Обсуждение: pgAdmin3 feedback
Some feedback regarding pgAdmin3 (pgadmin3-20030614-Win32 from http://snake.pgadmin.org/ on Win2K). Connecting: When connecting to a database as non-superuser, it shows the message: An error has occured: ERROR: pg_shadow: permissin denied I don't think this is really an error. It is expected behaviour, but the current message might confude people. Maybe it should say something like "Insufficient permissions to enumerate users.", but it would be really neat if the users entry in the left pane just greyed out and the message only appeared when somebody clicks on it. Vacuum: It would be nice if vacuum had the option to do a verbose vacuum and the result was displayed in screen. Also, a separate button for analyze would be nice. View data: I really like the inline editing of data in tables. Much better then what it was. However, horizontal and vertical scrollbars only appear when I maximize the window. When are you going to release an official alfa/beta version? Keep up the good work, Jochem
Hi Jochem, > -----Original Message----- > From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl] > Sent: 16 June 2003 10:22 > To: pgadmin-hackers@postgresql.org > Subject: [pgadmin-hackers] pgAdmin3 feedback > > > Some feedback regarding pgAdmin3 (pgadmin3-20030614-Win32 from > http://snake.pgadmin.org/ on Win2K). Always appreciated. > Connecting: > When connecting to a database as non-superuser, it shows the > message: An error has occured: > ERROR: pg_shadow: permissin denied > I don't think this is really an error. It is expected > behaviour, but the > current message might confude people. Maybe it should say > something like > "Insufficient permissions to enumerate users.", but it would > be really > neat if the users entry in the left pane just greyed out and > the message > only appeared when somebody clicks on it. I quite agree with this one, however it's a bit tricky (to do properly) with PostgreSQL 7.3 as the message will change for localised versions, so it's difficult to reliably detect what went wrong. When 7.4 is at a usable state we will have a better way of doing this as it will include error codes, alternatively we could hack it by only enumerating users for superusers, or those with select permissions on pg_shadow. > Vacuum: > It would be nice if vacuum had the option to do a verbose > vacuum and the > result was displayed in screen. Also, a separate button for analyze > would be nice. Agreed (and added to todo). > View data: > I really like the inline editing of data in tables. Much better then > what it was. However, horizontal and vertical scrollbars only appear > when I maximize the window. Yes, we've seen this error before. A simple resize normally makes the scrollbars appear. Andreas: did you get a chance to look at this at all? Can we just switch the scrollbars on permanently? 99% of the time they would probably be there anyway... > When are you going to release an official alfa/beta version? I was hoping we could aim to coincide with the release/beta for PostgreSQL 7.4, but that might be a bit ambitious. Thanks for the feedback :-) Regards, Dave.
Dave Page wrote: >>Connecting: >>When connecting to a database as non-superuser, it shows the >>message: An error has occured: >>ERROR: pg_shadow: permissin denied >>I don't think this is really an error. It is expected >>behaviour, but the >>current message might confude people. Maybe it should say >>something like >>"Insufficient permissions to enumerate users.", but it would >>be really >>neat if the users entry in the left pane just greyed out and >>the message >>only appeared when somebody clicks on it. >> >> > >I quite agree with this one, however it's a bit tricky (to do properly) >with PostgreSQL 7.3 as the message will change for localised versions, >so it's difficult to reliably detect what went wrong. > >When 7.4 is at a usable state we will have a better way of doing this as >it will include error codes, alternatively we could hack it by only >enumerating users for superusers, or those with select permissions on >pg_shadow. > I changed the schema code to use pg_user wherever possible, and checking for access privilege to pg_shadow to try to resolve the encrypted password. >>Vacuum: >>It would be nice if vacuum had the option to do a verbose >>vacuum and the >>result was displayed in screen. Also, a separate button for analyze >>would be nice. >> >> > >Agreed (and added to todo). > > > >>View data: >>I really like the inline editing of data in tables. Much better then >>what it was. However, horizontal and vertical scrollbars only appear >>when I maximize the window. >> >> > >Yes, we've seen this error before. A simple resize normally makes the >scrollbars appear. Andreas: did you get a chance to look at this at all? >Can we just switch the scrollbars on permanently? 99% of the time they >would probably be there anyway... > Nasty wx... There already was some code for non-win32, and I activated it for win32 now. wxWindows will optimize away the scrollbars if not initially necessary, and they will never come back... Now the window is forced extra-small so that scrollbars are necessary right from the start. >>When are you going to release an official alfa/beta version? >> >> > >I was hoping we could aim to coincide with the release/beta for >PostgreSQL 7.4, but that might be a bit ambitious. > This pessimism is illegal ;-), we should go ahead to match that goal. Regards, Andreas
> -----Original Message----- > From: Andreas Pflug [mailto:Andreas.Pflug@web.de] > Sent: 16 June 2003 13:57 > To: Dave Page > Cc: pgadmin-hackers@postgresql.org > Subject: Re: [pgadmin-hackers] pgAdmin3 feedback > > > Dave Page wrote: > > I changed the schema code to use pg_user wherever possible, > and checking > for access privilege to pg_shadow to try to resolve the > encrypted password. Cool. Not the most elegant solution, but until we get error codes... > >I was hoping we could aim to coincide with the release/beta for > >PostgreSQL 7.4, but that might be a bit ambitious. > > > This pessimism is illegal ;-), we should go ahead to match that goal. Sorry. I shall punish myself immediately by rebuilding one of our W2K DC's :-( Regards, Dave.
Dave Page wrote:
>
> Agreed (and added to todo).
Is the todo list public (so I don't duplicate requests)?
> I was hoping we could aim to coincide with the release/beta for
> PostgreSQL 7.4, but that might be a bit ambitious.
I don't think it is that ambitious. Maybe it wouldn't have all the
features, but it seems pretty stable and it degrades gracefully
("feature not implemented" messages).
Jochem
> -----Original Message-----
> From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl]
> Sent: 16 June 2003 21:40
> To: Dave Page
> Cc: pgadmin-hackers@postgresql.org
> Subject: Re: [pgadmin-hackers] pgAdmin3 feedback
>
>
> Dave Page wrote:
> >
> > Agreed (and added to todo).
>
> Is the todo list public (so I don't duplicate requests)?
>
>
> > I was hoping we could aim to coincide with the release/beta for
> > PostgreSQL 7.4, but that might be a bit ambitious.
>
> I don't think it is that ambitious. Maybe it wouldn't have all the
> features, but it seems pretty stable and it degrades gracefully
> ("feature not implemented" messages).
Should only be about one of those now iirc (operator class properties).
Regards, Dave.
> -----Original Message----- > From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl] > Sent: 16 June 2003 21:40 > To: Dave Page > Cc: pgadmin-hackers@postgresql.org > Subject: Re: [pgadmin-hackers] pgAdmin3 feedback > > > Dave Page wrote: > > > > Agreed (and added to todo). > > Is the todo list public (so I don't duplicate requests)? Missed this bit - yes it is, it's here http://cvs.pgadmin.org/cgi-bin/viewcvs.cgi/pgadmin3/ Regards, Dave.
On Wednesday 18 June 2003 09:18, Dave Page wrote: > Missed this bit - yes it is, it's here > http://cvs.pgadmin.org/cgi-bin/viewcvs.cgi/pgadmin3/ Thanks! Jean-Michel
On Wednesday 18 June 2003 09:18, Dave Page wrote: > Missed this bit - yes it is, it's here > http://cvs.pgadmin.org/cgi-bin/viewcvs.cgi/pgadmin3/ Pease disguard my last message. I deleted the symlink and rebuild pgAdmin3. Here is the output: checking for wxWindows... yes configure: error: "Could not find wx-config in /usr" error: Bad exit status from /var/tmp/rpm-tmp.67648 (%build) RPM build errors: Bad exit status from /var/tmp/rpm-tmp.67648 (%build) Uploading files It seems like the ./configure script does not look for /usr/bin/wxgtk2ud-2.5-config Cheers, Jean-Michel
On Wednesday 18 June 2003 15:05, Dave Page wrote:
> Yes, you need the symlink. It's there to provide a consistent name for
> the config script no matter what build options wx was compiled with. If
> you want to run with a non-standard wx installation, or the non-default
> build, perhaps you can persuade Adam to add --with-wx-config= option to
> override the script name?
Yes, but the name of the wxGTK library can be guessed from these options:
--enable-unicode => u
--enable-debug => d
/usr/bin/wxgtk2{u}{d}-2.5-config
Users can be several wxGTK versions at once. One symlink cannot point to all
version of wxGTK installed. For example, a user may need to install:
- wxGTK 2.4 in shared mode,
- wxGTK 2.5 cvs in shared mode with Unicode.
A Symlink cannot be part of a wxGTK2.5 or a pgAdmin3 package. Or it would
break existing wxGTK2.4 packages.
Cheers,
Jean-Michel
> -----Original Message-----
> From: Jean-Michel POURE [mailto:jm.poure@freesurf.fr]
> Sent: 18 June 2003 21:00
> To: pgadmin-hackers@postgresql.org
> Cc: Adam H. Pendleton; Jochem van Dieten; Dave Page
> Subject: Re: [pgadmin-hackers] pgAdmin3 feedback
>
>
> On Wednesday 18 June 2003 15:05, Dave Page wrote:
> > Yes, you need the symlink. It's there to provide a
> consistent name for
> > the config script no matter what build options wx was
> compiled with.
> > If you want to run with a non-standard wx installation, or the
> > non-default build, perhaps you can persuade Adam to add
> > --with-wx-config= option to override the script name?
>
> Yes, but the name of the wxGTK library can be guessed from
> these options: --enable-unicode => u --enable-debug => d
>
> /usr/bin/wxgtk2{u}{d}-2.5-config
>
> Users can be several wxGTK versions at once. One symlink
> cannot point to all
> version of wxGTK installed. For example, a user may need to install:
>
> - wxGTK 2.4 in shared mode,
> - wxGTK 2.5 cvs in shared mode with Unicode.
>
> A Symlink cannot be part of a wxGTK2.5 or a pgAdmin3 package.
> Or it would
> break existing wxGTK2.4 packages.
In which case an optional configure flag allows the user to select the
required config script. Whilst you can guess unicode and debug filenames
from the pgAdmin configure options, you cannot figure out what version
of wx is required (2.4 or 2.5), or if the user wants non-unicode linked
against GTK2.
For most people, the current behaviour fits with a standard wx
installation. For others, we can provide an override mechanism.
Regards, Dave.
Dave Page wrote: >From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl] >> >>Is the todo list public (so I don't duplicate requests)? > > Missed this bit - yes it is, it's here > http://cvs.pgadmin.org/cgi-bin/viewcvs.cgi/pgadmin3/ Some more feature requests (currently running pgadmin3-20030626-Win32): - refresh button in the "View data" window - SSL support - telepathy plugin - sort users alphabetically in tree view Jochem
> -----Original Message----- > From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl] > Sent: 27 June 2003 12:44 > To: pgadmin-hackers@postgresql.org > Subject: Re: [pgadmin-hackers] pgAdmin3 feedback > > > Dave Page wrote: > >From: Jochem van Dieten [mailto:jochemd@oli.tudelft.nl] > >> > >>Is the todo list public (so I don't duplicate requests)? > > > > Missed this bit - yes it is, it's here > > http://cvs.pgadmin.org/cgi-bin/viewcvs.cgi/pgadmin3/ > > Some more feature requests (currently running > pgadmin3-20030626-Win32): > - refresh button in the "View data" window Sounds reasonable. I'll add it to the todo. > - SSL support OK, should be there on the next Slackware builds. Jean-Michel, Andreas - can you enable ssl on your builds please? > - telepathy plugin Mark was working on that. As I can't sense how he's getting on, I'll assume the dev code has some bugs. > - sort users alphabetically in tree view Done (and groups). Regards, Dave.