Обсуждение: Crash in edit grid
Hi developers! Hi Dave! Testing 1.6.1 rev: 5709, client Win XP. Host Debian Sarge, PG 8.1.4 A new version is out. Congratulations on your good work! :) I have run a few tests, found nothing bad, and used it for serious work. Found something bad though, when I was not looking. Here is what I did: From within the edit grid I applied a new filter. _Many_ rows were fetched, so it took a _long_ time before (new) data would be displayed. Got distracted, came back, edited filter and reapplied before the result of the previous filter had been displayed. pgAdmin crashed on me, and with it a dozen windows with unfinished work. :( Now, I know that a note is being displayed at the bottom left: "Refreshing data, please wait" while data is being fetched. But it is easily overlooked, and that's what I did. Trying to reproduce this (no difference if I apply a new filter from within or I open the edit grid with the filter option), it crashes occasionally. Most of the time I just get an error of this kind: An error has occurred: server closed the connection unexpectedly This probably means the server terminated abnormally before or whileprocessing the request Possible cure: Lock the edit grid while waiting for data (hourglass?), or at least prevent pgAdmin from crashing. But please don't leave this footgun loaded for the occasionally poor of sight, like me. Regards Erwin
mike@thegodshalls.com wrote: > Were you doing that over a vpn connection by chance? > (...) Hi Mike! No, this time I have been working via plain TCP/IP to port 5433 in the company intranet.From outside the company I usually connect via SSH-tunnel and port forwarding, which works just as well. I have not tested vpn, yet. Regards Erwin
Were you doing that over a vpn connection by chance? Pgadmin frequently will get disconnect for me if it is idle for a while when using it through a vpn connection. I will still be logged into the vpn and able to browse my network folders after it has happened so I know the vpn connection was not completely lost. Mike On Thu, 2006-11-30 at 23:16 +0100, Erwin Brandstetter wrote: > Hi developers! Hi Dave! > > Testing 1.6.1 rev: 5709, client Win XP. Host Debian Sarge, PG 8.1.4 > > A new version is out. Congratulations on your good work! :) > > I have run a few tests, found nothing bad, and used it for serious work. > Found something bad though, when I was not looking. Here is what I did: > > From within the edit grid I applied a new filter. > _Many_ rows were fetched, so it took a _long_ time before (new) data > would be displayed. > Got distracted, came back, edited filter and reapplied before the result > of the previous filter had been displayed. > pgAdmin crashed on me, and with it a dozen windows with unfinished work. :( > Now, I know that a note is being displayed at the bottom left: > "Refreshing data, please wait" while data is being fetched. But it is > easily overlooked, and that's what I did. > > Trying to reproduce this (no difference if I apply a new filter from > within or I open the edit grid with the filter option), it crashes > occasionally. Most of the time I just get an error of this kind: > > An error has occurred: > > server closed the connection unexpectedly > This probably means the server terminated abnormally > before or while processing the request > > > Possible cure: > Lock the edit grid while waiting for data (hourglass?), or at least > prevent pgAdmin from crashing. > But please don't leave this footgun loaded for the occasionally poor of > sight, like me. > > > Regards > Erwin > > > ---------------------------(end of broadcast)--------------------------- > TIP 3: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faq
Erwin Brandstetter wrote: > Lock the edit grid while waiting for data (hourglass?), or at least > prevent pgAdmin from crashing. Yeah, that's related to those changes I made *way* back when you first got involved in the testing. I disabled the refresh whilst the data was loading, but didn't consider the Find/Filter dialogue which can trigger a refresh itself. Now fixed, thanks. Regards, Dave.