python / python/cpython

[Tool: c-analyzer] Print Helpful Error Messages for Invalid .tsv Files

未关闭
#116,826 0 条评论 1 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

type-feature
主要语言
Python
星标
77.2k
派生
36k
PR 指标待抓取

描述

Feature or enhancement

CI uses the c-analyzer tool (Tools/c-analyzer) to make sure we don't add any new unsupported C global variables to the code base. The tool makes use of .tsv files (tab-separated values) for some customization. The most import of these files are Tools/c-analyzer/cpython/ignored.tsv and .../globals-to-fix.tsv. If someone adds a line that has spaces instead of tabs (or spaces around the tabs), which is easy to do, then the tool unhelpfully fails as though the line weren't there.

(expand for an example)

From the CI check in gh-116413:

python3.12 ./Tools/c-analyzer/check-c-globals.py \
	--format summary \
	--traceback
analyzing files...
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ....
checking analysis results...

[extraneous text snipped]

Non-constant global variables are generally not supported
in the CPython repo.  We use a tool to analyze the C code
and report if any unsupported globals are found.  The tool
may be run manually with:

  ./python Tools/c-analyzer/check-c-globals.py --format summary [FILE]

Occasionally the tool is unable to parse updated code.
If this happens then add the file to the "EXCLUDED" list
in Tools/c-analyzer/cpython/_parser.py and create a new
issue for fixing the tool (and CC ericsnowcurrently
on the issue).

If the tool reports an unsupported global variable and
it is actually const (and thus supported) then first try
fixing the declaration appropriately in the code.  If that
doesn't work then add the variable to the "should be const"
section of Tools/c-analyzer/cpython/ignored.tsv.

If the tool otherwise reports an unsupported global variable
then first try to make it non-global, possibly adding to
PyInterpreterState (for core code) or module state (for
extension modules).  In an emergency, you can add the
variable to Tools/c-analyzer/cpython/globals-to-fix.tsv
to get CI passing, but doing so should be avoided.  If
this course it taken, be sure to create an issue for
eliminating the global (and CC ericsnowcurrently).

[extraneous text snipped]

make: *** [Makefile:3[16](https://github.com/python/cpython/actions/runs/8281492333/job/22660094475#step:16:17)1: check-c-globals] Error 1
Error: Process completed with exit code 2.

There's a similar situation when someone adds a line but forgets one of the row values, likely the last one which is usually a single hyphen. Such a line gets ignored without any indication to the user what went wrong or what to do about it.

(expand for an example)

From the CI check in gh-116413:

python3.12 ./Tools/c-analyzer/check-c-globals.py \
	--format summary \
	--traceback
analyzing files...
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....
..... ..... ..... ..... .....

..... ....
checking analysis results...

[extraneous text snipped]

Non-constant global variables are generally not supported
in the CPython repo.  We use a tool to analyze the C code
and report if any unsupported globals are found.  The tool
may be run manually with:

  ./python Tools/c-analyzer/check-c-globals.py --format summary [FILE]

Occasionally the tool is unable to parse updated code.
If this happens then add the file to the "EXCLUDED" list
in Tools/c-analyzer/cpython/_parser.py and create a new
issue for fixing the tool (and CC ericsnowcurrently
on the issue).

If the tool reports an unsupported global variable and
it is actually const (and thus supported) then first try
fixing the declaration appropriately in the code.  If that
doesn't work then add the variable to the "should be const"
section of Tools/c-analyzer/cpython/ignored.tsv.

If the tool otherwise reports an unsupported global variable
then first try to make it non-global, possibly adding to
PyInterpreterState (for core code) or module state (for
extension modules).  In an emergency, you can add the
variable to Tools/c-analyzer/cpython/globals-to-fix.tsv
to get CI passing, but doing so should be avoided.  If
this course it taken, be sure to create an issue for
eliminating the global (and CC ericsnowcurrently).

make: *** [Makefile:3[16](https://github.com/python/cpython/actions/runs/8280034808/job/22655683079#step:16:17)1: check-c-globals] Error 1
Error: Process completed with exit code 2.

It would be helpful if the tool failed with an appropriate specific message for the following cases:

  • spaces instead of tabs
  • spaces next to tabs
  • incomplete lines

CC @iritkatriel

贡献指南

打开贡献指南

从这里开始

  1. 先读完整个 Issue，再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库，在一个分支上完成修改。
  4. 提交 Pull Request，并在描述里引用这个 Issue 编号。

调研方向

从 Tools/c-analyzer 开始，检查 check-c-globals.py 如何读取 Tools/c-analyzer/cpython/ignored.tsv 和 globals-to-fix.tsv。针对格式错误的 TSV 示例运行 ./python Tools/c-analyzer/check-c-globals.py --format summary。当工具针对用空格代替制表符、制表符旁有空格以及不完整行报告具体错误时，即表示完成。

由索引模型根据 Issue 内容生成。

评估

技术栈
python
领域
tooling
Issue 类型
功能
难度
3/5
预计耗时
1-2 天
活跃度
停滞
描述清晰度
基本清楚
新手友好度
45/100

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。