New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Blank lines in ._pth file are not ignored #73512
Comments
If a python._pth file includes a blank line, it gets treated as '\n' which is then appended to the directory and used as an entry in sys.path. Empty lines should be ignored completely. |
I've attached a patch that makes the site package ignore empty lines in .pth files. Along with a test for this behavior. I'm not really familiar with the site machinery so I hope the test and code are fine. |
The fix for this issue is needed in PC/getpathp.c as this is a special feature in path generation on Windows. I'm not aware of any such issue with .pth files - the underscore in ._pth is deliberate. |
I've encountered this issue too. (FYI, the official 3.6.0 embeddable zip from https://www.python.org/downloads/windows/ does contain a blank line in its ._pth, so all its users get an invalid entry in sys.path). The patch is attached. I couldn't fix tests in more straightforward way because both of them were broken:
I've strengthened the check in 'nosite' test because it is trivial to fully emulate sys.path calculation logic in this case. I've preserved weaker checks in the other test because I didn't want to emulate site.py logic. |
I haven't had a chance to try the patch (underpth_blank_lines.diff) or apply it, but it looks right. If one of the other core devs gets to apply this before I do, feel free. Alexey, you don't need to sign a CLA for this fix, but if you're going to contribute more then you may want to: https://docs.python.org/devguide/committing.html#contributor-licensing-agreements It's all online, and it'll help save the core developers from having to evaluate whether you need it or not (which I can do, because it comes up often for me in my day job, but not everyone has that experience). And thanks for this patch! |
New changeset 54fea351e3f9 by Steve Dower in branch '3.6': New changeset a0ff777ab153 by Steve Dower in branch 'default': |
New changeset 6d90b135e36c5b2794a29f650ebebf8bbd96f3fb by Steve Dower in branch 'master': New changeset 3ba099c135cdbced6c2ff168a90a689a83b376db by Steve Dower in branch 'master': |
An identical issue *does* exist for .pth files. As you can see from examining site.addpackage, it does not ignore blank lines as the documentation says. Ammar's patch should also be applied. |
Please create a new issue. This one is closed. |
Hi. I guess this issue has to be reopen because Alexey erroneously broke adding empty paths to Official embeddable Pythons do contain a blank line in its ._pth, and this allowed users to import modules from script directory. Kind regards, |
I shall attach pull request soon. |
The python._pth file generated for the embeddable distribution includes ".", which allows you to import from the dist directory. The blank line should be ignored. If you want to misuse the embeddable distribution by not restricting its search paths, delete the ._pth file. |
Hi Steve. I'll try to explain what is my motivation. I need a reliable way to run Python (not matter embedded or not) in isolated mode, but still having current directory in Ironically I could misuse normal mode to simulate isolated mode with specified %PYTHONPATH% (which is not much handy, by the way), but hiding Windows Registry keys would look pretty hard (it is impossible to make a bat/cmd wrapper). Possibility to search %PYTHONPATH% in HKLM entries makes it even worse. Empty entry in # How would I fix it. Possibly we should allow another keyword for adding an empty entry (like we have an exclusion for Ideas are appreciated. |
FYI: I have created bpo-33689 for the non-Windows-specific issue. |
Vladimir, please open a new issue for this feature request. This issue is closed. |
pth
[WIP]. #7243Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
bugs.python.org fields:
The text was updated successfully, but these errors were encountered: