From ar at zeit.io  Fri May  3 14:58:38 2019
From: ar at zeit.io (Arup Rakshit)
Date: Sat, 4 May 2019 00:28:38 +0530
Subject: [Flask] Shouldn't is and == operator work same way while doing
 string comparisons?
Message-ID: <25DDD2C0-4ECA-4299-AFF0-4A64C51EADE6@zeit.io>

Hi,

I am facing a problem with `is` and `==` comparisons. The small below program gives different outputs.  When I use `if book['isbn'] is isbn]` I get an empty output, but `if book['isbn'] == isbn]` gives correct output. Why it is like that?

from flask import Flask, jsonify

app = Flask(__name__)

books = [
    {
        "name": 'Learning Python',
        'isbn': '1449355730',
        'price': 200
    },
    {
        "name": 'Practical Object-Oriented Design: An Agile Primer Using Ruby',
        'isbn': 'B07F88LY9M',
        'price': 210
    }
]


# ?.

@app.route('/books/<isbn>')
def get_book_by_isbn(isbn):
    return jsonify([{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] == isbn])


if __name__ == "__main__":
    app.run(debug=True)


Curl program:

~$ curl http://127.0.0.1:5000/books/B07F88LY9M
[]
~$ curl http://127.0.0.1:5000/books/B07F88LY9M
[
  {
    "name": "Practical Object-Oriented Design: An Agile Primer Using Ruby", 
    "price": 210
  }
]


Thanks,

Arup Rakshit
ar at zeit.io




From ar at zeit.io  Fri May  3 15:11:58 2019
From: ar at zeit.io (Arup Rakshit)
Date: Sat, 4 May 2019 00:41:58 +0530
Subject: [Flask] Shouldn't is and == operator work same way while doing
 string comparisons?
In-Reply-To: <CA+cbXRfdH5mnOJHnwWPDHZPf=UfhKO1s3aayacR5fhcx6+ra+w@mail.gmail.com>
References: <25DDD2C0-4ECA-4299-AFF0-4A64C51EADE6@zeit.io>
 <CA+cbXRfdH5mnOJHnwWPDHZPf=UfhKO1s3aayacR5fhcx6+ra+w@mail.gmail.com>
Message-ID: <9E9261F1-E2E8-44B0-8F12-6D0F0DE4F33A@zeit.io>

Hello Josh,

I tested the both while in REPL. They work same, why it is then different then flask app code?

Python 3.7.2 (v3.7.2:9a3ffc0492, Dec 24 2018, 02:44:43) 
[Clang 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> books = [
...     {
...         "name": 'Learning Python',
...         'isbn': '1449355730',
...         'price': 200
...     },
...     {
...         "name": 'Practical Object-Oriented Design: An Agile Primer Using Ruby',
...         'isbn': 'B07F88LY9M',
...         'price': 210
...     }
... ]
>>> [{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] is 'B07F88LY9M']
[{'name': 'Practical Object-Oriented Design: An Agile Primer Using Ruby', 'price': 210}]
>>> [{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] == 'B07F88LY9M']
[{'name': 'Practical Object-Oriented Design: An Agile Primer Using Ruby', 'price': 210}]
>>> isbn = 'B07F88LY9M'
>>> [{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] == isbn]
[{'name': 'Practical Object-Oriented Design: An Agile Primer Using Ruby', 'price': 210}]
>>> [{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] is isbn]
[{'name': 'Practical Object-Oriented Design: An Agile Primer Using Ruby', 'price': 210}]


Thanks,

Arup Rakshit
ar at zeit.io



> On 04-May-2019, at 12:35 AM, Josh Morrow <josh at jcmorrow.com> wrote:
> 
> https://stackoverflow.com/questions/132988/is-there-a-difference-between-and-is
> 
> On Fri, May 3, 2019 at 11:59 AM Arup Rakshit <ar at zeit.io> wrote:
> Hi,
> 
> I am facing a problem with `is` and `==` comparisons. The small below program gives different outputs.  When I use `if book['isbn'] is isbn]` I get an empty output, but `if book['isbn'] == isbn]` gives correct output. Why it is like that?
> 
> from flask import Flask, jsonify
> 
> app = Flask(__name__)
> 
> books = [
>     {
>         "name": 'Learning Python',
>         'isbn': '1449355730',
>         'price': 200
>     },
>     {
>         "name": 'Practical Object-Oriented Design: An Agile Primer Using Ruby',
>         'isbn': 'B07F88LY9M',
>         'price': 210
>     }
> ]
> 
> 
> # ?.
> 
> @app.route('/books/<isbn>')
> def get_book_by_isbn(isbn):
>     return jsonify([{'name': book['name'], 'price': book['price']} for book in books if book['isbn'] == isbn])
> 
> 
> if __name__ == "__main__":
>     app.run(debug=True)
> 
> 
> Curl program:
> 
> ~$ curl http://127.0.0.1:5000/books/B07F88LY9M
> []
> ~$ curl http://127.0.0.1:5000/books/B07F88LY9M
> [
>   {
>     "name": "Practical Object-Oriented Design: An Agile Primer Using Ruby", 
>     "price": 210
>   }
> ]
> 
> 
> Thanks,
> 
> Arup Rakshit
> ar at zeit.io
> 
> 
> 
> _______________________________________________
> Flask mailing list
> Flask at python.org
> https://mail.python.org/mailman/listinfo/flask


From ar at zeit.io  Fri May  3 15:23:46 2019
From: ar at zeit.io (Arup Rakshit)
Date: Sat, 4 May 2019 00:53:46 +0530
Subject: [Flask] flask-couchdb maintainers are around?
Message-ID: <FC46531D-83BC-42D3-BA16-44B34B9730B1@zeit.io>

I am recently messing around with Flask and couchDB. I used the package https://pythonhosted.org/Flask-CouchDB/#basic-use . But my program kept giving errors, which later fixed by someone from Reddit https://www.reddit.com/r/flask/comments/bk5g3w/python_flaskcouchdb_attributeerror_appctxglobals/emf6tyq/ . 

I followed what that Reddit guy said, and I met another problem https://www.reddit.com/r/flask/comments/bk5g3w/python_flaskcouchdb_attributeerror_appctxglobals/emfc6fl. Now he told me that it is related to incompatibility between python2 and python3 as per the PR https://bitbucket.org/leafstorm/flask-couchdb/pull-requests/2/fix-python-3-incompatibility/diff .

If maintainers are around would you mind to take a look at the project, and merge the PR. There is no other option I found to use couchdb with flask except this ORM.

Thanks,

Arup Rakshit
ar at zeit.io




From cs at cskk.id.au  Fri May  3 19:29:58 2019
From: cs at cskk.id.au (Cameron Simpson)
Date: Sat, 4 May 2019 09:29:58 +1000
Subject: [Flask] Shouldn't is and == operator work same way while doing
 string comparisons?
In-Reply-To: <9E9261F1-E2E8-44B0-8F12-6D0F0DE4F33A@zeit.io>
References: <9E9261F1-E2E8-44B0-8F12-6D0F0DE4F33A@zeit.io>
Message-ID: <20190503232958.GA2976@cskk.homeip.net>

On 04May2019 00:41, Arup Rakshit <ar at zeit.io> wrote:
>I tested the both while in REPL.

Testing only tells you what happens in your test, with a aprticular 
environment. It doesn't inform you about the language specification, 
except in so far as a failed test might tell you that your assumption 
was incorrect.

>They work same, why it is then different then flask app code?

Python interning small strings is an implementation detail, not part of 
the language specification. So there is no explicit guarantee that 2 
strings with the same value are the same string object, even for small 
strings.

If you want to act on the value of something, use ==. Using "is" is NOT 
FOR THAT. It is to check if 2 object references refer to the same 
object. This is rarely what you want; it is most useful with singletons 
such as sentinels.

Cheers,
Cameron Simpson <cs at cskk.id.au>

From gergely at polonkai.eu  Sat May  4 01:10:46 2019
From: gergely at polonkai.eu (Gergely Polonkai)
Date: Sat, 4 May 2019 07:10:46 +0200
Subject: [Flask] Shouldn't is and == operator work same way while doing
 string comparisons?
In-Reply-To: <20190503232958.GA2976@cskk.homeip.net>
References: <9E9261F1-E2E8-44B0-8F12-6D0F0DE4F33A@zeit.io>
 <20190503232958.GA2976@cskk.homeip.net>
Message-ID: <CACczBUK-LLVAP1EioP=aSe1ggh-_X6r=fmuCfC9y7axYKpDVNw@mail.gmail.com>

?in fact, i can easily give OP  a counterexample with strings:

>>> a = 'hello' + ' world'
>>> b = 'hello world'
>>> a
'hello world'
>>> b
'hello world'
>>> a == b
True
>>> a is b
False



Gergely Polonkai
[image: https://]about.me/gergely.polonkai
<https://about.me/gergely.polonkai?promo=email_sig&utm_source=email_sig&utm_medium=email_sig&utm_campaign=external_links>


lau., 4. ma? 2019 kl. 01:30 skrifa?i Cameron Simpson <cs at cskk.id.au>:

> On 04May2019 00:41, Arup Rakshit <ar at zeit.io> wrote:
> >I tested the both while in REPL.
>
> Testing only tells you what happens in your test, with a aprticular
> environment. It doesn't inform you about the language specification,
> except in so far as a failed test might tell you that your assumption
> was incorrect.
>
> >They work same, why it is then different then flask app code?
>
> Python interning small strings is an implementation detail, not part of
> the language specification. So there is no explicit guarantee that 2
> strings with the same value are the same string object, even for small
> strings.
>
> If you want to act on the value of something, use ==. Using "is" is NOT
> FOR THAT. It is to check if 2 object references refer to the same
> object. This is rarely what you want; it is most useful with singletons
> such as sentinels.
>
> Cheers,
> Cameron Simpson <cs at cskk.id.au>
> _______________________________________________
> Flask mailing list
> Flask at python.org
> https://mail.python.org/mailman/listinfo/flask
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/flask/attachments/20190504/1953da30/attachment-0001.html>

From tim at akwebsoft.com  Sun May  5 20:54:50 2019
From: tim at akwebsoft.com (Tim Johnson)
Date: Sun, 5 May 2019 16:54:50 -0800
Subject: [Flask] Shouldn't is and == operator work same way while doing
 string comparisons?
In-Reply-To: <CACczBUK-LLVAP1EioP=aSe1ggh-_X6r=fmuCfC9y7axYKpDVNw@mail.gmail.com>
References: <9E9261F1-E2E8-44B0-8F12-6D0F0DE4F33A@zeit.io>
 <20190503232958.GA2976@cskk.homeip.net>
 <CACczBUK-LLVAP1EioP=aSe1ggh-_X6r=fmuCfC9y7axYKpDVNw@mail.gmail.com>
Message-ID: <20190506005450.GJ2404@mail.akwebsoft.com>

* Gergely Polonkai <gergely at polonkai.eu> [190503 21:20]:
> ?in fact, i can easily give OP  a counterexample with strings:
> 
> >>> a = 'hello' + ' world'
> >>> b = 'hello world'
> >>> a
> 'hello world'
> >>> b
> 'hello world'
> >>> a == b
> True
> >>> a is b
> False
> 
> 
> 
> Gergely Polonkai
> [image: https://]about.me/gergely.polonkai
> <https://about.me/gergely.polonkai?promo=email_sig&utm_source=email_sig&utm_medium=email_sig&utm_campaign=external_links>
  Another way of looking at what Gergely puts it is that variables a
  and b may contain equal contents but they have different addresses
  in memory.
  
  They are in fact NOT the same object

-- 
Tim Johnson
http://www.tj49.com

From skip.montanaro at gmail.com  Tue May  7 14:48:25 2019
From: skip.montanaro at gmail.com (Skip Montanaro)
Date: Tue, 7 May 2019 13:48:25 -0500
Subject: [Flask] Getting a timestamp column wrong for an existing database
 table
Message-ID: <CANc-5Uy6NQFfrwaOeNV+6O5gv1bnA5DVrPnjvzdrY=zSr1mRUw@mail.gmail.com>

I'm trying to model an existing SQL Server table. It has a stamp
column of type datetime. I tried declaring it like so:

class Summary(DB.Model):
    ...
    stamp = DB.Column(DB.TIMESTAMP)
    ...

When I start my Flask server, it complains:

Attribute 'stamp' on class <class 'app.models.Summary'> appears to be
a non-schema 'sqlalchemy.sql.column()' object; this won't be part of
the declarative mapping

I've tried the following, all to no avail:

* DB.Column(DB.DATETIME)
* Changed the column in the database from datetime to datetime2
* DB.Column(SMALLDATETIME), where that comes from sqlalchemy.dialects.mssql.

How do I deal with this? I'm currently unable to use stamp as a query
constraint (at least I think so).

Skip Montanaro

From arunim29 at gmail.com  Mon May 13 05:36:40 2019
From: arunim29 at gmail.com (Arunim Chopra)
Date: Mon, 13 May 2019 15:06:40 +0530
Subject: [Flask] Python 2.6.9 with Flask 1.0.2
Message-ID: <CALniYZ-C1K93mvjJdmRq2jAD87ouuQALhP46yvCMA0diWyHUHg@mail.gmail.com>

Hi,
I need to work with python 2.6.9 . I wanted to know what will break if I
use Flask 1.0.2 with python 2.6.9 as it is no longer supported with this
Flask version.
Or is it better to use Flask 0.12 with python 2.6.9

Regards,
Arunim Chopra
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/flask/attachments/20190513/6963d880/attachment.html>

From lists at localguru.de  Tue May 21 06:42:40 2019
From: lists at localguru.de (Marcus Schopen)
Date: Tue, 21 May 2019 12:42:40 +0200
Subject: [Flask] get environment variables from mod_shib
Message-ID: <ac1534afd226d708f5a95116a5ada322a31e7bf4.camel@localguru.de>

Hi,

I'm using mod_shib for authentication and run my flask app via wsgi.
How do I securely access to shib attributes like "uid" or "displayName"
in my app? If I understood it correctly, it is insecure to pass the
attributes with RequestHeader and read them in Flask with
request.environ.get.

Ciao
Marcus



From lists at localguru.de  Tue May 21 07:56:20 2019
From: lists at localguru.de (Marcus Schopen)
Date: Tue, 21 May 2019 13:56:20 +0200
Subject: [Flask] get environment variables from mod_shib
In-Reply-To: <ac1534afd226d708f5a95116a5ada322a31e7bf4.camel@localguru.de>
References: <ac1534afd226d708f5a95116a5ada322a31e7bf4.camel@localguru.de>
Message-ID: <7c29a3af2369d670256a24a16579c1e6f28c2142.camel@localguru.de>

Am Dienstag, den 21.05.2019, 12:42 +0200 schrieb Marcus Schopen:
> Hi,
> 
> I'm using mod_shib for authentication and run my flask app via wsgi.
> How do I securely access to shib attributes like "uid" or
> "displayName"
> in my app? If I understood it correctly, it is insecure to pass the
> attributes with RequestHeader and read them in Flask with
> request.environ.get.

This seems to work:

   request.environ['displayName']


Is there a further spoof checking necessary?

Ciao
Marcus



From coreybrett at gmail.com  Sat May 25 09:50:39 2019
From: coreybrett at gmail.com (Corey Boyle)
Date: Sat, 25 May 2019 09:50:39 -0400
Subject: [Flask] Uwsgi emperor socket creation permissions
Message-ID: <CAD=ssw35zWjxLuRuv45xxuo83SE0KjL1e0CVD38S5iq_7GhE0Q@mail.gmail.com>

So I am trying to setup Uwsgi / Emperor for multiple apps each running
under their own user account.

I have the Emperor running as root in tyrant mode, and it's starting
the vassels and running them with their own accounts.

The trouble I have is the permissions on the sockets created by the
vassels. If I set chmod-socket = 666 everything works fine. The socket
is created with the vassels user and group, but Nginx (and everyone
else) is able to read/write because of the world permissions.

I just don't think that's a good situation.

What I can't figure out is how to have the vassels create the sockit
with permissions that will allow the appuser(uwsgi) and Nginx to
read/write, but keep everyone else out.

I tried using chown-socket = appuser:www-data, but that doesn't work
because appuser is not a member of the www-data group and therefore
can't set is as group.


Any suggestions?

From coreybrett at gmail.com  Sat May 25 12:10:48 2019
From: coreybrett at gmail.com (Corey Boyle)
Date: Sat, 25 May 2019 12:10:48 -0400
Subject: [Flask] Uwsgi emperor socket creation permissions
In-Reply-To: <CAE9kuxNMK4LqUEPFG8vku45rt3=gNwn2wXoHN+BubqcEM_UXkQ@mail.gmail.com>
References: <CAD=ssw35zWjxLuRuv45xxuo83SE0KjL1e0CVD38S5iq_7GhE0Q@mail.gmail.com>
 <CAE9kuxNMK4LqUEPFG8vku45rt3=gNwn2wXoHN+BubqcEM_UXkQ@mail.gmail.com>
Message-ID: <CAD=ssw1sjH4+_rmV2d6KFiqrTo6pZ_U7BLJsy3sgCSYbTMo2Zg@mail.gmail.com>

Thank you for the suggestion. I've read that the uwsgi binary protocol via
a Unix sockit is more efficient than HTTP. So I'd like to have my cake and
eat it too. ;-) HTTP may be my fallback plan if I can't figure this out.

On Sat, May 25, 2019 at 11:52 AM Rodrigo Ram?rez Norambuena <
decipher.hk at gmail.com> wrote:

> On Sat, May 25, 2019 at 9:51 AM Corey Boyle <coreybrett at gmail.com> wrote:
> >
> >  [...]
> >
> > I tried using chown-socket = appuser:www-data, but that doesn't work
> > because appuser is not a member of the www-data group and therefore
> > can't set is as group.
> >
> >
> > Any suggestions?
>
> You could running a wsgi process in http mode and Nginx get from this
> backend as proxy pass.
>
> Here a little example:
>
> [uwsgi]
> socket= 0.0.0.0:5000
> protocol=http
> master = true
> venv = /usr/local/apps/qpanel/env
> chdir = /usr/local/apps/qpanel
> mount = /qpanel=start.wsgi
> manage-script-name = true
> uid = x
> gid = x
> plugins = python
>
> Nginx
>     location /qpanel {
>
>         proxy_pass       http://localhost:5000/qpanel;
>         proxy_set_header Host      $host;
>         proxy_set_header X-Real-IP $remote_addr;
>
>     }
> Regards!
> --
> Rodrigo Ram?rez Norambuena
> http://www.rodrigoramirez.com/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/flask/attachments/20190525/e679e700/attachment.html>

From decipher.hk at gmail.com  Sat May 25 17:08:50 2019
From: decipher.hk at gmail.com (=?UTF-8?Q?Rodrigo_Ram=C3=ADrez_Norambuena?=)
Date: Sat, 25 May 2019 17:08:50 -0400
Subject: [Flask] Uwsgi emperor socket creation permissions
In-Reply-To: <CAD=ssw1sjH4+_rmV2d6KFiqrTo6pZ_U7BLJsy3sgCSYbTMo2Zg@mail.gmail.com>
References: <CAD=ssw35zWjxLuRuv45xxuo83SE0KjL1e0CVD38S5iq_7GhE0Q@mail.gmail.com>
 <CAE9kuxNMK4LqUEPFG8vku45rt3=gNwn2wXoHN+BubqcEM_UXkQ@mail.gmail.com>
 <CAD=ssw1sjH4+_rmV2d6KFiqrTo6pZ_U7BLJsy3sgCSYbTMo2Zg@mail.gmail.com>
Message-ID: <CAE9kuxNpFrb6P-s4CJRsEqgsWN9rap_r2u6Jh3sKS18chyuBtA@mail.gmail.com>

On Sat, May 25, 2019 at 12:11 PM Corey Boyle <coreybrett at gmail.com> wrote:
>
> Thank you for the suggestion. I've read that the uwsgi binary protocol via a Unix sockit is more efficient than HTTP. So I'd like to have my cake and eat it too. ;-) HTTP may be my fallback plan if I can't figure this out.
>

Should be more efficient...but really need the extra performance?

For this point, the uwsgi.ini file should be came a same permission
user:group.  I remember having a problem in Centos 7 for this
situation.

If run for the  normal users system you can move the socket another
places, example /var/run/.... with the Nginx user/group names
permission.  Or add the users in a group for uwsgi where www-data be
part of this group too

So, if you can share some information about your environment system
(O.S.), Python Version, some configuration could be help to take a
look
Regards!
-- 
Rodrigo Ram?rez Norambuena
http://www.rodrigoramirez.com/

From coreybrett at gmail.com  Sun May 26 10:11:40 2019
From: coreybrett at gmail.com (Corey Boyle)
Date: Sun, 26 May 2019 10:11:40 -0400
Subject: [Flask] Uwsgi emperor socket creation permissions
In-Reply-To: <CAE9kuxNpFrb6P-s4CJRsEqgsWN9rap_r2u6Jh3sKS18chyuBtA@mail.gmail.com>
References: <CAD=ssw35zWjxLuRuv45xxuo83SE0KjL1e0CVD38S5iq_7GhE0Q@mail.gmail.com>
 <CAE9kuxNMK4LqUEPFG8vku45rt3=gNwn2wXoHN+BubqcEM_UXkQ@mail.gmail.com>
 <CAD=ssw1sjH4+_rmV2d6KFiqrTo6pZ_U7BLJsy3sgCSYbTMo2Zg@mail.gmail.com>
 <CAE9kuxNpFrb6P-s4CJRsEqgsWN9rap_r2u6Jh3sKS18chyuBtA@mail.gmail.com>
Message-ID: <CAD=ssw0Ki8b-wQ9BZu4-itOpw+6mkPC64CdtSRTM1=x_7rEMYQ@mail.gmail.com>

Took me some experimentation, but found something that appears to work.

I created a new group and made appuser and www-data members.

Then I tried using "chown-socket=appuser:newgroup", but that didn't
work and made the Emperor curse the vassel.

I had to change the ownership of the vassel's ini file to
appuser:newgroup, and that did the trick.

I am using "chmod-socket=660" in the ini, and that seems to work fine.

Having to create the special group seems like an extra step, but I
guess that could be automated.

On Sat, May 25, 2019 at 5:09 PM Rodrigo Ram?rez Norambuena
<decipher.hk at gmail.com> wrote:
>
> On Sat, May 25, 2019 at 12:11 PM Corey Boyle <coreybrett at gmail.com> wrote:
> >
> > Thank you for the suggestion. I've read that the uwsgi binary protocol via a Unix sockit is more efficient than HTTP. So I'd like to have my cake and eat it too. ;-) HTTP may be my fallback plan if I can't figure this out.
> >
>
> Should be more efficient...but really need the extra performance?
>
> For this point, the uwsgi.ini file should be came a same permission
> user:group.  I remember having a problem in Centos 7 for this
> situation.
>
> If run for the  normal users system you can move the socket another
> places, example /var/run/.... with the Nginx user/group names
> permission.  Or add the users in a group for uwsgi where www-data be
> part of this group too
>
> So, if you can share some information about your environment system
> (O.S.), Python Version, some configuration could be help to take a
> look
> Regards!
> --
> Rodrigo Ram?rez Norambuena
> http://www.rodrigoramirez.com/