Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=209.85.210.173; helo=mail-pf1-f173.google.com; envelope-from=peter@netlandish.com; receiver=<UNKNOWN> 
Date: Mon, 21 Dec 2020 17:16:18 -0800
From: Peter Sanchez <peter@netlandish.com>
To: =?utf-8?B?0JjQu9GM0Y8=?= <mortas.11@gmail.com>
Cc: ~petersanchez/public-inbox@lists.code.netlandish.com
Subject: Re: [PATCH django-impersonate] optimization | preserve request.user
 laziness
Message-ID: <20201222011618.r4g5lhd7y5rhffsp@thinkpad>
References: <CAJF=k3=D3vguoC6ZhGJDQaRinJTU0H60zRyCQoT+fFVq7nonzQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="ax6bvpqfd5xurhhl"
Content-Disposition: inline
In-Reply-To: <CAJF=k3=D3vguoC6ZhGJDQaRinJTU0H60zRyCQoT+fFVq7nonzQ@mail.gmail.com>


--ax6bvpqfd5xurhhl
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Interesting catch. I have some thoughts below.

On 12/21, =D0=98=D0=BB=D1=8C=D1=8F wrote:
>class ImpersonateMiddleware(MiddlewareMixin):
>    def process_request(self, request):
>        request.user.is_impersonate =3D False  # <----- this one
>        request.impersonator =3D None
>
>        if is_authenticated(request.user) and \  # <----- and this one
>           '_impersonate' in request.session:
>            new_user_id =3D request.session['_impersonate']
>            .....
>
>As a quick-and-dirty fix we can just remove the first line - but in
>client code we would access this param as `getattr(request.user,
>'is_impersonate', False)`.

The issue with changing this is it will break the "API" (if we want to
call it that) of the app and no doubt break deployments for end users
who are expecting this value.

>The second one is solved by placing `'_impersonate' in
>request.session` as a first condition.

This makes sense. I've pushed this change.

>I think for many people impersonated requests are a very little amount
>of the whole load so it could help to improve performance.

Agreed that every query counts. Though the single query probably isn't
the end of the world for 99.9% of cases.

--ax6bvpqfd5xurhhl
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=cyXu
-----END PGP SIGNATURE-----

--ax6bvpqfd5xurhhl--