Re: SQLAlchemy updates for check tables.

On Tue, Jul 20, 2021 at 12:15 PM Khushboo Vashi <
khushboo(dot)vashi(at)enterprisedb(dot)com> wrote:

>
>
> On Mon, Jul 19, 2021 at 3:14 PM Dave Page <dpage(at)pgadmin(dot)org> wrote:
>
>>
>>
>> On Mon, Jul 19, 2021 at 9:58 AM Khushboo Vashi <
>> khushboo(dot)vashi(at)enterprisedb(dot)com> wrote:
>>
>>>
>>>
>>> On Mon, Jul 19, 2021 at 2:17 PM Dave Page <dpage(at)pgadmin(dot)org> wrote:
>>>
>>>> Hi
>>>>
>>>> On Mon, Jul 19, 2021 at 6:37 AM Nikhil Mohite <
>>>> nikhil(dot)mohite(at)enterprisedb(dot)com> wrote:
>>>>
>>>>> Hi Hackers,
>>>>>
>>>>> Please find the attached patch for SQLAlchemy updates for check table
>>>>> is present in the database or not. (This will resolve load and dump server.)
>>>>>
>>>>
>>>> How come this upstream change didn't fail the regression tests?
>>>>
>>> Flask-SQLAlchemy is dependent on SQLAlchemy and which is an indirect
>>> dependency of pgAdmin, so if the installed version of Flask-SQLAlchemy is
>>> the latest one, it will be skipped.
>>>
>>
>> Sure, but the regression test runs on the buildfarm build the venv from
>> scratch on every run (as happens when we build the packages themselves). So
>> I can see why local regression runs might have passed (as developers
>> generally don't recreate their venv's from scratch before testing), but I
>> would expect to have seen failures on the buildfarm.
>>
>> The reason for not failing the test cases is, the old version of
> SQLAlchemy is being installed as a Flask-Migration dependency through the
> requirements.txt file. So, if we explicitly install Flask-SQLAlchemy, then
> only this issue is reproducible.
>

OK, so how did an installed copy of pgAdmin on a user machine get into that
state?

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

EDB: https://www.enterprisedb.com