Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtp (Exim 4.84_2) (envelope-from ) id 1d7gGz-0002Q4-4l for pgadmin-hackers@arkaria.postgresql.org; Mon, 08 May 2017 10:52:13 +0000 Received: from localhost ([127.0.0.1] helo=postgresql.org) by malur.postgresql.org with smtp (Exim 4.84_2) (envelope-from ) id 1d7gGy-00013j-NY for pgadmin-hackers@arkaria.postgresql.org; Mon, 08 May 2017 10:52:12 +0000 Received: from makus.postgresql.org ([2001:4800:1501:1::229]) by malur.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256) (Exim 4.84_2) (envelope-from ) id 1d7gGh-0000dI-8t for pgadmin-hackers@postgresql.org; Mon, 08 May 2017 10:51:55 +0000 Received: from mail-wm0-x22f.google.com ([2a00:1450:400c:c09::22f]) by makus.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA1:256) (Exim 4.84_2) (envelope-from ) id 1d7gGY-0000gL-R5 for pgadmin-hackers@postgresql.org; Mon, 08 May 2017 10:51:53 +0000 Received: by mail-wm0-x22f.google.com with SMTP id b84so49912317wmh.0 for ; Mon, 08 May 2017 03:51:46 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=enterprisedb-com.20150623.gappssmtp.com; s=20150623; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc; bh=Smkee8e+LU14mWKY8yfQPJjdj9hQsuV9NTgn9jGEj5o=; b=Y0FIrn9LpMB68FuhN3f8AayVbaX/VYxWyEmq5JbjQ8m82jd5h/JVIbqu6r3phO9Kc/ yTOnPDDH5P9c/2agysfFZSS11D8qpnr0hy08exor0kzwCxvrp6hwczJFTUMHEv0Cr1lj WyLqRT8RBntoGuIq444Wka0nM55adQFXRfq0wPQC6YDWCvACxge7pg5Fo+jPRmXRQ9W/ Be47N0G8t5oHP7YD87s2OSdnw9ku820LzVkdJ/otfHlIA9dUjjlAsDbuQPnWKZCyqgwX R6Nf9Y4udPnnCOPvC/w2lE1r55v0gnvepXhcQNAyM4zbixUfBWjeCD8StrMnHjWy/xds 3V7A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc; bh=Smkee8e+LU14mWKY8yfQPJjdj9hQsuV9NTgn9jGEj5o=; b=gTpgdGJ2V/xWS9sBbCtxAuKz8h/y0B5U+E69w+z1+LA9WH7+YhWviYUzWfOCOxpfOJ jQk/iIcMhgo6A2MJhCxRtsDDNDNEQ9cL7YRz9/C43Jc0kVbwggR0Oil9EfO37Q1pZ5s2 3FlTW+sBE30dGmhrdXIXcmJ+sSdzANJkWIZS1E78w2f6SpwTRXImzHu2hmc1+xghNX5U bgF6HHSfiKdigFvMQR7wlDqRY7IE+/LeLX6DujdNRZVcz3QX2hM2KTdxIQNyT8c/JtcS yi5Z0wtvJWWn91Gop0onQjwzS+dsLXgETK41aItb/hzLMtNQnXAqP6yjfwY5tgDfICfJ oiqA== X-Gm-Message-State: AN3rC/4y9oEqhrnkXRy3rfiWWv+ZqJpYSfHvsxsWJmR7h/WWNFHY68ND lQJG0MCB5bcRneRoeNFbORXZf8OVR7Bj X-Received: by 10.28.109.81 with SMTP id i78mr12064203wmc.24.1494240704706; Mon, 08 May 2017 03:51:44 -0700 (PDT) MIME-Version: 1.0 Received: by 10.28.19.138 with HTTP; Mon, 8 May 2017 03:51:14 -0700 (PDT) In-Reply-To: References: From: Surinder Kumar Date: Mon, 8 May 2017 16:21:14 +0530 Message-ID: Subject: Re: [pgAdmin4][Patch][RM2257]: Query tool - Insert row doesn't use default values To: Dave Page Cc: pgadmin-hackers Content-Type: multipart/alternative; boundary=001a1147c65070feb1054f01068a X-Pg-Spam-Score: -2.3 (--) List-Archive: List-Help: List-ID: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: X-Mailing-List: pgadmin-hackers Precedence: bulk Sender: pgadmin-hackers-owner@postgresql.org --001a1147c65070feb1054f01068a Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Hi On Mon, May 8, 2017 at 3:51 PM, Dave Page wrote: > Hi > > On Mon, May 8, 2017 at 11:13 AM, Surinder Kumar < > surinder.kumar@enterprisedb.com> wrote: > >> Hi Dave, >> >> On Mon, May 8, 2017 at 3:28 PM, Dave Page wrote: >> >>> Hi >>> >>> On Fri, May 5, 2017 at 12:52 PM, Surinder Kumar < >>> surinder.kumar@enterprisedb.com> wrote: >>> >>>> Hi Dave, >>>> >>>> The support to handle [null] and [default] values is added for >>>> following formatters: >>>> - JsonFormatter >>>> - NumbersFormatter >>>> - CheckmarkFormatter >>>> - TextFormatter >>>> >>>> Previously when a new row is added, it was not validating each and >>>> every cell for columns with attribute not_null =3D true. >>>> Introduced a new function validateRow(...) which is called on adding a >>>> new row, it validates each cell with column having >>>> attribute(not_null=3Dtrue). the corresponding cell will highlighted an= d save >>>> button will be disabled if value is [null]. >>>> >>> >>> I'm not sure that behaviour is right. What I now see (given the table >>> below) is that: >>> >>> - If I click in the id field of a new row, it forces me to enter a valu= e >>> or hit escape. Why is it trying to force me? It's a serial field, so wi= ll >>> get a value on save anyway. >>> >> =E2=80=8BYes, It is because I am validating all cell values after it ren= ders. It >> will reverted back. >> >>> >>> - If I do enter a value in the ID field, focus jumps over all the other >>> fields with either a default or no 'not null' option to the data_no_nul= ls >>> column. Focus should always go to the next field. >>> >>> I think this addition can just be removed - I'm pretty sure the previou= s >>> behaviour would have been what we want, with the additional formatters >>> fixed. >>> >> =E2=80=8BBut If i remove this addition, the value for column like 'data_= no_nulls' >> =E2=80=8Bwill be set to '' (blank string), then on save its value will b= e validated >> on the server side and whatever the error message is will be returned ba= ck >> (the same behaviour as in pgAdmin3) >> > > Which is fine I think. If you want to leave the validation there, that's > also fine - but, it a) shouldn't require me to press Esc if I decide not = to > fill in that value yet, and b) shouldn't change the tab order of the > fields. It's also broken of course, in that it was trying to force me to > enter a value for a not-null field with a default. > =E2=80=8BYes, I will check if we just highlight the field and don't force t= he user to =E2=80=8Benter value. This will enable user to edit any field using TAB key even if required field is highlighted red. Should the save button remains disable untill user enters any valid value in 'data_no_nulls' column ? > > >>> >>>> >>>> Now I will add more feature test cases for remaining formatters. Will >>>> send separate patch for feature test cases once completed. >>>> >>> >>> Thanks. >>> >>> >>>> >>>> Please review updated patch. >>>> >>>> >>>> On Tue, May 2, 2017 at 5:57 PM, Surinder Kumar < >>>> surinder.kumar@enterprisedb.com> wrote: >>>> >>>>> Hi Dave, >>>>> >>>>> On Tue, May 2, 2017 at 5:21 PM, Dave Page wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> This is looking much better now :-). Couple of thoughts and a bug: >>>>>> >>>>>> - Only the TextFormatter seems to handle both [null] and [default] >>>>>> values. Shouldn't all formatters do so (including Json and Checkmark= )? >>>>>> >>>>> =E2=80=8BYes, I will apply the same changes for other formatters too.= =E2=80=8B >>>>> >>>>>> For example, "serial" columns currently get displayed as [null] when >>>>>> left blank, but I would expect to see [default]. >>>>>> >>>>>> - I would suggest we put [null] and [default] in a lighter colour - >>>>>> #999999. >>>>>> >>>>>> - With the feature test patch added, I seem to be consistently >>>>>> getting the following failure (immediately after your new tests run)= : >>>>>> >>>>>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >>>>>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >>>>>> ERROR: runTest (pgadmin.feature_tests.xss_che >>>>>> cks_panels_and_query_tool_test.CheckForXssFeatureTest) >>>>>> Test XSS check for panels and query tool >>>>>> ------------------------------------------------------------ >>>>>> ---------- >>>>>> Traceback (most recent call last): >>>>>> File "/Users/dpage/git/pgadmin4/web/regression/feature_utils/base_= feature_test.py", >>>>>> line 42, in setUp >>>>>> self._screenshot() >>>>>> File "/Users/dpage/git/pgadmin4/web/regression/feature_utils/base_= feature_test.py", >>>>>> line 92, in _screenshot >>>>>> python_version)) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 802, in >>>>>> get_screenshot_as_file >>>>>> png =3D self.get_screenshot_as_png() >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 821, in >>>>>> get_screenshot_as_png >>>>>> return base64.b64decode(self.get_scre >>>>>> enshot_as_base64().encode('ascii')) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 831, in >>>>>> get_screenshot_as_base64 >>>>>> return self.execute(Command.SCREENSHOT)['value'] >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 238, in execute >>>>>> self.error_handler.check_response(response) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/errorhandler.py", line 193, in >>>>>> check_response >>>>>> raise exception_class(message, screen, stacktrace) >>>>>> UnexpectedAlertPresentException: Alert Text: None >>>>>> Message: unexpected alert open: {Alert text : Are you sure you wish >>>>>> to close the pgAdmin 4 browser?} >>>>>> (Session info: chrome=3D58.0.3029.81) >>>>>> (Driver info: chromedriver=3D2.29.461585 >>>>>> (0be2cd95f834e9ee7c46bcc7cf405b483f5ae83b),platform=3DMac OS X 10.12= .3 >>>>>> x86_64) >>>>>> >>>>>> >>>>>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >>>>>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >>>>>> ERROR: runTest (pgadmin.feature_tests.xss_che >>>>>> cks_pgadmin_debugger_test.CheckDebuggerForXssFeatureTest) >>>>>> Test table DDL generation >>>>>> ------------------------------------------------------------ >>>>>> ---------- >>>>>> Traceback (most recent call last): >>>>>> File "/Users/dpage/git/pgadmin4/web/regression/feature_utils/base_= feature_test.py", >>>>>> line 42, in setUp >>>>>> self._screenshot() >>>>>> File "/Users/dpage/git/pgadmin4/web/regression/feature_utils/base_= feature_test.py", >>>>>> line 92, in _screenshot >>>>>> python_version)) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 802, in >>>>>> get_screenshot_as_file >>>>>> png =3D self.get_screenshot_as_png() >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 821, in >>>>>> get_screenshot_as_png >>>>>> return base64.b64decode(self.get_scre >>>>>> enshot_as_base64().encode('ascii')) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 831, in >>>>>> get_screenshot_as_base64 >>>>>> return self.execute(Command.SCREENSHOT)['value'] >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/webdriver.py", line 238, in execute >>>>>> self.error_handler.check_response(response) >>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packa >>>>>> ges/selenium/webdriver/remote/errorhandler.py", line 193, in >>>>>> check_response >>>>>> raise exception_class(message, screen, stacktrace) >>>>>> UnexpectedAlertPresentException: Alert Text: None >>>>>> Message: unexpected alert open: {Alert text : Are you sure you wish >>>>>> to close the pgAdmin 4 browser?} >>>>>> (Session info: chrome=3D58.0.3029.81) >>>>>> (Driver info: chromedriver=3D2.29.461585 >>>>>> (0be2cd95f834e9ee7c46bcc7cf405b483f5ae83b),platform=3DMac OS X 10.12= .3 >>>>>> x86_64) >>>>>> >>>>> =E2=80=8BSure. I will fix this.=E2=80=8B >>>>> >>>>>> >>>>>> Thanks! >>>>>> >>>>>> >>>>>> On Fri, Apr 28, 2017 at 10:19 AM, Surinder Kumar < >>>>>> surinder.kumar@enterprisedb.com> wrote: >>>>>> >>>>>>> Hi Dave, >>>>>>> >>>>>>> Please find updated patch for RM case and a separate patch for >>>>>>> Feature tests. >>>>>>> >>>>>>> *Python:* >>>>>>> >>>>>>> - Added [default] label for cells with default values while >>>>>>> inserting a new row. >>>>>>> >>>>>>> - Introduced a FieldValidator function for cells that don't allow >>>>>>> null values. If user tries to insert null value, field with be high= lighted >>>>>>> with red borders around. >>>>>>> >>>>>>> =E2=80=8B- >>>>>>> If a cell contains blank string('') and when we set it to null, the >>>>>>> change into the cell is not detected. It was because the comparison >>>>>>> for (defaultValue =3D=3D null) return true if defaultValue is undef= ined. >>>>>>> Hence _.isNull(value) is used to fix this. >>>>>>> >>>>>>> *Feature Test cases:* >>>>>>> >>>>>>> - Introduced a new method create_table_with_query(server, db_name, >>>>>>> query) in test_utils.py which executes the given query on >>>>>>> connected server. >>>>>>> >>>>>>> - Added a new file test_data.json that has test data for test case= s. >>>>>>> >>>>>>> >>>>>>> On Fri, Apr 7, 2017 at 2:21 PM, Dave Page wrote= : >>>>>>> >>>>>>>> Hi >>>>>>>> >>>>>>>> On Sat, Apr 1, 2017 at 12:45 PM, Surinder Kumar >>>>>>>> wrote: >>>>>>>> > Hi >>>>>>>> > >>>>>>>> > Issues fixed: >>>>>>>> > >>>>>>>> > 1. If a column is defined with a default modifier, there is now >>>>>>>> way to >>>>>>>> > insert the row with those defaults. >>>>>>>> > The column will be left blank and it will take default value >>>>>>>> automatically. >>>>>>>> > >>>>>>>> > 2. If a column has a not-null constraint then an error is >>>>>>>> returned and the >>>>>>>> > row is not inserted. >>>>>>>> > The column will be left blank >>>>>>>> > >>>>>>>> > The default values for new added rows will be displayed on >>>>>>>> refresh/execute. >>>>>>>> > >>>>>>>> > Please find attached patch and review. >>>>>>>> >>>>>>>> This largely works as expected, but there is some weirdness. I hav= e >>>>>>>> a >>>>>>>> test table that looks like this: >>>>>>>> >>>>>>>> CREATE TABLE public.defaults >>>>>>>> ( >>>>>>>> id bigint NOT NULL DEFAULT nextval('defaults_id_seq'::reg >>>>>>>> class), >>>>>>>> data_default_nulls text COLLATE pg_catalog."default" DEFAULT >>>>>>>> 'abc123'::text, >>>>>>>> data_default_no_nulls text COLLATE pg_catalog."default" NOT NU= LL >>>>>>>> DEFAULT 'def456'::text, >>>>>>>> data_nulls text COLLATE pg_catalog."default", >>>>>>>> data_no_nulls text COLLATE pg_catalog."default" NOT NULL, >>>>>>>> CONSTRAINT defaults_pkey PRIMARY KEY (id) >>>>>>>> ) >>>>>>>> >>>>>>>> Remember that the expected behaviour is: >>>>>>>> >>>>>>>> - Set a value to empty to update the column to null. >>>>>>>> - Set a value to '' to update the column to an empty string >>>>>>>> - Set a value to anything else to update the column to that value >>>>>>>> >>>>>>>> 1) In a row with values in each column, if I try to set the value = of >>>>>>>> data_default_nulls to null, the query executed is: >>>>>>>> >>>>>>>> UPDATE public.defaults SET >>>>>>>> data_default_nulls =3D '' WHERE >>>>>>>> id =3D '2'; >>>>>>>> >>>>>>>> 2) If I do the same in the data_nulls column, the value is >>>>>>>> immediately >>>>>>>> shown as [null] and the query executed is: >>>>>>>> >>>>>>>> UPDATE public.defaults SET >>>>>>>> data_nulls =3D NULL WHERE >>>>>>>> id =3D '2'; >>>>>>>> >>>>>>>> 3) If I then edit the value in data_default_nulls, it shows the >>>>>>>> current value as ''. Removing the quotes (to set it to null) doesn= 't >>>>>>>> get detected as a change. >>>>>>>> >>>>>>> =E2=80=8B=E2=80=8BTaken care. >>>>>>> >>>>>>>> >>>>>>>> 4) When I manually executed "update defaults set data_default_null= s >>>>>>>> =3D >>>>>>>> null where id =3D 2" in a query tool window, I got: >>>>>>>> >>>>>>>> 2017-04-07 09:43:02,987: INFO werkzeug: 127.0.0.1 - - [07/Apr/2017 >>>>>>>> 09:43:02] "GET /sqleditor/columns/8745675 HTTP/1.1" 500 - >>>>>>>> Traceback (most recent call last): >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 2000, in __call__ >>>>>>>> return self.wsgi_app(environ, start_response) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1991, in wsgi_app >>>>>>>> response =3D self.make_response(self.handle_exception(e)) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1567, in handle_exception >>>>>>>> reraise(exc_type, exc_value, tb) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1988, in wsgi_app >>>>>>>> response =3D self.full_dispatch_request() >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1641, in full_dispatch_request >>>>>>>> rv =3D self.handle_user_exception(e) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1544, in handle_user_exception >>>>>>>> reraise(exc_type, exc_value, tb) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1639, in full_dispatch_request >>>>>>>> rv =3D self.dispatch_request() >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask/app.py", >>>>>>>> line 1625, in dispatch_request >>>>>>>> return self.view_functions[rule.endpoint](**req.view_args) >>>>>>>> File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-pack= a >>>>>>>> ges/flask_login.py", >>>>>>>> line 792, in decorated_view >>>>>>>> return func(*args, **kwargs) >>>>>>>> File "/Users/dpage/git/pgadmin4/web/pgadmin/tools/sqleditor/__in= i >>>>>>>> t__.py", >>>>>>>> line 452, in get_columns >>>>>>>> tid=3Dcommand_obj.obj_id) >>>>>>>> AttributeError: 'QueryToolCommand' object has no attribute 'obj_id= ' >>>>>>>> >>>>>>> =E2=80=8BFixed.=E2=80=8B >>>>>>> >>>>>>>> >>>>>>>> 5) When I run the query again in pgAdmin III, then refresh the dat= a >>>>>>>> in >>>>>>>> pgAdmin 4, the data_default_nulls column is displayed without the >>>>>>>> [null] marker (despite having a null value, which I confirmed in >>>>>>>> pgAdmin 3). >>>>>>>> >>>>>>> =E2=80=8BFixed.=E2=80=8B >>>>>>> >>>>>>>> >>>>>>>> I'm sure there are other combinations of issues here. Please fix a= nd >>>>>>>> thoroughly re-test to ensure behaviour is consistent - and to avoi= d >>>>>>>> future issues, please add some appropriate feature tests to check >>>>>>>> nulls, defaults and empty strings are properly handled in view, >>>>>>>> insert >>>>>>>> and updates. Murtuza recently wrote some feature tests for the que= ry >>>>>>>> tool - you should be able to use those as a starting point. >>>>>>>> >>>>>>> =E2=80=8BAdded feature tests=E2=80=8B >>>>>>> >>>>>>>> >>>>>>>> Thanks. >>>>>>>> >>>>>>>> -- >>>>>>>> Dave Page >>>>>>>> Blog: http://pgsnake.blogspot.com >>>>>>>> Twitter: @pgsnake >>>>>>>> >>>>>>>> EnterpriseDB UK: http://www.enterprisedb.com >>>>>>>> The Enterprise PostgreSQL Company >>>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Dave Page >>>>>> Blog: http://pgsnake.blogspot.com >>>>>> Twitter: @pgsnake >>>>>> >>>>>> EnterpriseDB UK: http://www.enterprisedb.com >>>>>> The Enterprise PostgreSQL Company >>>>>> >>>>> >>>>> >>>> >>> >>> >>> -- >>> Dave Page >>> Blog: http://pgsnake.blogspot.com >>> Twitter: @pgsnake >>> >>> EnterpriseDB UK: http://www.enterprisedb.com >>> The Enterprise PostgreSQL Company >>> >> >> > > > -- > Dave Page > Blog: http://pgsnake.blogspot.com > Twitter: @pgsnake > > EnterpriseDB UK: http://www.enterprisedb.com > The Enterprise PostgreSQL Company > --001a1147c65070feb1054f01068a Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Hi<= /div>

On Mon, May = 8, 2017 at 3:51 PM, Dave Page <dpage@pgadmin.org> wrote:
=
Hi

On Mon,= May 8, 2017 at 11:13 AM, Surinder Kumar <surinder.kumar@ent= erprisedb.com> wrote:
Hi Dave,

On Mon, May 8, 2017 at 3:28 PM, Dave Page &= lt;dpage@pgadmin.org= > wrote:
Hi

On Fri,= May 5, 2017 at 12:52 PM, Surinder Kumar <surinder.kumar@ent= erprisedb.com> wrote:
Hi Dave,

The support to ha= ndle [null] and [default] values is added for following formatters:
=C2=A0-=C2=A0JsonFormatter
=C2=A0-=C2=A0NumbersFormatter
=C2=A0-=C2=A0CheckmarkFormatter
=C2=A0-=C2=A0TextFormatter

<= div style=3D"font-size:small">Previously when a new row is added, it was no= t validating each and every cell for columns with attribute not_null =3D tr= ue.
Introduced a new function=C2=A0validateRow(...)=C2=A0which= is called on adding a new row, it validates each cell with column having a= ttribute(not_null=3Dtrue). the corresponding cell will highlighted and save= button will be disabled if value is [null].

I'm not sure that behaviour is right. What I now s= ee (given the table below) is that:

- If I click i= n the id field of a new row, it forces me to enter a value or hit escape. W= hy is it trying to force me? It's a serial field, so will get a value o= n save anyway.
=E2=80=8BYes, It is because I am validating all cell values a= fter it renders. It will reverted back.

- If I do enter a value in the ID field, focus jumps over all the = other fields with either a default or no 'not null' option to the d= ata_no_nulls column. Focus should always go to the next field.
I think this addition can just be removed - I'm pretty sur= e the previous behaviour would have been what we want, with the additional = formatters fixed.=C2=A0
=E2=80=8BBut If i remove this addition, the value fo= r column like 'data_no_nulls' =E2=80=8Bwill be set to '' (b= lank string), then on save its value will be validated on the server side a= nd whatever the error message is will be returned back (the same behaviour = as in pgAdmin3)

<= div>Which is fine I think. If you want to leave the validation there, that&= #39;s also fine - but, it a) shouldn't require me to press Esc if I dec= ide not to fill in that value yet, and b) shouldn't change the tab orde= r of the fields. It's also broken of course, in that it was trying to f= orce me to enter a value for a not-null field with a default.
=E2=80=8BYes, I will check if we just highlight the field and don't = force the user to =E2=80=8Benter value.
This will enable user to edit any field using TAB = key even if required field is highlighted red.
Should the save button remains disable unti= ll user enters any valid value in 'data_no_nulls' column ?
=C2=A0
=
<= span>
=C2=A0

Now I will add more feature test cases for remaining formatters. Will send= separate patch for feature test cases once completed.

Thanks.
=C2=A0

Please review upda= ted patch.


On Tue, May 2, 2017 at 5:57 PM, = Surinder Kumar <surinder.kumar@enterprisedb.com>= wrote:
Hi Dave,

On Tue, May 2, 2017 at 5:21 PM, Dave Page <= dpage@pgadmin.org> wrote:
Hi,

=
This is looking much better now :-). Couple of thoughts and a bug:

- Only the TextFormatter seems to handle both [null] = and [default] values. Shouldn't all formatters do so (including Json an= d Checkmark)?
For example, "serial" columns= currently get displayed as [null] when left blank, but I would expect to s= ee [default].=C2=A0

- I would suggest we put [null= ] and [default] in a lighter colour - #999999.

- W= ith the feature test patch added, I seem to be consistently getting the fol= lowing failure (immediately after your new tests run):

=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D
ERROR: runTest (pgadmin.feature_tests.xss_checks_p= anels_and_query_tool_test.CheckForXssFeatureTest)
Test XSS c= heck for panels and query tool
----------------------------------------------------------------------
Traceback (m= ost recent call last):
=C2=A0 File "/Users/dpage/git/= pgadmin4/web/regression/feature_utils/base_feature_test.py",= line 42, in setUp
=C2=A0 =C2=A0 self._screenshot()
=C2= =A0 File "/Users/dpage/git/pgadmin4/web/regression/feature_utils/= base_feature_test.py", line 92, in _screenshot
=C2=A0 = =C2=A0 python_version))
=C2=A0 File "/Users/dpage/.virtualen= vs/pgadmin4/lib/python2.7/site-packages/selenium/webdriver/remote= /webdriver.py", line 802, in get_screenshot_as_file
=C2= =A0 =C2=A0 png =3D self.get_screenshot_as_png()
=C2=A0 File "= ;/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packages/s= elenium/webdriver/remote/webdriver.py", line 821, in get_screensh= ot_as_png
=C2=A0 =C2=A0 return base64.b64decode(self.get_screenshot_as_base64().encode('ascii'))
=C2=A0 File &qu= ot;/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packages= /selenium/webdriver/remote/webdriver.py", line 831, in get_screen= shot_as_base64
=C2=A0 =C2=A0 return self.execute(Command.SCREENSH= OT)['value']
=C2=A0 File "/Users/dpage/.virtual= envs/pgadmin4/lib/python2.7/site-packages/selenium/webdriver/remo= te/webdriver.py", line 238, in execute
=C2=A0 =C2=A0 se= lf.error_handler.check_response(response)
=C2=A0 File "= /Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packages/se= lenium/webdriver/remote/errorhandler.py", line 193, in check_resp= onse
=C2=A0 =C2=A0 raise exception_class(message, screen, stacktr= ace)
UnexpectedAlertPresentException: Alert Text: None
=
Message: unexpected alert open: {Alert text : Are you sure you wish to= close the pgAdmin 4 browser?}
=C2=A0 (Session info: chrome=3D58.= 0.3029.81)
=C2=A0 (Driver info: chromedriver=3D2.29.461585 (0be2c= d95f834e9ee7c46bcc7cf405b483f5ae83b),platform=3DMac OS X 10.12.3 x86_6= 4)


=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
ERROR: runTest (pgadmin= .feature_tests.xss_checks_pgadmin_debugger_test.CheckDebuggerForX= ssFeatureTest)
Test table DDL generation
--------------= --------------------------------------------------------
Traceback (most recent call last):
=C2=A0 File &qu= ot;/Users/dpage/git/pgadmin4/web/regression/feature_utils/base_fe= ature_test.py", line 42, in setUp
=C2=A0 =C2=A0 self._screen= shot()
=C2=A0 File "/Users/dpage/git/pgadmin4/web/regre= ssion/feature_utils/base_feature_test.py", line 92, in _screensho= t
=C2=A0 =C2=A0 python_version))
=C2=A0 File "/Use= rs/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packages/seleni= um/webdriver/remote/webdriver.py", line 802, in get_screenshot_as= _file
=C2=A0 =C2=A0 png =3D self.get_screenshot_as_png()
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/s= ite-packages/selenium/webdriver/remote/webdriver.py", line 8= 21, in get_screenshot_as_png
=C2=A0 =C2=A0 return base64.b64decod= e(self.get_screenshot_as_base64().encode('ascii'))
<= div>=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7= /site-packages/selenium/webdriver/remote/webdriver.py", line= 831, in get_screenshot_as_base64
=C2=A0 =C2=A0 return self.execu= te(Command.SCREENSHOT)['value']
=C2=A0 File "/U= sers/dpage/.virtualenvs/pgadmin4/lib/python2.7/site-packages/sele= nium/webdriver/remote/webdriver.py", line 238, in execute
=C2=A0 =C2=A0 self.error_handler.check_response(response)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/si= te-packages/selenium/webdriver/remote/errorhandler.py", line= 193, in check_response

Thanks!


On Fri, Apr 28, 2017 at 10:= 19 AM, Surinder Kumar <surinder.kumar@enterprisedb.com<= /a>> wrote:
Hi Dave,

Please find updated patch for RM case and a separ= ate patch for Feature tests.

=
Python:

- Added=C2=A0[default] label for cells with default values while insert= ing a new row.

- Introduced a FieldValidator function for cells th= at don't allow null values. If user tries to insert null value, field w= ith be highlighted with red =C2=A0 =C2=A0borders around.

= =E2=80=8B-=C2=A0
If a cell contains blank string('') and when = we set it to null, the change into the cell is not detected. It was because= the comparison
for (defaultValue =3D=3D null) return true if defaultVal= ue is undefined. Hence _.isNull(value) is used to fix this.

Featu= re Test cases:

=C2=A0- Int= roduced a new method=C2=A0create_table_with_query(server, db_= name, query)=C2=A0 in test_utils.py which executes the given query o= n connected server.

=C2=A0- Added a new file test_= data.json that has test data for test cases.


On Fri, Apr 7, 2017 at 2:21 PM, Dave Page <dpage@pgadmin.org><= /span> wrote:
Hi

On Sat, Apr 1, 2017 at 12:45 PM, Surinder Kumar
<su= rinder.kumar@enterprisedb.com> wrote:
> Hi
>
> Issues fixed:
>
> 1. If a column is defined with a default modifier, there is now way to=
> insert the row with those defaults.
> The column will be left blank and it will take default value automatic= ally.
>
> 2. If a column has a not-null constraint then an error is returned and= the
> row is not inserted.
> The column will be left blank
>
> The default values for new added rows will be displayed on refresh/exe= cute.
>
> Please find attached patch and review.

This largely works as expected, but there is some weirdness. I have = a
test table that looks like this:

CREATE TABLE public.defaults
(
=C2=A0 =C2=A0 id bigint NOT NULL DEFAULT nextval('defaults_id_seq':= :regclass),
=C2=A0 =C2=A0 data_default_nulls text COLLATE pg_catalog."default"= ; DEFAULT 'abc123'::text,
=C2=A0 =C2=A0 data_default_no_nulls text COLLATE pg_catalog."default&q= uot; NOT NULL
DEFAULT 'def456'::text,
=C2=A0 =C2=A0 data_nulls text COLLATE pg_catalog."default",
=C2=A0 =C2=A0 data_no_nulls text COLLATE pg_catalog."default" NOT= NULL,
=C2=A0 =C2=A0 CONSTRAINT defaults_pkey PRIMARY KEY (id)
)

Remember that the expected behaviour is:

- Set a value to empty to update the column to null.
- Set a value to '' to update the column to an empty string
- Set a value to anything else to update the column to that value

1) In a row with values in each column, if I try to set the value of
data_default_nulls to null, the query executed is:

UPDATE public.defaults SET
data_default_nulls =3D '' WHERE
id =3D '2';

2) If I do the same in the data_nulls column, the value is immediately
shown as [null] and the query executed is:

UPDATE public.defaults SET
data_nulls =3D NULL WHERE
id =3D '2';

3) If I then edit the value in data_default_nulls, it shows the
current value as ''. Removing the quotes (to set it to null) doesn&= #39;t
get detected as a change.
=E2=80=8B=E2=80=8BTaken care.=C2=A0

4) When I manually executed "update defaults set data_default_nulls = =3D
null where id =3D 2" in a query tool window, I got:

2017-04-07 09:43:02,987: INFO werkzeug: 127.0.0.1 - - [07/Apr/2017
09:43:02] "GET /sqleditor/columns/8745675 HTTP/1.1" 500 -
Traceback (most recent call last):
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 2000, in __call__
=C2=A0 =C2=A0 return self.wsgi_app(environ, start_response)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1991, in wsgi_app
=C2=A0 =C2=A0 response =3D self.make_response(self.handle_exception(e)= )
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1567, in handle_exception
=C2=A0 =C2=A0 reraise(exc_type, exc_value, tb)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1988, in wsgi_app
=C2=A0 =C2=A0 response =3D self.full_dispatch_request()
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1641, in full_dispatch_request
=C2=A0 =C2=A0 rv =3D self.handle_user_exception(e)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1544, in handle_user_exception
=C2=A0 =C2=A0 reraise(exc_type, exc_value, tb)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1639, in full_dispatch_request
=C2=A0 =C2=A0 rv =3D self.dispatch_request()
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask/app.py",
line 1625, in dispatch_request
=C2=A0 =C2=A0 return self.view_functions[rule.endpoint](**req.view_arg= s)
=C2=A0 File "/Users/dpage/.virtualenvs/pgadmin4/lib/python2.7/sit= e-packages/flask_login.py",
line 792, in decorated_view
=C2=A0 =C2=A0 return func(*args, **kwargs)
=C2=A0 File "/Users/dpage/git/pgadmin4/web/pgadmin/tools/sqledito= r/__init__.py",
line 452, in get_columns
=C2=A0 =C2=A0 tid=3Dcommand_obj.obj_id)
AttributeError: 'QueryToolCommand' object has no attribute 'obj= _id'
=E2=80= =8BFixed.=E2=80=8B

5) When I run the query again in pgAdmin III, then refresh the data in
pgAdmin 4, the data_default_nulls column is displayed without the
[null] marker (despite having a null value, which I confirmed in
pgAdmin 3).
=E2=80=8B= Fixed.=E2=80=8B

I'm sure there are other combinations of issues here. Please fix and thoroughly re-test to ensure behaviour is consistent - and to avoid
future issues, please add some appropriate feature tests to check
nulls, defaults and empty strings are properly handled in view, insert
and updates. Murtuza recently wrote some feature tests for the query
tool - you should be able to use those as a starting point.
=E2=80=8BAdded feature tests=E2=80= =8B

Thanks.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company




--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsn= ake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Co= mpany





--
http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseD= B UK: http://www.= enterprisedb.com
The Enterprise PostgreSQL Company




--
Dave Page
Blog: = http://pgsnake.bl= ogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com<= br>The Enterprise PostgreSQL Company

--001a1147c65070feb1054f01068a--