Aug 25, 2020, 2:59 PM by dp...@pgadmin.org:

>
>
> On Fri, Aug 21, 2020 at 11:23 AM <> tutilu...@tutanota.com> > wrote:
>  
>
>> Please consider testing your software before releasing it.
>>
>
> https://pgsnake.blogspot.com/2020/08/testing-pgadmin.html>  
>

This certainly convinces me that there is quite a lot of testing happening, so 
that's at least reassuring in some sense. However, since it's necessary (in 
practice) to create a separate, dedicated browser profile for pgAdmin (since 
otherwise, it forgets the entire "state" every time you clear your browser data 
or close the browser, which happens constantly), breaking the "browser command" 
in a new version is quite remarkable.

Also, it should be noted that my found work-around, to find the pgAdmin icon in 
the Notification area, right click it and then click "New pgAdmin window", only 
works once you are actually running it. When I start my machine, pgAdmin isn't 
running, so I first have to launch it using my normal Taskbar icon, which up 
until the latest version opened the correct pgAdmin browser profile. Now, it 
instead loads for some time and finally opens in the default browser (obviously 
with forgotten "state"). I then have to close it and then start it with the 
Notification area work-around.

Yes, I could make it run on boot, to save myself another click and some 
waiting, but again, the problem isn't that I cannot find a way at all to use 
pgAdmin -- the issue is that such an "obvious" thing broke. It really makes me 
wonder how anyone could be running pgAdmin in their standard browser profile. I 
guess they never or rarely clear their browser data and never have to close all 
browser windows. I frequently need to do that for many reasons besides privacy, 
including updates, freezes/crashes (most frequently caused by pgAdmin, 
ironically), getting the "right order" of grouped windows of different browser 
profiles, etc.

As I explained previously, it's impossible for me to use a "supported browser" 
because Chrome (and all its "skins" which pretend to be browsers) as well as 
Firefox are pure spyware. I don't say that without reason, but I'm not going to 
go into detail about that again here. At the end of the day, I'm forced to use 
Pale Moon or nothing at this point, and pgAdmin either hangs entirely or 
freezes for many, many seconds (half a minute or more is not uncommon) if I 
forget myself and try to click and resize the object tree pane to make me able 
to see what it contains. I have to actively remember to just scroll 
horizontally or else I can say "good bye" to that entire pgAdmin session. Which 
has many times caused loss of work/state for me. The same thing happens even if 
I just maximize/restore the window. The most likely cause is some JavaScript 
code used to "redraw" or "recalculate" the view.

As you can see, I have extremely good reasons for wanting pgAdmin to ship with 
its own GUI/webview, and I frankly don't understand the stated reasons for why 
this is not done. I don't think you're lying, but NW.js (for example) uses 
Chrome/Chromium's engine and should not be possible to have any issues 
rendering and handling pgAdmin on all supported OSes. (I don't mention Electron 
because its developer is extremely toxic.)

Yes, I'm aware that pgAdmin can be run in a "hosted" manner, so it still has to 
support "other browsers" (whatever that means at this point with Google's 
engine having a total monopoly besides a minimal Firefox and Safari user base), 
but then you could at least say that there's always the option to download the 
"stand-alone" version of pgAdmin which comes with a nice GUI/webview and never 
has to interfere in any way with existing browsers and all the nightmares that 
entails.

The fact that you, the developers, don't see this as the #1 priority makes me 
wonder how it's possible that my "workflow" is apparently so fundamentally 
different from yours. Note that I'm not bashing the entire concept of "web 
apps", as this is what I have the most experience with myself, but simply the 
reluctance of packaging it in such a manner that it can be used without 
piggybacking on other software.

I actually remember trying it out early on when it was still a stand-alone 
thing, and while it was horribly slow and buggy, I never attributed this to the 
fact that it ran in some kind of webview. That makes no sense to me. My browser 
is an old fork of Firefox, maintained by "some guy in his basement", and I use 
it solely out of having no other choice. How can a webview which simply uses 
the Chromium engine (as evil as I find it, but that's a different problem) 
possibly be slower at rendering pgAdmin, which was made to support Chrome? 
Something about that doesn't add up.

pgAdmin III is entirely unusable at this point, and none of the "alternatives" 
to pgAdmin 4 are usable (for a number of reasons which are also pointless to 
list). I thus consider pgAdmin 4 to be the "official" and *only* software to 
administrate PostgreSQL databases. The CLI tool shipping with PG is not usable 
for somebody like me who cannot memorize syntax, and comes with all the 
limitations and problems of any CLI tool, and the last thing I want to do (or 
have time for) is sit and code my own tool just for myself. There are a number 
of things in pgAdmin 4 which would be a massive pain to reimplement.

To say something positive about pgAdmin 4, one of the best things ever, which 
was sorely lacking in pgAdmin III, is the ability to mark rows for deletion and 
to delete them from any "result view", as well as making edits of cells in a 
natural way.

It was probably going too far when I claimed at some point that the pgAdmin 
developers are doing this "on purpose, out of sadism", but it sometimes really 
feels like that when software authors do various things which seem just beyond 
all rhyme and reason. I think that many users of software in general would 
agree with me in that we want stability far, far, *far* more than "new 
features", once a minimum working environment has been accomplished. For 
example, Windows 10 is an ever-changing nightmare of bloat and broken nonsense. 
They just keep piling on garbage when they should have long since gone back to 
Windows 95-era polish, consistency and quality control.

Reply via email to