sphinx-doc/sphinx

Cannot recover from MemoryErrors

已关闭

#7,635 创建于 2020年5月8日

在 GitHub 查看贡献者指南
 (2 条评论) (0 个反应) (0 位负责人)Python (1,985 个派生)batch import
help wantedtype:bug

仓库指标

星标
 (5,625 个星标)
PR 合并指标
 (平均合并 10天 11小时) (30 天内合并 11 个 PR)

描述

We are using recommonmark with sphinx to parse 2.3k+ md files and experiencing excesive memory usage. Running 64 bit we see a peak of about 19gb of memory usage and an execution time around 30 minutes. On many machines (or running 32 bit) this is not doable. The majority of the markdown files have an rst label, and some of them have references to these labels. Are there configuration options that may be affecting this? Is there a way to review what's actually taking up this memory?

贡献者指南

研究方向
Investigate memory usage during parsing of many markdown files with recommonmark. Use memory profiler or tracemalloc to identify peak memory allocations. Compare with direct Python memory usage and explore Sphinx's handling of large document sets. Look at recommonmark's parser and Sphinx's build process in sphinx/builders/ for potential optimizations or configuration options to reduce memory footprint.
技术栈
python
领域
documentationperformance
议题类型
缺陷
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
4
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
超过 1 周
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
较久未更新
清晰度议题是否清楚说明期望改动、验收标准和下一步。
需要先调研
前置要求
PythonMemory profilingSphinx internalsCPython debugging
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
10