perf: optimize upload

[j2rong4cn]

添加配置选项 max_buffer_limitMB

• 限制单个buffer的大小
  自动： -1 (默认)：5%的总内存
  关闭：0
  自定义：大于0 单位为MB
• 现阶段主要用于上传，如果单个分片或单文件超过该大小就会先缓存到临时文件再上传

优化以下驱动在上传时的内存和存储占用

115open、123、123open、aliyun_open、google_drive、cloudreve、cloudreve_v4、onedrive、onedrive_app、doubao、189pc

[github-actions]

⚠️ PR 标题需以 feat(): , docs(): , fix(): , style(): , refactor(): 其中之一开头，例如：feat(component): 新增功能。⚠️ The PR title must start with feat(): , docs(): , fix(): , style(): , or refactor(): . For example: feat(component): add new feature.

[Copilot]

Pull Request Overview

This PR adds a configurable buffer limit feature to optimize memory usage during uploads and improves memory/storage efficiency for multiple drivers. The main change introduces a max_buffer_limitMB configuration option that automatically defaults to 5% of total memory, can be disabled with 0, or set to a custom value in MB.

• Added max_buffer_limitMB configuration with automatic memory-based sizing
• Optimized upload operations for 11 drivers to use streaming readers instead of memory buffers
• Replaced string HTTP method literals with constants for consistency

Reviewed Changes

Copilot reviewed 43 out of 43 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
internal/conf/config.go	Added MaxBufferLimit configuration field with default -1
internal/bootstrap/config.go	Implemented automatic buffer limit calculation based on system memory
internal/stream/util.go	Added StreamSectionReader for efficient memory management during uploads
drivers/*/util.go	Replaced memory buffering with streaming section readers for upload operations
pkg/errgroup/errgroup.go	Enhanced error group with ordered execution and lifecycle management
pkg/singleflight/var.go	Changed from ErrorGroup to AnyGroup for better type flexibility

Comments suppressed due to low confidence (1)

internal/stream/util.go:199

• [nitpick] The parameter name 'bufMaxLen' is unclear. Consider renaming to 'maxBufferSize' or 'bufferSizeLimit' to better indicate it represents the maximum buffer size.

func NewStreamSectionReader(file model.FileStreamer, bufMaxLen int) (*StreamSectionReader, error) {

[Copilot]

[nitpick] The error message could be more descriptive about why this condition matters. Consider adding context about the streaming nature of the reader and why sequential access is required.

Suggested change

	return nil, fmt.Errorf("stream not cached: request offset %d != current offset %d", off, ss.off)
	return nil, fmt.Errorf("stream not cached: sequential access required for streaming reader (requested offset %d, current offset %d). Random access is only supported when the stream is fully cached.", off, ss.off)