unescaped slashes for entry names/paths

[doganoo]

Hey there,

thank you for creating this awesome library!

I have noticed that entry names are not escaped. This leads to wrong path names, as you can see here:

Test/Entry is a password in Test/datenbank within the path ../Datenbanken/

Is it possible to escape those names? Or any known workarounds for this?

Thanks in advance!

EDIT

database = PyKeePass("", "") 

for group in database.groups:
    print(group.name)
    print("----------------")
    for entry in group.entries:
        print(entry.path)

[pschmitt]

How do you create these entries?

[doganoo]

How do you create these entries?

using KeepassXC "new group"

[KeyWeeUsr]

I'm thinking it's not an issue of PyKeePass but rather of KeepassXC and it's way of inserting entries.

>>> from pykeepass import PyKeePass as P
>>> kp = P("<somewhere>", password="pass")
>>> kp.entries[17].group.group.group.group.name
'Root'
>>> kp.entries[17].group.group.group.name
'foobar_group'
>>> kp.entries[17].group.group.name
'subgroup/name'
>>> kp.entries[17].group.name
'hello/hi'
>>> kp.entries[17].title
'My/super/entry'
>>> kp.entries[17].group.path
'foobar_group/subgroup/name/hello/hi/'
>>> kp.entries[17].group.group.path
'foobar_group/subgroup/name/'
>>> kp.entries[17].group.group.group.path
'foobar_group/'
>>> kp.entries[17].group.group.group.group.path
'/'
>>> kp.add_entry(kp.entries[17].group, title="path/test", username="user/name", password="pass/word")
Entry: "foobar_group/subgroup/name/hello/hi/path/test (user/name)"
>>> kp.entries[18]
Entry: "foobar_group/subgroup/name/hello/hi/path/test (user/name)"
>>> kp.entries[18].username
'user/name'
>>> kp.entries[18].password
'pass/word'
>>> kp.entries[18].title
'path/test'
>>> kp.entries[18].path
'foobar_group/subgroup/name/hello/hi/path/test'
>>> kp.entries[18].group.path
'foobar_group/subgroup/name/hello/hi/'
>>> kp.find_groups(name="hello/hi", first=True)
Group: "foobar_group/subgroup/name/hello/hi/"

[Evidlo]

The KeePass spec allows / inside of entry titles and group names, and since we use it as a delimiter, things can be a bit confusing.

[KeyWeeUsr]

But then it's as good delimiter as any other due to keepass allowing almost everything in there. We can do spaces between the paths or choose a new delimiter (e.g. > )and the same issue will be opened later.

We might also do some bracketing for visibility e.g.:

[Root]/[Group/with/slash]/[Title]

[Evidlo]

The safest thing would be to return a list instead of a string, but this is a breaking change. What are you using paths for?

…

On Oct 10, 2020 2:29 AM, "Peter Badida" ***@***.***> wrote: But then it's as good delimiter as any other due to keepass allowing almost everything in there. We can do spaces between the paths or choose a new delimiter (e.g. > )and the same issue will be opened later. We might also do some bracketing for visibility e.g.: [Root]/[Group/with/slash]/[Title] — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#205 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABJT7YP46EFRGHP2YXOWVKDSKAEOFANCNFSM4SKA5A5A> .

[KeyWeeUsr]

@Evidlo Not necessarily, we can just add a new attribute and keep the original path in place. This way it won't be a breaking change and people who'd want to use the list would be able to.

[doganoo]

@Evidlo whats about represeting as a graph?

[KeyWeeUsr]

@doganoo Do you mean as a CLI tool like in #207 ?

[doganoo]

@doganoo Do you mean as a CLI tool like in #207 ?

@KeyWeeUsr Yes exactly :-)

When do you plan to merge this?

[Evidlo]

I think I'll make this a breaking change and release it with the next major version.

[doganoo]

cool, do you have a timeline for that?

[Evidlo]

Check out https://github.com/libkeepass/pykeepass/tree/pathlist and let me know what you think.

I may do a major release in the next couple of weeks.

[doganoo]

@Evidlo any updates here? :)

[Evidlo]

pathlist is now merged into master. Just waiting on #204 before making a release.

[spaetz]

As a downstream user I have to say that I would have preferred a fixed up path2 attribute instead of a breaking change in order to avoid version dependent-behavior changes. Especially, given that the version variable itself is not guaranteed to be there (as it was just introduced a version earlier).
That having said, we can certainly cope with the change...