source does not support the use of html as the source entry

[StringKe]

• I have searched the issues of this repo and believe that this is not a duplicate.
• I have searched the FAQ and general documentation and believe that my question is not already covered.

Feature Request

source does not support the use of html as the source entry, the spliced file address is wrong!

The correct URL should be https://modelscope.oss-cn-beijing.aliyuncs.com/releases/v1.3/modelscope-1.3.1-py3-none-any.whl

➜ poetry lock
Updating dependencies
Resolving dependencies... (2.6s)

404 Client Error: Not Found for url: https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html/v1.3/modelscope-1.3.1-py3-none-any.whl

[tool.poetry.dependencies]
python = "^3.11"
fastapi = "^0.93.0"
uvicorn = { version = "^0.20.0", extras = ["standard"] }
modelscope = { version = "^1", extras = ["cv"], source = "modelscope" }

[[tool.poetry.source]]
name = "modelscope"
url = "https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html"
secondary = true

[dimbleby]

because

poetry/src/poetry/repositories/link_sources/html.py

Lines 60 to 61 in 40061f9

	if not url.endswith("/"):
	url += "/"

if you want to submit an MR then that's the bit that needs to change

[PabloEmidio]

The solution would be something like:

if not (url.endswith("/") or url.endswith('html')):
    url += "/"

---

If anyone already have submitted anything untill tonigh, I'll purpose a PR with the changes and understand how could the tests be modify to attest this behavior.

[neersighted]

Duplicate #6687.

[dimbleby]

I think not a duplicate of #6687

this issue is about download, not upload. The proposed solution in #6687 is to normalize the URL by adding a trailing slash: but poetry attempting that normalisation is exactly the cause of the bug here.

[neersighted]

My bad, I had the wrong issue on the clipboard when closing. #6885 is what I meant to link/I thought I pasted; unless I'm mistaken, this is a clear duplicate of that issue.

[dimbleby]

ah yes, that does look duplicate.

gosh that's a lot of talk for an issue which I expect should be solved by removing two lines of code. Perhaps I should read through it and understand why I'm wrong...

[neersighted]

The discussion mostly centers around how we should detect a file terminating a URL, as simply removing the appending of a trailing slash is likely to break existing users. I guess the other way to look at it is that they likely should not have depended on such a feature, and a one-time break is an improvement. In any case, feel free to weigh in there.

[github-actions]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.