Re: [pgadmin-support] making me love pgAdmin III - Mailing list pgadmin-support

On Fri, May 19, 2017 at 10:56 AM, Tomek <tomek@apostata.org> wrote:
>>> You don't care what You users say about Your software??? That's a nice statement...
>>
>> Twisting my words is a great way to make your point. Kudos.
>
> Twisting... Saying "I'm not going to lose sleep over it" means what - in case of someones concerns?

No, it means I'm not going to lose sleep if someone doesn't like
pgAdmin, as you cannot please everyone - that's a fact of life. That
doesn't mean I don't care if users provide useful feedback.

>>> I've posted here a list of what is missing/wrong in new pgAdmin - did You even commented on that?
>>
>> Posted here? The website is quite clear that bugs and feature requests
>> should be logged as tickets on our Redmine instance. There is no
>> guarantee developers will read every message here.
>
> I get the bugtracker - but my first list contains 19 cases (from one day of using) - cases that
> every other dbms has, and previous version had...

OK, so that means we have to use a different process for you? There
are very good reasons why we ask people to use the bug tracker - not
least because it allows us to properly track them.

>> Feedback along the lines of "It's crap and I hate it" is unhelpful and
>> will likely be ignored.
>
> Nobody said this - we all gave examples what is wrong with it...

Perhaps you should re-read the thread.

>>> How 4 is more stable than 3? Please explain it to us...
>>
>> Because it doesn't crash every 5 minutes? Because it doesn't go nuts
>> if it loses an open database connection?
>
> That doesn't mean more stable - that means You've fixed the bug (which everybody got used to)...

Not breaking unexpectedly is the very definition of "stable".

>>> In 1.4 - query SELECT 2/0; returns
>>> successfully...
>>
>> Oh? I haven't seen that logged. When I try it, I get the result in the
>> attached screenshot.
>
> https://ibb.co/eBbtdF

In 1.4?

>>> And please, please explain to us how 'less and slower' is better than 'more and faster'...
>>
>> The vast majority of the "more" wasn't used by anyone. Did you ever
>> create an Operator Class? Or use pgScript for example?
>
> Look at the list above...

That doesn't answer or invalidate my question at all.

>> As for the speed, yes, it's a little slower due to the architecture,
>> but we're talking fractions of a second (at least on all the machines
>> and virtual machines I've tested on). I know I still can't keep up
>> with it. The only exception I know of is the query tool with very
>> large results sets, which we've been working on and already have a WIP
>> patch that makes it ~20x *faster* than pgAdmin 3 in some large test
>> cases.
>
> I don't see it - I see query that is running 14 seconds in v4 but 3.5 seconds in v3, seconds pass
> before new query tab opens, seconds pass before browse tab opens, click properties count to 10 -
> properties appear, table properties - You can see switches flipping on columns tab, try moving any
> window - drags behind mouse, try put back accidentally detached tab, and so on...

With the exception of the query tool, which I noted above, and the
couple of seconds it takes to initialise, all of the above is so close
to instant for me that I can't begin to time it.

> In v3 3 dbs connected 4 query windows, 2 browse windows: private memory: 35 MBytes
> In v4 1 db connected 1 query window: private memory: 600 MBytes
> In v4 constant CPU usage around 10% (damn dashboard)

https://www.pgadmin.org/faq/#5

>> Oh? You have a better idea for something that we wanted to be able to
>> run over the web? Java or Flash perhaps?
>
> My objection for html is for standalone app - not webapp.
> You know that every major databases does not use web as primary dbms yes? But even if we use it as
> a server side only (like phppgadmin) it is still slow...

I'm sorry - I thought we were free to write and give away at no cost
whatever software we wanted.

Seriously, this is Open Source. Contributors work on what they want,
and clearly (based on the number of contributors we have now) more
people want to work on newer technology that can be run in multiple
different modes - and based on the download numbers, most users aren't
disagreeing with our choices. We could have just quit and gone and
done something else instead but instead, people have put in well over
20,000 hours of effort at this point to give you something for free.

I'm more than happy to help when bug reports are logged, and to make
improvements where issues can be demonstrated and reproduced. However,
if that isn't good enough for you then I'd suggest that maybe you
would prefer to use an alternative tool.