = source.Length - start;
        if (length > 0)
        {
            if (U8Info.IsContinuationByte(in source.UnsafeRefAdd(start)))
            {
                ThrowHelpers.InvalidSplit();
            }

            return new(source._value, source.Offset + start, length);
        }

        return default;
    }

    /// <summary>
    /// Retrieves a substring from this instance. The substring starts at a specified
    /// character position and has a specified length.
    /// </summary>
    /// <param name="start">The zero-based starting character position of a substring in this instance.</param>
    /// <param name="length">The number of bytes in the substring.</param>
    /// <returns>A substring view that begins at <paramref name="start"/> and has <paramref name="length"/> bytes.</returns>
    /// <exception cref="ArgumentOutOfRangeException">
    /// <paramref name="start"/> or <paramref name="length"/> is less than zero, or the sum of <paramref name="start"/> and <paramref name="length"/> is greater than the length of the current instance.
    /// </exception>
    /// <exception cref="ArgumentException">
    /// The resulting substring splits at a UTF-8 code point boundary and would result in an invalid UTF-8 string.
    /// </exception>
    public U8String Slice(int start, int length)
    {
        var source = this;
        // From ReadOnly/Span<T> Slice(int, int) implementation
        if ((ulong)(uint)start + (ulong)(uint)length > (ulong)(uint)source.Length)
        {
            ThrowHelpers.ArgumentOutOfRange();
        }

        var result = default(U8String);
        if (length > 0)
        {
            // TODO: Is there really no way to get rid of length < source.Length when checking the last+1 byte?
            if ((start > 0 && U8Info.IsContinuationByte(source.UnsafeRefAdd(start))) || (
                length < source.Length && U8Info.IsContinuationByte(source.UnsafeRefAdd(start + length))))
            {
                // TODO: Exception message UX
                ThrowHelpers.InvalidSplit();
            }

            result = new(source._value, source.Offset + start, length);
        }

        return result;
    }

    /// <summary>
    /// Removes all leading and trailing whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all whitespace characters
    /// are removed from the start and end of the current string.
    /// </returns>
    public U8String Trim()
    {
        // TODO: Optimize fast path on no whitespace
        // TODO 2: Do not convert to runes and have proper
        // whitespace LUT to evaluate code points in a branchless way
        var source = this;
        if (!source.IsEmpty)
        {
            ref var ptr = ref source.UnsafeRef;

            var start = 0;
            while (start < source.Length)
            {
                if (!U8Info.IsWhitespaceRune(ref ptr.Add(start), out var size))
                {
                    break;
                }
                start += size;
            }

            var end = source.Length - 1;
            for (var endSearch = end; endSearch >= start; endSearch--)
            {
                var b = ptr.Add(endSearch);
                if (!U8Info.IsContinuationByte(b))
                {
                    if (U8Info.IsAsciiByte(b)
                        ? U8Info.IsAsciiWhitespace(b)
                        : U8Info.IsNonAsciiWhitespace(ref ptr.Add(end), out _))
                    {
                        // Save the last found whitespace code point offset and continue searching
                        // for more whitspace byte sequences from their end. If we don't do this,
                        // we will end up trimming away continuation bytes at the end of the string.
                        end = endSearch - 1;
                    }
                    else
                    {
                        break;
                    }
                }
            }

            return U8Marshal.Slice(source, start, end - start + 1);
        }

        return default;
    }

    /// <summary>
    /// Removes all leading whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all whitespace characters
    /// are removed from the start of the current string.
    /// </returns>
    public U8String TrimStart()
    {
        var source = this;
        if (!source.IsEmpty)
        {
            ref var ptr = ref source.UnsafeRef;
            var b = ptr;

            if (U8Info.IsAsciiByte(b) && !U8Info.IsAsciiWhitespace(b))
            {
                return source;
            }

            var start = 0;
            while (start < source.Length)
            {
                if (!U8Info.IsWhitespaceRune(ref ptr.Add(start), out var size))
                {
                    break;
                }
                start += size;
            }

            return U8Marshal.Slice(source, start);
        }

        return default;
    }

    /// <summary>
    /// Removes all trailing whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all whitespace characters
    /// are removed from the end of the current string.
    /// </returns>
    public U8String TrimEnd()
    {
        var source = this;
        if (!source.IsEmpty)
        {
            ref var ptr = ref source.UnsafeRef;

            var end = source.Length - 1;
            for (var endSearch = end; endSearch >= 0; endSearch--)
            {
                var b = ptr.Add(endSearch);
                if (!U8Info.IsContinuationByte(b))
                {
                    if (U8Info.IsAsciiByte(b)
                        ? U8Info.IsAsciiWhitespace(b)
                        : U8Info.IsNonAsciiWhitespace(ref ptr.Add(end), out _))
                    {
                        end = endSearch - 1;
                    }
                    else
                    {
                        break;
                    }
                }
            }

            return U8Marshal.Slice(source, 0, end + 1);
        }

        return default;
    }

    /// <summary>
    /// Removes all leading and trailing ASCII whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all ASCII whitespace characters
    /// are removed from the start and end of the current string.
    /// </returns>
    public U8String TrimAscii()
    {
        var source = this;
        var range = Ascii.Trim(source);

        return U8Marshal.Slice(source, range);
    }

    /// <summary>
    /// Removes all the leading ASCII whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all whitespace characters
    /// are removed from the start of the current string.
    /// </returns>
    public U8String TrimStartAscii()
    {
        var source = this;
        var range = Ascii.TrimStart(source);

        return U8Marshal.Slice(source, range);
    }

    /// <summary>
    /// Removes all the trailing ASCII whitespace characters from the current string.
    /// </summary>
    /// <returns>
    /// A sub-slice that remains after all whitespace characters
    /// are removed from the end of the current string.
    /// </returns>
    public U8String TrimEndAscii()
    {
        var source = this;
        var range = Ascii.TrimEnd(source);

        return U8Marshal.Slice(source, range);
    }

    // TODO:
    // - Complete impl. depends on porting of InlineArray-based array builder for letters
    // which have different lengths in upper/lower case.
    // - Remove/rename to ToLowerFallback or move to something like "FallbackInvariantComparer"
    // clearly indicating it being slower and inferior alternative to proper implementations
    // which call into ICU/NLS/Hybrid-provided case change exports.
    public U8String ToLower()
    {
        var source = this;
        if (source.Length > 0)
        {
            var lowercase = new byte[source.Length + 3];
            var destination = lowercase.AsSpan();
            ref var dst = ref destination.AsRef();

            var result = Ascii.ToLower(source, destination, out var consumed);
            if (result is OperationStatus.InvalidData)
            {
                foreach (var rune in U8Marshal.Slice(source, consumed).Runes)
                {
                    var lower = Rune.ToLowerInvariant(rune);
                    var scalar = U8Scalar.Create(lower);
                    if (consumed + 4 > destination.Length)
                    {
                        [DoesNotReturn]
                        static void Unimpl()
                        {
                            throw new NotImplementedException();
                        }

                        Unimpl();
                    }

                    scalar.StoreUnsafe(ref dst.Add(consumed));
                    consumed += scalar.Size;
                }
            }

            return new(lowercase, 0, consumed);
        }

        return default;
    }

    public U8String ToUpper()
    {
        var source = this;
        if (source.Length > 0)
        {
            var uppercase = new byte[source.Length + 3];
            var destination = uppercase.AsSpan();
            ref var dst = ref destination.AsRef();

            var result = Ascii.ToUpper(source, destination, out var consumed);
            if (result is OperationStatus.InvalidData)
            {
                foreach (var rune in U8Marshal.Slice(source, consumed).Runes)
                {
                    var upper = Rune.ToUpperInvariant(rune);
                    var scalar = U8Scalar.Create(upper);
                    if (consumed + 4 > destination.Length)
                    {
                        [DoesNotReturn]
                        static void Unimpl()
                        {
                            throw new NotImplementedException();
                        }

                        Unimpl();
                    }

                    scalar.StoreUnsafe(ref dst.Add(consumed));
                    consumed += scalar.Size;
                }
            }

            return new(uppercase, 0, consumed);
        }

        return default;
    }

    // TODO: docs
    public U8String ToLowerAscii()
    {
        var source = this;
        if (source.Length > 0)
        {
            var destination = new byte[source.Length];

            U8Manipulation.ToLowerAscii(
                ref source.UnsafeRef,
                ref MemoryMarshal.GetArrayDataReference(destination),
                (uint)source.Length);

            return new(destination, 0, source.Length);
        }

        return default;
    }

    public U8String ToUpperAscii()
    {
        var source = this;
        if (source.Length > 0)
        {
            var destination = new byte[source.Length];

            U8Manipulation.ToUpperAscii(
                ref source.UnsafeRef,
                ref MemoryMarshal.GetArrayDataReference(destination),
                (uint)source.Length);

            return new(destination, 0, source.Length);
        }

        return default;
    }
}

1	using System.Buffers;
2	using System.Diagnostics.CodeAnalysis;
3	using System.Runtime.InteropServices;
4	using System.Text;
5	using U8Primitives.InteropServices;
6
7	namespace U8Primitives;
8
9	#pragma warning disable IDE0046, IDE0057 // Why: range slicing and ternary expressions do not produce desired codegen
10	public readonly partial struct U8String
11	{
12	// TODO: Optimize/deduplicate Concat variants
13	// TODO: Investigate if it is possible fold validation for u8 literals
14	public static U8String Concat(U8String left, U8String right)
15	{
16	if (!left.IsEmpty)	×
17	{
18	if (!right.IsEmpty)	×
19	{
20	return U8Manipulation.ConcatUnchecked(	×
21	left.UnsafeSpan,	×
22	right.UnsafeSpan);	×
23	}
24
25	return left;	×
26	}
27
28	return right;	×
29	}
30
31	public static U8String Concat(U8String left, ReadOnlySpan<byte> right)
32	{
33	if (!right.IsEmpty)	×
34	{
35	Validate(right);	×
36	if (!left.IsEmpty)	×
37	{
38	return U8Manipulation.ConcatUnchecked(left.UnsafeSpan, right);	×
39	}
40
41	return new U8String(right, skipValidation: true);	×
42	}
43
44	return left;	×
45	}
46
47	public static U8String Concat(ReadOnlySpan<byte> left, U8String right)
48	{
49	if (!left.IsEmpty)	×
50	{
51	Validate(left);	×
52	if (!right.IsEmpty)	×
53	{
54	return U8Manipulation.ConcatUnchecked(left, right.UnsafeSpan);	×
55	}
56
57	return new U8String(left, skipValidation: true);	×
58	}
59
60	return right;	×
61	}
62
63	public static U8String Concat(ReadOnlySpan<byte> left, ReadOnlySpan<byte> right)
64	{
65	var length = left.Length + right.Length;	×
66	if (length != 0)	×
67	{
68	var value = new byte[length];	×
69
70	left.CopyTo(value);	×
71	right.CopyTo(value.SliceUnsafe(left.Length, right.Length));	×
72
73	Validate(value);	×
74	return new U8String(value, 0, length);	×
75	}
76
77	return default;	×
78	}
79
80	/// <inheritdoc />
81	public void CopyTo(byte[] destination, int index)
82	{
83	var src = this;	×
84	var dst = destination.AsSpan()[index..];	×
85	if (src.Length > dst.Length)	×
86	{
87	ThrowHelpers.ArgumentOutOfRange(nameof(index));	×
88	}
89
90	src.UnsafeSpan.CopyTo(dst);	×
91	}	×
92
93	/// <summary>
94	/// Normalizes current <see cref="U8String"/> to the specified Unicode normalization form (default: <see cref="NormalizationForm.FormC"/>).
95	/// </summary>
96	/// <returns>A new <see cref="U8String"/> normalized to the specified form.</returns>
97	public U8String Normalize(NormalizationForm form = NormalizationForm.FormC)
98	{
99	throw new NotImplementedException();	×
100	}
101
102	[MethodImpl(MethodImplOptions.AggressiveInlining)]
103	public U8String Replace(byte oldValue, byte newValue)
104	{
105	return U8Manipulation.Replace(this, oldValue, newValue);	×
106	}
107
108	[MethodImpl(MethodImplOptions.AggressiveInlining)]
109	public U8String Replace(char oldValue, char newValue)
110	{
111	return U8Manipulation.Replace(this, oldValue, newValue);	×
112	}
113
114	[MethodImpl(MethodImplOptions.AggressiveInlining)]
115	public U8String Replace(Rune oldValue, Rune newValue)
116	{
117	return U8Manipulation.Replace(this, oldValue, newValue);	×
118	}
119
120	[MethodImpl(MethodImplOptions.AggressiveInlining)]
121	public U8String Replace(ReadOnlySpan<byte> oldValue, ReadOnlySpan<byte> newValue)
122	{
123	return U8Manipulation.Replace(this, oldValue, newValue);	×
124	}
125
126	[MethodImpl(MethodImplOptions.AggressiveInlining)]
127	public U8String Replace(U8String oldValue, U8String newValue)
128	{
129	return U8Manipulation.Replace(this, oldValue, newValue);	×
130	}
131
132	public U8String ReplaceLineEndings()
133	{
134	var source = this;	×
135	if (!source.IsEmpty)	×
136	{
137	if (!OperatingSystem.IsWindows())	×
138	{
139	return U8Manipulation.ReplaceCore(	×
140	source, "\r\n"u8, "\n"u8, validate: false);	×
141	}
142
143	throw new NotImplementedException();	×
144	}
145
146	return source;	×
147	}
148
149	/// <summary>
150	/// Retrieves a substring from this instance. The substring starts at a specified
151	/// character position and continues to the end of the string.
152	/// </summary>
153	/// <param name="start">The zero-based starting character position of a substring in this instance.</param>
154	/// <returns>A substring view that begins at <paramref name="start"/>.</returns>
155	/// <exception cref="ArgumentOutOfRangeException">
156	/// <paramref name="start"/> is less than zero or greater than the length of this instance.
157	/// </exception>
158	/// <exception cref="ArgumentException">
159	/// The resulting substring splits at a UTF-8 code point boundary and would result in an invalid UTF-8 string.
160	/// </exception>
161	public U8String Slice(int start)
162	{
163	var source = this;	×
164	// From ReadOnly/Span<T> Slice(int) implementation
165	if ((ulong)(uint)start > (ulong)(uint)source.Length)	×
166	{
167	ThrowHelpers.ArgumentOutOfRange();	×
168	}
169
170	var length = source.Length - start;	×
171	if (length > 0)	×
172	{
173	if (U8Info.IsContinuationByte(in source.UnsafeRefAdd(start)))	×
174	{
175	ThrowHelpers.InvalidSplit();	×
176	}
177
178	return new(source._value, source.Offset + start, length);	×
179	}
180
181	return default;	×
182	}
183
184	/// <summary>
185	/// Retrieves a substring from this instance. The substring starts at a specified
186	/// character position and has a specified length.
187	/// </summary>
188	/// <param name="start">The zero-based starting character position of a substring in this instance.</param>
189	/// <param name="length">The number of bytes in the substring.</param>
190	/// <returns>A substring view that begins at <paramref name="start"/> and has <paramref name="length"/> bytes.</returns>
191	/// <exception cref="ArgumentOutOfRangeException">
192	/// <paramref name="start"/> or <paramref name="length"/> is less than zero, or the sum of <paramref name="start"/> and <paramref name="length"/> is greater than the length of the current instance.
193	/// </exception>
194	/// <exception cref="ArgumentException">
195	/// The resulting substring splits at a UTF-8 code point boundary and would result in an invalid UTF-8 string.
196	/// </exception>
197	public U8String Slice(int start, int length)
198	{
199	var source = this;	×
200	// From ReadOnly/Span<T> Slice(int, int) implementation
201	if ((ulong)(uint)start + (ulong)(uint)length > (ulong)(uint)source.Length)	×
202	{
203	ThrowHelpers.ArgumentOutOfRange();	×
204	}
205
206	var result = default(U8String);	×
207	if (length > 0)	×
208	{
209	// TODO: Is there really no way to get rid of length < source.Length when checking the last+1 byte?
210	if ((start > 0 && U8Info.IsContinuationByte(source.UnsafeRefAdd(start))) \|\| (	×
211	length < source.Length && U8Info.IsContinuationByte(source.UnsafeRefAdd(start + length))))	×
212	{
213	// TODO: Exception message UX
214	ThrowHelpers.InvalidSplit();	×
215	}
216
217	result = new(source._value, source.Offset + start, length);	×
218	}
219
220	return result;	×
221	}
222
223	/// <summary>
224	/// Removes all leading and trailing whitespace characters from the current string.
225	/// </summary>
226	/// <returns>
227	/// A sub-slice that remains after all whitespace characters
228	/// are removed from the start and end of the current string.
229	/// </returns>
230	public U8String Trim()
231	{
232	// TODO: Optimize fast path on no whitespace
233	// TODO 2: Do not convert to runes and have proper
234	// whitespace LUT to evaluate code points in a branchless way
235	var source = this;	×
236	if (!source.IsEmpty)	×
237	{
238	ref var ptr = ref source.UnsafeRef;	×
239
240	var start = 0;	×
241	while (start < source.Length)	×
242	{
243	if (!U8Info.IsWhitespaceRune(ref ptr.Add(start), out var size))	×
244	{
245	break;
246	}
247	start += size;	×
248	}
249
250	var end = source.Length - 1;	×
251	for (var endSearch = end; endSearch >= start; endSearch--)	×
252	{
253	var b = ptr.Add(endSearch);	×
254	if (!U8Info.IsContinuationByte(b))	×
255	{
256	if (U8Info.IsAsciiByte(b)	×
257	? U8Info.IsAsciiWhitespace(b)	×
258	: U8Info.IsNonAsciiWhitespace(ref ptr.Add(end), out _))	×
259	{
260	// Save the last found whitespace code point offset and continue searching
261	// for more whitspace byte sequences from their end. If we don't do this,
262	// we will end up trimming away continuation bytes at the end of the string.
263	end = endSearch - 1;	×
264	}
265	else
266	{
267	break;
268	}
269	}
270	}
271
272	return U8Marshal.Slice(source, start, end - start + 1);	×
273	}
274
275	return default;	×
276	}
277
278	/// <summary>
279	/// Removes all leading whitespace characters from the current string.
280	/// </summary>
281	/// <returns>
282	/// A sub-slice that remains after all whitespace characters
283	/// are removed from the start of the current string.
284	/// </returns>
285	public U8String TrimStart()
286	{
287	var source = this;	×
288	if (!source.IsEmpty)	×
289	{
290	ref var ptr = ref source.UnsafeRef;	×
291	var b = ptr;	×
292
293	if (U8Info.IsAsciiByte(b) && !U8Info.IsAsciiWhitespace(b))	×
294	{
295	return source;	×
296	}
297
298	var start = 0;	×
299	while (start < source.Length)	×
300	{
301	if (!U8Info.IsWhitespaceRune(ref ptr.Add(start), out var size))	×
302	{
303	break;
304	}
305	start += size;	×
306	}
307
308	return U8Marshal.Slice(source, start);	×
309	}
310
311	return default;	×
312	}
313
314	/// <summary>
315	/// Removes all trailing whitespace characters from the current string.
316	/// </summary>
317	/// <returns>
318	/// A sub-slice that remains after all whitespace characters
319	/// are removed from the end of the current string.
320	/// </returns>
321	public U8String TrimEnd()
322	{
323	var source = this;	×
324	if (!source.IsEmpty)	×
325	{
326	ref var ptr = ref source.UnsafeRef;	×
327
328	var end = source.Length - 1;	×
329	for (var endSearch = end; endSearch >= 0; endSearch--)	×
330	{
331	var b = ptr.Add(endSearch);	×
332	if (!U8Info.IsContinuationByte(b))	×
333	{
334	if (U8Info.IsAsciiByte(b)	×
335	? U8Info.IsAsciiWhitespace(b)	×
336	: U8Info.IsNonAsciiWhitespace(ref ptr.Add(end), out _))	×
337	{
338	end = endSearch - 1;	×
339	}
340	else
341	{
342	break;
343	}
344	}
345	}
346
347	return U8Marshal.Slice(source, 0, end + 1);	×
348	}
349
350	return default;	×
351	}
352
353	/// <summary>
354	/// Removes all leading and trailing ASCII whitespace characters from the current string.
355	/// </summary>
356	/// <returns>
357	/// A sub-slice that remains after all ASCII whitespace characters
358	/// are removed from the start and end of the current string.
359	/// </returns>
360	public U8String TrimAscii()
361	{
362	var source = this;	×
363	var range = Ascii.Trim(source);	×
364
365	return U8Marshal.Slice(source, range);	×
366	}
367
368	/// <summary>
369	/// Removes all the leading ASCII whitespace characters from the current string.
370	/// </summary>
371	/// <returns>
372	/// A sub-slice that remains after all whitespace characters
373	/// are removed from the start of the current string.
374	/// </returns>
375	public U8String TrimStartAscii()
376	{
377	var source = this;	×
378	var range = Ascii.TrimStart(source);	×
379
380	return U8Marshal.Slice(source, range);	×
381	}
382
383	/// <summary>
384	/// Removes all the trailing ASCII whitespace characters from the current string.
385	/// </summary>
386	/// <returns>
387	/// A sub-slice that remains after all whitespace characters
388	/// are removed from the end of the current string.
389	/// </returns>
390	public U8String TrimEndAscii()
391	{
392	var source = this;	×
393	var range = Ascii.TrimEnd(source);	×
394
395	return U8Marshal.Slice(source, range);	×
396	}
397
398	// TODO:
399	// - Complete impl. depends on porting of InlineArray-based array builder for letters
400	// which have different lengths in upper/lower case.
401	// - Remove/rename to ToLowerFallback or move to something like "FallbackInvariantComparer"
402	// clearly indicating it being slower and inferior alternative to proper implementations
403	// which call into ICU/NLS/Hybrid-provided case change exports.
404	public U8String ToLower()
405	{
406	var source = this;	×
407	if (source.Length > 0)	×
408	{
409	var lowercase = new byte[source.Length + 3];	×
410	var destination = lowercase.AsSpan();	×
411	ref var dst = ref destination.AsRef();	×
412
413	var result = Ascii.ToLower(source, destination, out var consumed);	×
414	if (result is OperationStatus.InvalidData)	×
415	{
416	foreach (var rune in U8Marshal.Slice(source, consumed).Runes)	×
417	{
418	var lower = Rune.ToLowerInvariant(rune);	×
419	var scalar = U8Scalar.Create(lower);	×
420	if (consumed + 4 > destination.Length)	×
421	{
422	[DoesNotReturn]
423	static void Unimpl()
424	{
425	throw new NotImplementedException();	×
426	}
427
428	Unimpl();	×
429	}
430
431	scalar.StoreUnsafe(ref dst.Add(consumed));	×
432	consumed += scalar.Size;	×
433	}
434	}
435
436	return new(lowercase, 0, consumed);	×
437	}
438
439	return default;	×
440	}
441
442	public U8String ToUpper()
443	{
444	var source = this;	×
445	if (source.Length > 0)	×
446	{
447	var uppercase = new byte[source.Length + 3];	×
448	var destination = uppercase.AsSpan();	×
449	ref var dst = ref destination.AsRef();	×
450
451	var result = Ascii.ToUpper(source, destination, out var consumed);	×
452	if (result is OperationStatus.InvalidData)	×
453	{
454	foreach (var rune in U8Marshal.Slice(source, consumed).Runes)	×
455	{
456	var upper = Rune.ToUpperInvariant(rune);	×
457	var scalar = U8Scalar.Create(upper);	×
458	if (consumed + 4 > destination.Length)	×
459	{
460	[DoesNotReturn]
461	static void Unimpl()
462	{
463	throw new NotImplementedException();	×
464	}
465
466	Unimpl();	×
467	}
468
469	scalar.StoreUnsafe(ref dst.Add(consumed));	×
470	consumed += scalar.Size;	×
471	}
472	}
473
474	return new(uppercase, 0, consumed);	×
475	}
476
477	return default;	×
478	}
479
480	// TODO: docs
481	public U8String ToLowerAscii()
482	{
483	var source = this;	×
484	if (source.Length > 0)	×
485	{
486	var destination = new byte[source.Length];	×
487
488	U8Manipulation.ToLowerAscii(	×
489	ref source.UnsafeRef,	×
490	ref MemoryMarshal.GetArrayDataReference(destination),	×
491	(uint)source.Length);	×
492
493	return new(destination, 0, source.Length);	×
494	}
495
496	return default;	×
497	}
498
499	public U8String ToUpperAscii()
500	{
501	var source = this;	×
502	if (source.Length > 0)	×
503	{
504	var destination = new byte[source.Length];	×
505
506	U8Manipulation.ToUpperAscii(	×
507	ref source.UnsafeRef,	×
508	ref MemoryMarshal.GetArrayDataReference(destination),	×
509	(uint)source.Length);	×
510
511	return new(destination, 0, source.Length);	×
512	}
513
514	return default;	×
515	}
516	}

neon-sunset / U8String / 5950717515

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous