LlosLab Software Development BlogEngineering notes for PDF, Delphi, and document software developers.

Техническая статья

HotPDF: Direct File API processing for large PDFs в Delphi

HotPDF Component

HotPDF — нативная VCL PDF-библиотека для приложений Delphi и C++Builder, которым нужны прямое создание и редактирование PDF, формы, аннотации, шифрование, цифровые подписи, Unicode-шрифты, вывод с учетом стандартов и preflight-отчеты без внешнего PDF-runtime.

Эта статья предназначена для developers processing large statements, archives, drawings, or customer bundles в Delphi. Она рассматривает Direct File API processing for large PDFs как промышленную инженерию документов, а не как одиночный вызов компонента.

Практический риск состоит в том, что a workflow that is acceptable for a small PDF can exhaust memory, leave partial files, or become impossible to support when documents reach hundreds of megabytes. Поэтому процессу нужны письменный контракт, наблюдаемая диагностика и реалистичные регрессионные файлы.

Архитектурные решения

Treat storage as part of the PDF pipeline. maximum input size, page count, and temporary storage budget / page-range validation rules and whether ranges are user-supplied or policy-derived

  • maximum input size, page count, and temporary storage budget
  • page-range validation rules and whether ranges are user-supplied or policy-derived
  • output naming, atomic replacement, rollback, and partial-result retention
  • progress reporting, cancellation behavior, and support bundle contents

Порядок реализации

Plan page ranges and output targets before opening the file. The order below keeps the workflow reviewable for Delphi and C++Builder teams.

  1. validate the file path, size, page count, and page-range request before processing
  2. choose a direct-read strategy and allocate temporary files in a controlled location
  3. stream output to a new file and avoid replacing the source until validation passes
  4. record page mappings, skipped ranges, warnings, and elapsed time per stage
  5. delete or retain temporary artifacts according to the support policy

Доказательства проверки

Operational evidence for large-file jobs. Keep these fields with the output or support record.

  • input size, page count, selected ranges, output size, and peak memory estimate
  • temporary file paths, cleanup status, cancellation point, and final disposition
  • warnings for damaged objects, unsupported compression, or repaired cross references
  • hashes for input and output files when customer support needs reproducibility

Memory pressure is usually a design issue

Direct file access is most useful when the workflow knows which pages, objects, and metadata need to move. The application should avoid loading the whole document as a convenience layer when the business operation only needs a bounded subset.

Customer-visible behavior

Users do not see internal call order. They see whether the file opens, validates, prints, edits, imports, or gets rejected. The workflow should translate Direct File API processing for large PDFs results into states users can act on.

  • validate the file path, size, page count, and page-range request before processing
  • choose a direct-read strategy and allocate temporary files in a controlled location
  • stream output to a new file and avoid replacing the source until validation passes
  • network paths and antivirus filters can change latency more than PDF parsing does
  • page ranges should be checked before output begins to avoid empty deliverables

Замечания для инженерного ревью по Direct File API processing for large PDFs

Используйте эти замечания, чтобы убедиться, что функция вышла за рамки демо и может быть обоснована на релизе, в поддержке и при эскалации клиента

  • Решение: maximum input size, page count, and temporary storage budget. Точка приложения при реализации: choose a direct-read strategy and allocate temporary files in a controlled location. Доказательство приемки: warnings for damaged objects, unsupported compression, or repaired cross references. Триггер регрессии: linearized or incrementally saved files may contain revisions the user did not expect
  • Решение: page-range validation rules and whether ranges are user-supplied or policy-derived. Точка приложения при реализации: stream output to a new file and avoid replacing the source until validation passes. Доказательство приемки: hashes for input and output files when customer support needs reproducibility. Триггер регрессии: network paths and antivirus filters can change latency more than PDF parsing does
  • Решение: output naming, atomic replacement, rollback, and partial-result retention. Точка приложения при реализации: record page mappings, skipped ranges, warnings, and elapsed time per stage. Доказательство приемки: input size, page count, selected ranges, output size, and peak memory estimate. Триггер регрессии: page ranges should be checked before output begins to avoid empty deliverables
  • Решение: progress reporting, cancellation behavior, and support bundle contents. Точка приложения при реализации: delete or retain temporary artifacts according to the support policy. Доказательство приемки: temporary file paths, cleanup status, cancellation point, and final disposition. Триггер регрессии: partial output should never overwrite a known-good source file
  • Решение: maximum input size, page count, and temporary storage budget. Точка приложения при реализации: validate the file path, size, page count, and page-range request before processing. Доказательство приемки: warnings for damaged objects, unsupported compression, or repaired cross references. Триггер регрессии: linearized or incrementally saved files may contain revisions the user did not expect
  • Решение: page-range validation rules and whether ranges are user-supplied or policy-derived. Точка приложения при реализации: choose a direct-read strategy and allocate temporary files in a controlled location. Доказательство приемки: hashes for input and output files when customer support needs reproducibility. Триггер регрессии: network paths and antivirus filters can change latency more than PDF parsing does

Пограничные случаи

  • network paths and antivirus filters can change latency more than PDF parsing does
  • page ranges should be checked before output begins to avoid empty deliverables
  • partial output should never overwrite a known-good source file
  • linearized or incrementally saved files may contain revisions the user did not expect

Примечания по Delphi / C++Builder

HotPDF Component should sit behind a small service boundary that receives files, streams, profiles, and credentials, then returns output paths, warnings, metrics, and validation status. Важные термины включают Direct File API, large PDF, page range, streaming, temporary file, rollback.

Пример кода Delphi

Следующий эскиз Delphi показывает практическую границу сервиса для этой темы. Оставляйте проверки политики, журналирование и валидацию вне узкого блока вызова продукта, чтобы сценарий было проще тестировать.

procedure CopyLargePdfForIntake(const SourceFile, OutputFile: string);
var
  Pdf: THotPDF;
  PageCount: Integer;
begin
  Pdf := THotPDF.Create(nil);
  try
    if Pdf.DACopyFile(SourceFile, OutputFile, PageCount) <> 1 then
      raise EInvalidOperation.Create('Direct copy failed');
    LogDirectAccessCopy(SourceFile, OutputFile, PageCount);
    VerifyCopiedBytes(SourceFile, OutputFile);
  finally
    Pdf.Free;
  end;
end;

Производственный чек-лист

  • Запускайте сценарий на пустом файле, обычном клиентском файле и файле худшего случая
  • Открывайте сгенерированный PDF в целевом просмотрщике, валидаторе, принтере или downstream-приложении
  • Записывайте версию продукта, версию профиля, хэш входа, путь вывода, затраченное время и число предупреждений
  • Храните пароли, сертификаты, временные файлы и данные клиентов по явным правилам хранения
  • Добавляйте регрессионные документы, когда клиентский файл выявляет новый граничный случай

Документация по продукту

HotPDF Component

Дополнительные примеры кода

// Structural copy: validate-and-move without parsing the object tree
Status := Pdf.DACopyFile('incoming\statement.pdf', 'verified\statement.pdf');
LogDirectFileStatus('copy', Status);

// Decrypt while copying: the Direct File route into protected inputs
Status := Pdf.DecryptFile('incoming\protected.pdf',
  'verified\plain.pdf', 'batch-password');
LogDirectFileStatus('decrypt-copy', Status);

// Encrypt while copying: protect an output without a full load
Status := Pdf.EncryptFile('verified\statement.pdf',
  'outbound\statement.pdf', 'owner-secret', '', aes256, [prPrint]);
LogDirectFileStatus('encrypt-copy', Status);
// Append an audit page to a large archive without rewriting it
Pdf.BeginIncrementalUpdate('archive-2026-06.pdf');
Pdf.AddPage;
Pdf.CurrentPage.SetFont('Arial', [], 10);
Pdf.CurrentPage.TextOut(50, 760, 0, 'Processed by intake service 2026-06-11');
Pdf.SaveIncrementalUpdate('archive-2026-06-stamped.pdf');  // original bytes + delta