Documents need attachments and generated files (SPEC 9.4) before the UI of M3 can show them, and the spec forbids external libraries for PDF and spreadsheets. - VolumeFileStore keeps files under <data>/files/<COMPANY>/<yyyy>/<guid>, writing a .part file in the same folder with size limit and SHA-256, then quota and scanner checks (IFileScanner hook, no-op by default). - attachments table (migration 007) with kind, version series, lock, trash and the uploading user; AttachmentService uploads, adds generated files as the next version, lists, opens, zips, locks, trashes, restores and purges after 30 days (new system job system.attachments-trash). - FileGenerationService runs IFileGenerator implementations into the store; three generators for the fictitious test document (PDF, CSV, XLSX). - CsvWriter (Italian Excel defaults, formula-injection guard), XlsxWriter, ZipStreamWriter and PdfWriter (standard Type1 fonts, WinAnsi, AFM widths, JPEG, Flate, xref) only write asynchronously, so they can stream to an HTTP response. - ADR 0021 records versioning, storage layout, PDF scope and CSV defaults. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
212 lines
9.7 KiB
C#
212 lines
9.7 KiB
C#
using System.Globalization;
|
|
using System.IO.Compression;
|
|
using System.Text;
|
|
using System.Text.RegularExpressions;
|
|
using Khaemhat.Infrastructure.Files;
|
|
|
|
namespace Khaemhat.Infrastructure.Tests.Files;
|
|
|
|
/// <summary>The in-house PDF writer, first version (SPEC §2 "PDF e file", issue #34).</summary>
|
|
[TestClass]
|
|
public sealed partial class PdfWriterTests
|
|
{
|
|
private static readonly Encoding Latin1 = Encoding.Latin1;
|
|
|
|
[TestMethod]
|
|
public async Task Document_HasHeaderTrailerAndXrefOffsetsThatPointAtTheObjects()
|
|
{
|
|
var bytes = await WriteAsync(async pdf =>
|
|
{
|
|
var first = await pdf.AddPageAsync();
|
|
first.DrawText("Prima pagina", 50, 780, PdfFont.HelveticaBold, 16);
|
|
first.DrawLine(50, 770, 545, 770, 1, PdfColor.FromHex("#1F5F8B"));
|
|
first.DrawRectangle(50, 700, 200, 40, PdfColor.Black, PdfColor.Gray(230));
|
|
var second = await pdf.AddPageAsync(PdfWriter.A4Height, PdfWriter.A4Width);
|
|
second.DrawTextRight("€ 1.234,56", 800, 500, PdfFont.TimesRoman, 10);
|
|
await pdf.AddPageAsync();
|
|
Assert.AreEqual(3, pdf.PageCount);
|
|
});
|
|
var text = Latin1.GetString(bytes);
|
|
|
|
StringAssert.StartsWith(text, "%PDF-1.7\n");
|
|
StringAssert.EndsWith(text, "%%EOF\n");
|
|
StringAssert.Contains(text, "/Type /Pages /Kids [");
|
|
StringAssert.Contains(text, "/Count 3");
|
|
Assert.AreEqual(3, PageObject().Count(text));
|
|
StringAssert.Contains(text, "/MediaBox [0 0 841.89 595.28]");
|
|
|
|
var startxref = long.Parse(StartXref().Match(text).Groups[1].Value, CultureInfo.InvariantCulture);
|
|
Assert.AreEqual("xref", text.Substring((int)startxref, 4));
|
|
var header = XrefHeader().Match(text, (int)startxref);
|
|
var size = int.Parse(header.Groups[1].Value, CultureInfo.InvariantCulture);
|
|
var entries = XrefEntry().Matches(text, header.Index + header.Length).Take(size).ToList();
|
|
Assert.HasCount(size, entries);
|
|
Assert.AreEqual("f", entries[0].Groups[3].Value);
|
|
for (var i = 1; i < size; i++)
|
|
{
|
|
var offset = int.Parse(entries[i].Groups[1].Value, CultureInfo.InvariantCulture);
|
|
StringAssert.StartsWith(text[offset..], $"{i} 0 obj\n", $"xref entry {i}");
|
|
}
|
|
|
|
StringAssert.Contains(text, $"/Size {size} /Root 1 0 R /Info 3 0 R");
|
|
}
|
|
|
|
[TestMethod]
|
|
public async Task Text_IsWinAnsiEncoded_EuroAndAccentsIncluded()
|
|
{
|
|
var bytes = await WriteAsync(
|
|
async pdf =>
|
|
{
|
|
var page = await pdf.AddPageAsync();
|
|
page.DrawText("Totale € città (prova) \\ fine", 50, 700, PdfFont.Helvetica, 12, PdfColor.FromHex("#FF0000"));
|
|
},
|
|
compress: false);
|
|
var text = Latin1.GetString(bytes);
|
|
|
|
// € is 0x80 (octal 200), à is 0xE0 (octal 340); parentheses and backslash are escaped.
|
|
StringAssert.Contains(text, "(Totale \\200 citt\\340 \\(prova\\) \\\\ fine) Tj");
|
|
StringAssert.Contains(text, "1 0 0 rg BT /F1 12 Tf 1 0 0 1 50 700 Tm");
|
|
StringAssert.Contains(text, "/BaseFont /Helvetica /Encoding /WinAnsiEncoding");
|
|
CollectionAssert.AreEqual(new byte[] { 0x80, 0xE0, 0xE8, 0xE9, 0xEC, 0xF2, 0xF9, (byte)'?' }, PdfTextLayout.EncodeWinAnsi("€àèéìòù漢"));
|
|
}
|
|
|
|
[TestMethod]
|
|
public async Task ContentStreams_AreFlateCompressed()
|
|
{
|
|
var bytes = await WriteAsync(async pdf => (await pdf.AddPageAsync()).DrawText("Compresso è meglio", 50, 700, PdfFont.Courier, 10));
|
|
var text = Latin1.GetString(bytes);
|
|
var match = FlateStream().Match(text);
|
|
Assert.IsTrue(match.Success);
|
|
|
|
var length = int.Parse(match.Groups[1].Value, CultureInfo.InvariantCulture);
|
|
var start = match.Index + match.Length;
|
|
using var zlib = new ZLibStream(new MemoryStream(bytes, start, length), CompressionMode.Decompress);
|
|
using var reader = new StreamReader(zlib, Latin1);
|
|
StringAssert.Contains(reader.ReadToEnd(), "(Compresso \\350 meglio) Tj");
|
|
}
|
|
|
|
[TestMethod]
|
|
public async Task DocumentInfo_IsWritten_WithUnicodeAsUtf16()
|
|
{
|
|
var info = new PdfDocumentInfo { Title = "Fattura n. 7", Author = "Società Prova", CreationDate = new DateTimeOffset(2026, 11, 11, 10, 30, 0, TimeSpan.Zero) };
|
|
var output = new AsyncOnlyStream();
|
|
await using (var pdf = new PdfWriter(output, info, leaveOpen: true))
|
|
{
|
|
await pdf.AddPageAsync();
|
|
}
|
|
|
|
var text = Latin1.GetString(output.ToArray());
|
|
StringAssert.Contains(text, "/Title (Fattura n. 7)");
|
|
StringAssert.Contains(text, "/Author <FEFF0053006F0063006900650074" + "00E0");
|
|
StringAssert.Contains(text, "/Creator (Khaemhat)");
|
|
StringAssert.Contains(text, "/CreationDate (D:20261111103000Z)");
|
|
}
|
|
|
|
[TestMethod]
|
|
public async Task EmptyDocument_GetsOneBlankPage()
|
|
{
|
|
var text = Latin1.GetString(await WriteAsync(_ => Task.CompletedTask));
|
|
|
|
StringAssert.Contains(text, "/Count 1");
|
|
}
|
|
|
|
[TestMethod]
|
|
public void Jpeg_SizeAndComponents_AreReadFromTheFrameHeader()
|
|
{
|
|
var info = JpegInfo.Read(TinyJpeg(width: 64, height: 32, components: 3));
|
|
|
|
Assert.AreEqual(new JpegInfo(64, 32, 3, 8, false), info);
|
|
Assert.AreEqual(1, JpegInfo.Read(TinyJpeg(10, 10, 1)).Components);
|
|
Assert.IsTrue(JpegInfo.Read(TinyJpeg(10, 10, 4, adobe: true)).AdobeInverted);
|
|
Assert.ThrowsExactly<InvalidDataException>(() => JpegInfo.Read("not a jpeg"u8));
|
|
Assert.ThrowsExactly<InvalidDataException>(() => JpegInfo.Read(new byte[] { 0xFF, 0xD8, 0xFF, 0xD9 }));
|
|
}
|
|
|
|
[TestMethod]
|
|
public async Task Jpeg_IsEmbeddedAsIs_WithDctDecode()
|
|
{
|
|
var jpeg = TinyJpeg(64, 32, 3);
|
|
var bytes = await WriteAsync(async pdf =>
|
|
{
|
|
var image = pdf.AddJpeg(jpeg);
|
|
var page = await pdf.AddPageAsync();
|
|
page.DrawImage(image, 50, 600, 128, 64);
|
|
Assert.AreEqual(64, image.Width);
|
|
});
|
|
var text = Latin1.GetString(bytes);
|
|
|
|
StringAssert.Contains(text, $"/Subtype /Image /Width 64 /Height 32 /ColorSpace /DeviceRGB /BitsPerComponent 8 /Filter /DCTDecode /Length {jpeg.Length} >>");
|
|
StringAssert.Contains(text, "/XObject << /Im1 ");
|
|
Assert.IsGreaterThanOrEqualTo(0, text.IndexOf(Latin1.GetString(jpeg), StringComparison.Ordinal));
|
|
}
|
|
|
|
[TestMethod]
|
|
public void Measure_UsesTheAfmWidths_AndWrapKeepsLinesInsideTheWidth()
|
|
{
|
|
// Helvetica: "A" = 667, "i" = 222; Helvetica-Bold: "A" = 722; Courier: 600 for every character.
|
|
Assert.AreEqual(6.67, PdfTextLayout.Measure("A", PdfFont.Helvetica, 10), 1e-9);
|
|
Assert.AreEqual(8.89, PdfTextLayout.Measure("Ai", PdfFont.Helvetica, 10), 1e-9);
|
|
Assert.AreEqual(7.22, PdfTextLayout.Measure("A", PdfFont.HelveticaBold, 10), 1e-9);
|
|
Assert.AreEqual(5.56, PdfTextLayout.Measure("€", PdfFont.Helvetica, 10), 1e-9);
|
|
Assert.AreEqual(18, PdfTextLayout.Measure("abc", PdfFont.Courier, 10), 1e-9);
|
|
Assert.AreEqual(PdfTextLayout.Measure("e", PdfFont.Helvetica, 10), PdfTextLayout.Measure("è", PdfFont.Helvetica, 10));
|
|
|
|
var text = "Questo è un paragrafo abbastanza lungo da andare a capo più volte, con una parolaincredibilmentelunghissimachenonstainriga.";
|
|
var lines = PdfTextLayout.Wrap(text, PdfFont.Helvetica, 10, 120);
|
|
Assert.IsGreaterThan(2, lines.Count);
|
|
Assert.IsTrue(lines.All(l => PdfTextLayout.Measure(l, PdfFont.Helvetica, 10) <= 120));
|
|
Assert.AreEqual(text.Replace(" ", string.Empty, StringComparison.Ordinal), string.Concat(lines).Replace(" ", string.Empty, StringComparison.Ordinal));
|
|
Assert.HasCount(2, PdfTextLayout.Wrap("riga uno\nriga due", PdfFont.Helvetica, 10, 500));
|
|
StringAssert.EndsWith(PdfTextLayout.Fit("Descrizione molto lunga di un articolo", PdfFont.Helvetica, 10, 60), "…");
|
|
}
|
|
|
|
private static async Task<byte[]> WriteAsync(Func<PdfWriter, Task> draw, bool compress = true)
|
|
{
|
|
var output = new AsyncOnlyStream();
|
|
var pdf = new PdfWriter(output, new PdfDocumentInfo { Title = "Prova" }, leaveOpen: true) { CompressContent = compress };
|
|
await using (pdf)
|
|
{
|
|
await draw(pdf);
|
|
await pdf.CloseAsync();
|
|
}
|
|
|
|
return output.ToArray();
|
|
}
|
|
|
|
/// <summary>The header of a JPEG file built by hand: SOI, APP0 (JFIF), optional APP14 (Adobe), SOF0, EOI. No pixels.</summary>
|
|
private static byte[] TinyJpeg(int width, int height, int components, bool adobe = false)
|
|
{
|
|
var bytes = new List<byte> { 0xFF, 0xD8 };
|
|
bytes.AddRange([0xFF, 0xE0, 0x00, 0x10, (byte)'J', (byte)'F', (byte)'I', (byte)'F', 0x00, 0x01, 0x01, 0x00, 0x00, 0x01, 0x00, 0x01, 0x00, 0x00]);
|
|
if (adobe)
|
|
{
|
|
bytes.AddRange([0xFF, 0xEE, 0x00, 0x0E, (byte)'A', (byte)'d', (byte)'o', (byte)'b', (byte)'e', 0x00, 0x64, 0x00, 0x00, 0x00, 0x00, 0x02]);
|
|
}
|
|
|
|
var length = 8 + (3 * components);
|
|
bytes.AddRange([0xFF, 0xC0, (byte)(length >> 8), (byte)length, 0x08, (byte)(height >> 8), (byte)height, (byte)(width >> 8), (byte)width, (byte)components]);
|
|
for (var c = 1; c <= components; c++)
|
|
{
|
|
bytes.AddRange([(byte)c, 0x11, 0x00]);
|
|
}
|
|
|
|
bytes.AddRange([0xFF, 0xD9]);
|
|
return [.. bytes];
|
|
}
|
|
|
|
[GeneratedRegex(@"/Type /Page /Parent")]
|
|
private static partial Regex PageObject();
|
|
|
|
[GeneratedRegex(@"startxref\n(\d+)\n%%EOF")]
|
|
private static partial Regex StartXref();
|
|
|
|
[GeneratedRegex(@"xref\n0 (\d+)\n")]
|
|
private static partial Regex XrefHeader();
|
|
|
|
[GeneratedRegex(@"(\d{10}) (\d{5}) ([nf])\r\n")]
|
|
private static partial Regex XrefEntry();
|
|
|
|
[GeneratedRegex(@"<< /Length (\d+) /Filter /FlateDecode >>\nstream\n")]
|
|
private static partial Regex FlateStream();
|
|
}
|