doc: note misleading error code in uv_os_get_passwd
#3926
Conversation
| @@ -541,6 +541,10 @@ API | |||
| memory allocated to `pwd` needs to be freed with | |||
| :c:func:`uv_os_free_passwd`. | |||
|
|
|||
| .. note:: | |||
| If the effective uid does not have an entry in the password file this | |||
| function will return ENOENT even if the file exists. | |||
There was a problem hiding this comment.
| function will return ENOENT even if the file exists. | |
| function will return `UV_ENOENT` even if the file exists. |
I kind of feel we shouldn't be using the phrase "password file" since there doesn't have to be one (e.g., on systems that use LDAP-based authentication) but I'm not sure what to replace "password file entry" on line 536 with.
There was a problem hiding this comment.
I think "user source" or "user database" would fit well
There was a problem hiding this comment.
Yes, that's a good suggestion.
There was a problem hiding this comment.
I took a look at the windows source, and it looks like the specific ENOENT issue is only relevant for unix. Also, the term password file is used all over the documentation of this function.
Maybe prefixing this block with On Unix systems, if the effective uid... would be the way to go
There was a problem hiding this comment.
I'd consider UV_ENOENT being Unix-only an implementation detail. It's something that can change in the future. I wouldn't mention that.
There was a problem hiding this comment.
What can be particularly confusing here is that the authoritative information for linux systems now comes not directly from /etc/password but from /etc/nsswitch.conf which specifies a list of dynamic library (such as libnss_files.so) to open to search for pwnam info corresponding to the given uid. The man page has this to say about the error:
0 or ENOENT or ESRCH or EBADF or EPERM or ...
The given name or uid was not found.
If an effective uid does not have an entry in the passwd file, the
uv_os_get_passwdfunction will returnENOENTwhich has misled our debugging efforts to look for missing files even if they were present. I've added a short note to the documentation to help others encountering similar issues in the future.