Receiving warning when I'm not calling auth.getSession anywhere: Using is potentially insecure as it loads data directly from the storage medium (typically cookies) which may not be authentic. Prefer using supabase.auth.getUser() instead. To suppress this warning call supabase.auth.getUser() before you call supabase.auth.getSession() #755

williamlmao · 2024-03-29T03:17:43Z

Bug report

[x ] I confirm this is a bug with Supabase, not with my own application.
I confirm I have searched the Docs, GitHub Discussions, and Discord.

Describe the bug

Despite not calling supabase.auth.getSession() anywhere in my code (as seen in screenshot), I get this warning nonstop in my server side terminal logs Using is potentially insecure as it loads data directly from the storage medium (typically cookies) which may not be authentic. Prefer using supabase.auth.getUser() instead. To suppress this warning call supabase.auth.getUser() before you call supabase.auth.getSession()

To Reproduce

import './globals.css'
import { Metadata } from 'next/types'
import { cookies } from 'next/headers'
import { createServerClient } from '@supabase/ssr'

export default async function RootLayout({
  children,
}: {
  children: React.ReactNode
}) {
  const cookieStore = cookies()

  const supabase = createServerClient(
    process.env.NEXT_PUBLIC_SUPABASE_URL!,
    process.env.NEXT_PUBLIC_SUPABASE_ANON_KEY!,
    {
      cookies: {
        get(name: string) {
          return cookieStore.get(name)?.value
        },
      },
    }
  )

  const user = await supabase.auth.getUser()

  return (
    <html
      lang="en"
    >
      <div className="">text</div>
    </html>
  )
}

Expected behavior

I expect it to not give me the error since I'm using the function anywhere. It would be nice if it told me where in my code it detected the issue.

System information

OS: macOs
"@supabase/ssr": "^0.1.0",
"@supabase/supabase-js": "2.39.1",

The text was updated successfully, but these errors were encountered:

sahmed007 · 2024-03-29T14:26:32Z

Was just about to make an issue about this as well. Getting the same warnings despite not using it.

beelzick · 2024-03-29T20:56:49Z

I have the same issue

przemyslaw-paziewski · 2024-03-29T21:19:13Z

Same here

AlanKeown · 2024-03-30T03:23:50Z

I'm getting the warning but I am using it according to the docs for SSR in Sveltekit.

caanyp24 · 2024-03-31T00:09:12Z

I'm getting the same warning and i am not using supabase.auth.getSession()

Cheveniko · 2024-03-31T09:04:42Z

Same issue here, happened just after I updated the Nextjs and eslint-config-next version from 14.0.4 to 14.1.4.
Before that I wasn't getting the warning.

jdgamble555 · 2024-03-31T21:16:18Z

This is a huge problem and a pain. I don't want to contact the Supabase server to see if a user is logged in. This will slow down my app!

This is a supabase core problem it seems, as I can't find the error message in this package.

J

Pluriscient · 2024-03-31T21:20:47Z

+1 to the problem, only call getSession in the hooks.server.ts with the code from the tutorial (adapted to prevent setting cookies after request has been returned)

  let called = false;
  /**
   * A convenience helper so we can just call await getSession() instead const { data: { session } } = await supabase.auth.getSession()
   */

  event.locals.getSession = async () => {
    called = true;
    const {
      data: { user },
    } = await event.locals.supabase.auth.getUser();
    let {
      data: { session },
    } = await event.locals.supabase.auth.getSession();

    // solving the case if the user was deleted from the database but the browser still has a cookie/loggedin user
    // +layout.server.js will delete the cookie if the session is null
    if (user == null) {
      session = null;
    }

    return session;
  };
  if (!called) {
    await event.locals.getSession();
  }

jdgamble555 · 2024-03-31T21:38:31Z

https://github.com/supabase/auth-js/blob/92fefbd49f25e20793ca74d5b83142a1bb805a18/src/GoTrueClient.ts#L936 - Here is the culprit. I just created a new issue and linked it.

@Pluriscient - I'm not sure why you're doing that, but it doesn't work. You're just calling getSession() an extra time? What is the goal here?

J

Pluriscient · 2024-03-31T21:42:55Z

https://github.com/supabase/auth-js/blob/92fefbd49f25e20793ca74d5b83142a1bb805a18/src/GoTrueClient.ts#L936 - Here is the culprit. I just created a new issue and linked it.

@Pluriscient - I'm not sure why you're doing that, but it doesn't work. You're just calling getSession() an extra time? What is the goal here?

J

The console log actually asked to call getUser before getSession to suppress the warning. The if (!called) escape was included due to this issue

jdgamble555 · 2024-03-31T22:07:58Z

Looks like that fix doesn't work

Pluriscient · 2024-03-31T22:26:30Z

Looks like that fix doesn't work

YMMV, but I haven't seen a crash due to setCookies since. This error popped up the day after when we ran a fresh npm install, though I don't see the ssr package being updated there.

jdgamble555 · 2024-03-31T22:30:51Z

For my purposes this hack solves it. However, I am only using session to detect if there is a session, not get session data.

event.locals.getSession = async () => {

  const { data: { user } } = await event.locals.supabase.auth.getUser();

  if (user === null) {
    return null;
  }

  const session = { user };

  return session as Session;
};

My problem with this approach is that I need to check for a user session on my non logged in page. So I can give different results whether or not a user is logged in (likes, votes, etc on a post). This means even for non-logged-in users, I have to send an extra fetch request to Supabase in order to display my data. Perhaps it should only call getUser() if there is a session, and then verify it on the Supabase server.

This is problematic for sure.

J

cowboycodr · 2024-04-09T20:46:24Z

I have the same issue as well. Makes debugging in the console an absolute nightmare.

hmnd · 2024-04-09T22:34:58Z

Just chiming in with my vote in support of fixing this issue... here's what I said in another thread:

I understand the intention with this warning, but it's impossible to silence when I know I'm properly verifying with getUser. In my case, I need to use it with createSupabaseLoadClient from auth-helpers, and because it serializes the session as JSON, I get a warning every time I call that function or eg. spread the session into a new object, etc.

At the very least, could the warning only be displayed once per session or could we please have an option to disable it? I see that now it's gated behind a non-existent __suppressUserWarning key on session.

For reference the warning I'm getting is below. It's a different one from the one this discussion originally addresses:

Using the user object as returned from supabase.auth.getSession() or from some supabase.auth.onAuthStateChange() events could be insecure! This value comes directly from the storage medium (usually cookies on the server) and many not be authentic. Use supabase.auth.getUser() instead which authenticates the data by contacting the Supabase Auth server.

ixxie · 2024-05-05T08:57:15Z

As somebody just setting up a SvelteKit project with Supabase, this is quite confusing: copy-pasting code from your tutorial results in a warning about best practices! One more vote for a fix!

j4w8n · 2024-05-09T09:04:04Z

Oops! I created an issue for this a couple of weeks after yours was created: supabase/auth-js#888

sgriff96 · 2024-05-26T05:10:31Z

Should we open a separate ticket about addressing the incorrect code in the documentation for sveltekit? As @ixxie said it's very confusing that the documentation is incorrect and has us do something wrong/unsafe.

j4w8n · 2024-05-26T09:47:46Z

Should we open a separate ticket about addressing the incorrect code in the documentation for sveltekit? As @ixxie said it's very confusing that the documentation is incorrect and has us do something wrong/unsafe.

You can, but there is currently no Supabase-recommended way to set up Sveltekit while avoiding these warnings. It's unfortunate, but that's where we are right now.

And you aren't doing anything wrong or unsafe, but rather, there are some unintended side-effects of their warning implementation.

sgriff96 · 2024-05-26T16:45:48Z

You can, but there is currently no Supabase-recommended way to set up Sveltekit while avoiding these warnings. It's unfortunate, but that's where we are right now.

And you aren't doing anything wrong or unsafe, but rather, there are some unintended side-effects of their warning implementation.

Oh okay. So it is safe to access the session from locals whenever I want now because we used safeGetSession in the hooks.server.ts?

j4w8n · 2024-05-26T19:01:28Z

You can, but there is currently no Supabase-recommended way to set up Sveltekit while avoiding these warnings. It's unfortunate, but that's where we are right now.
And you aren't doing anything wrong or unsafe, but rather, there are some unintended side-effects of their warning implementation.

Oh okay. So it is safe to access the session from locals whenever I want now because we used safeGetSession in the hooks.server.ts?

It's safe as long as you heed the warning (that some of us shouldn't be seeing): don't trust info from session.user on the server side - even after calling safeGetSession (SvelteKit). What they're implying with that function is you should use the returned user data instead of session.user.

I don't want the scope of this issue to grow beyond it's original purpose, but if you're wanting to understand why all of this has come up, read this. If you have further questions or comments about it, I'd recommend asking them on the discussion rather than this issue.

sgriff96 · 2024-05-29T16:26:16Z

It's safe as long as you heed the warning (that some of us shouldn't be seeing): don't trust info from session.user on the server side - even after calling safeGetSession (SvelteKit). What they're implying with that function is you should use the returned user data instead of session.user.

I don't want the scope of this issue to grow beyond it's original purpose, but if you're wanting to understand why all of this has come up, read this. If you have further questions or comments about it, I'd recommend asking them on the discussion rather than this issue.

Ohh understood then, thank you for the clarification!

thedelanyo · 2024-06-22T21:34:36Z

It's safe as long as you heed the warning (that some of us shouldn't be seeing): don't trust info from session.user on the server side - even after calling safeGetSession (SvelteKit). What they're implying with that function is you should use the returned user data instead of session.user.
I don't want the scope of this issue to grow beyond it's original purpose, but if you're wanting to understand why all of this has come up, read this. If you have further questions or comments about it, I'd recommend asking them on the discussion rather than this issue.

Ohh understood then, thank you for the clarification!

But how do we to least make the warning disappears? It makes reading (debugging) using server logs, a nightmare.

j4w8n · 2024-06-22T23:18:00Z

But how do we to least make the warning disappears? It makes reading (debugging) using server logs, a nightmare.

@delanyoyoko, there’s no native Supabase way to do this if you're using SvelteKit. Some people have created ways to silence the logs with custom console.log code. Another option is to verify and decode the access token, then return your own validated session - which I have an example of at https://github.com/j4w8n/sveltekit-supabase-ssr

jdgamble555 · 2024-06-23T00:50:49Z

@j4w8n - Then your code is flawed, not SvelteKit. Make an option to turn it off. There is a reason there are probably over 100 complaints to this over three different Supabase packages. If you can't make a warning that properly works due to the limitations of SvelteKit, then don't make the warning, or make an option to silence it. At the VERY LEAST, it should not be able to display the warning more than once.

It was a HUGE oversight for the Supabase team to build this without proper testing.

J

j4w8n · 2024-06-23T01:28:08Z

@jdgamble555 I'm with you, and I understand the frustration. There should've been more testing, and, in my opinion, this warning should've been rolled back by now.

Having said that, I was not involved in this code decision. I've contributed to the auth codebase, but I'm not part of the Supabase team; nor am I a collaborator. I'm sorry if I've implied or lead anyone to believe that I'm more involved than I actually am. Seeing my contributor tag, and possibly heavy involvement on Discord, can understandbly cause people to conclude what you have; because it's the same tag hf has, yet they are heavily involved with code decisions and are a part of the auth team.

I hope this gets resolved soon because it's clearly harming dx.

jdgamble555 · 2024-06-23T02:39:28Z

@j4w8n - Fair enough. I wasn't necessarily venting towards you, just at the Supabase team in general for letting this issue get out of hand. I would imagine there is a way to solve this issue without just patching the problem with hacks. I think most people use the code from the examples, not insecurely.

J

thedelanyo · 2024-06-23T16:22:22Z

But how do we to least make the warning disappears? It makes reading (debugging) using server logs, a nightmare.

@delanyoyoko, there’s no native Supabase way to do this if you're using SvelteKit. Some people have created ways to silence the logs with custom console.log code. Another option is to verify and decode the access token, then return your own validated session - which I have an example of at https://github.com/j4w8n/sveltekit-supabase-ssr

This works fine. The logs are gone, session is validated and my dev server seem pretty fast. Thanks.

mstibbard · 2024-06-25T05:58:13Z

But how do we to least make the warning disappears? It makes reading (debugging) using server logs, a nightmare.

@delanyoyoko, there’s no native Supabase way to do this if you're using SvelteKit. Some people have created ways to silence the logs with custom console.log code. Another option is to verify and decode the access token, then return your own validated session - which I have an example of at https://github.com/j4w8n/sveltekit-supabase-ssr

I understand the warning is showing unnecessarily, but I'm confused how & why @j4w8n's implementation manages to not trigger the warnings when the SSR SvelteKit guide code does? It appears both implementations call getSession in a similar manner and neither implementation uses the unvalidated output.

thedelanyo · 2024-06-25T08:03:07Z

But how do we to least make the warning disappears? It makes reading (debugging) using server logs, a nightmare.

@delanyoyoko, there’s no native Supabase way to do this if you're using SvelteKit. Some people have created ways to silence the logs with custom console.log code. Another option is to verify and decode the access token, then return your own validated session - which I have an example of at https://github.com/j4w8n/sveltekit-supabase-ssr

I understand the warning is showing unnecessarily, but I'm confused how & why @j4w8n's implementation manages to not trigger the warnings when the SSR SvelteKit guide code does? It appears both implementations call getSession in a similar manner and neither implementation uses the unvalidated output.

Well I think it's not about calling the "getSession" method. Whether getSession is called or not, the message will be logged.

Even when you solely rely on user, thereby just calling getUser, your console will be flooded.

I guess the log is a bug sort, probably supposed to logged once, if the user is triggered on getSession.

j4w8n · 2024-06-25T11:36:53Z

I understand the warning is showing unnecessarily, but I'm confused how & why @j4w8n's implementation manages to not trigger the warnings when the SSR SvelteKit guide code does? It appears both implementations call getSession in a similar manner and neither implementation uses the unvalidated output.

The short answer is: Because both SSR and SvelteKit code are accessing session.user at some point. The Supabase guide is preserving the Proxied version of session, allowing the logs to trigger; while I'm creating my own version of it.

SSR is doing it by calling JSON.stringify(session) in +layout.ts
SvelteKit is doing it when validating that we're returning POJOs from certain files

REF: supabase/auth-js#873 (comment)

jdgamble555 · 2024-06-25T12:51:01Z

I don't understand why this can't be fixed under the hood.

EmilienKopp · 2024-07-01T11:45:43Z

Leaving this here for whoever does not want to do a custom token validation workaround or still hasn't reworked their code.
This will at least silence the warnings. Importing that in hooks.server.ts did the trick.

const originalConsoleWarn = console.warn;

console.warn = function (...args) {
  const shouldLog = args.every((arg) => {
    if (typeof arg === 'string') {
      return !arg.includes('Using the user object');
    }
    return true;
  });
  if(shouldLog) {
    originalConsoleWarn.apply(console, args);
  }
};

williamlmao added the bug Something isn't working label Mar 29, 2024

jdgamble555 mentioned this issue Mar 31, 2024

Can't get rid of getUser() warning supabase/auth-js#873

Open

2 tasks

j4w8n mentioned this issue Apr 10, 2024

Code in supabase-js calls getSession(), which now throws a warning when using in combo with ssr supabase/supabase-js#1010

Closed

2 tasks

cowboycodr mentioned this issue Apr 12, 2024

Call getUser instead of getSession in vulnerable cases cowboycodr/BlueQ#54

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

williamlmao commented Mar 29, 2024

sahmed007 commented Mar 29, 2024

beelzick commented Mar 29, 2024

przemyslaw-paziewski commented Mar 29, 2024

AlanKeown commented Mar 30, 2024

caanyp24 commented Mar 31, 2024

Cheveniko commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024 •

edited

Loading

jdgamble555 commented Mar 31, 2024

cowboycodr commented Apr 9, 2024

hmnd commented Apr 9, 2024

ixxie commented May 5, 2024

j4w8n commented May 9, 2024

sgriff96 commented May 26, 2024

j4w8n commented May 26, 2024

sgriff96 commented May 26, 2024

j4w8n commented May 26, 2024

sgriff96 commented May 29, 2024 •

edited

Loading

thedelanyo commented Jun 22, 2024

j4w8n commented Jun 22, 2024

jdgamble555 commented Jun 23, 2024

j4w8n commented Jun 23, 2024

jdgamble555 commented Jun 23, 2024

thedelanyo commented Jun 23, 2024

mstibbard commented Jun 25, 2024

thedelanyo commented Jun 25, 2024

j4w8n commented Jun 25, 2024

jdgamble555 commented Jun 25, 2024

EmilienKopp commented Jul 1, 2024 •

edited

Loading

Comments

williamlmao commented Mar 29, 2024

Bug report

Describe the bug

To Reproduce

Expected behavior

System information

sahmed007 commented Mar 29, 2024

beelzick commented Mar 29, 2024

przemyslaw-paziewski commented Mar 29, 2024

AlanKeown commented Mar 30, 2024

caanyp24 commented Mar 31, 2024

Cheveniko commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024

jdgamble555 commented Mar 31, 2024

Pluriscient commented Mar 31, 2024 • edited Loading

jdgamble555 commented Mar 31, 2024

cowboycodr commented Apr 9, 2024

hmnd commented Apr 9, 2024

ixxie commented May 5, 2024

j4w8n commented May 9, 2024

sgriff96 commented May 26, 2024

j4w8n commented May 26, 2024

sgriff96 commented May 26, 2024

j4w8n commented May 26, 2024

sgriff96 commented May 29, 2024 • edited Loading

thedelanyo commented Jun 22, 2024

j4w8n commented Jun 22, 2024

jdgamble555 commented Jun 23, 2024

j4w8n commented Jun 23, 2024

jdgamble555 commented Jun 23, 2024

thedelanyo commented Jun 23, 2024

mstibbard commented Jun 25, 2024

thedelanyo commented Jun 25, 2024

j4w8n commented Jun 25, 2024

jdgamble555 commented Jun 25, 2024

EmilienKopp commented Jul 1, 2024 • edited Loading

Pluriscient commented Mar 31, 2024 •

edited

Loading

sgriff96 commented May 29, 2024 •

edited

Loading

EmilienKopp commented Jul 1, 2024 •

edited

Loading