Deprecation of ldapjs

[wesley3295]

Any ideas on what we can do in replacement of ldapjs? This package is something my company relies on heavily and with it not being maintained anymore throws a huge plot twist into our current work with our client. Is there another open source ldap package that can replace ldapjs or something I/we can contribute to?

[shaozi]

It is a very sad that some stupid coward threatened the maintainer and now the project is archived.

So far I have not found anything yet. We should find some people to take the project over since it is open source.

[mattkrins]

I don't know of any alternatives sadly... It was one of, if not the only decent ldap packages for node.
I've never seen such an unhinged deprecation readme before.
It was clearly a troll/spam email but I guess the author wanted an excuse to stop maintaining it.

[stalkerg]

I think we can easily port it to https://github.com/jjhoughton/napi-ldap or make it abstract.

[stalkerg]

or maybe it will be better just to reimplement such features by Rust? https://github.com/inejge/ldap3

[shaozi]

or maybe it will be better just to reimplement such features by Rust? https://github.com/inejge/ldap3

I like this idea. Any pointers on how?

[stalkerg]

I like this idea. Any pointers on how?

I think the easier way to make binding by this
https://github.com/neon-bindings/neon

It should be relatively easy because we do not need many features.
If you want, I can try/help next week.

PS Of course, true JS implementation has some benefits, but we do not have easy solutions in the current situation.

[stalkerg]

It seems like we are using one filter: https://github.com/ldapjs/filter. Do we need it? As I understand, it's IR, and in our case, (${userAttr}=${userName}) it's overkill.

[stalkerg]

For the rust idea, I think it will be easier to implement high-level functions - _searchUser and _searchUserGroups.

[shaozi]

It seems like we are using one filter: https://github.com/ldapjs/filter. Do we need it? As I understand, it's IR, and in our case, (${userAttr}=${userName}) it's overkill.

What is IR? Now I look into the code, it does seem over complicated in the ldapjs/filter implementation. But I wonder if we simply use string substitute, do we need to do any pre-processing of the user input to ease any security injection concern? Of course, I am not sure the filter lib does that, but I just assume it does something.

[shaozi]

The neon-bindings is a really nice lib. I searched and there is another one called napi-rs. Have you compared these two?

[stalkerg]

What is IR?

https://en.wikipedia.org/wiki/Intermediate_representation

do we need to do any pre-processing of the user input to ease any security injection concern?

Yes, sure, but it is trivial; all literals should be escaped. Rust lib already supports it: https://docs.rs/ldap3/0.11.3/ldap3/fn.ldap_escape.html

The neon-bindings is a really nice lib. I searched and there is another one called napi-rs. Have you compared these two?

Yeah, it seems like napi-rs even better!
https://napi.rs/docs/more/neon

[shaozi]

Cool let me try it out next week and see how is it going.

[stalkerg]

do we have any updates on a such hot topic?

[kedom1337]

Has anybody already evaluated ldapts as an alternative?

[shaozi]

do we have any updates on a such hot topic?

I have done some testing with rust ldap3 with napi-rs and the result is great!

Here is the plan forward:

I will maintain two branches:

• v3 which is the current one based on ldapjs
• v4 which will based on napi-rs and rust ldap3. v4 will have a slightly incompatible result structure than v3, mainly the binary part, will be all base64 encoded string. otherwise I will make it as close to v3 as possible.

[shaozi]

Has anybody already evaluated ldapts as an alternative?

This one looks nice. As I said, I will keep v3 as pure js/ts, so this one may fit there.

[stalkerg]

Yeah, ldapts also looks good, maybe even easier for a short term.