Thread: OK button on restore
This may be more of a feature request/suggestion than a bug. But here's the issue: After finishing a restore using the Windows version (tested in 1.6.3, but was also present in 1.4.x), if you click the "OK" button when you're done, the restore process runs a 2d time--which is not desirable. I've used the program enough to know I should select "CANCEL" instead. But the distinction between OK and CANCEL (or "DONE", which I believe I've seen in some versions) is not totally clear. The bottom line is it's far too easy for a user to click OK thinking they're doing the right thing, when in fact it's not the right thing to do. I'm sure there are multiple ways to address this issue. One thought I had would be to throw a message whenever the OK button is selected. Something along the lines of "Clicking OK will cause the restore operation to be run using the specified backup file. Do you wish to continue? Y/N. HINT: To exit the restore screen select CANCEL." Best regards, Pierce -- Pierce Tyler
Pierce Tyler wrote: > This may be more of a feature request/suggestion than a bug. But here's > the issue: After finishing a restore using the Windows version (tested > in 1.6.3, but was also present in 1.4.x), if you click the "OK" button > when you're done, the restore process runs a 2d time--which is not > desirable. I've used the program enough to know I should select "CANCEL" > instead. But the distinction between OK and CANCEL (or "DONE", which I > believe I've seen in some versions) is not totally clear. The bottom > line is it's far too easy for a user to click OK thinking they're doing > the right thing, when in fact it's not the right thing to do. If the restore doesn't return a success code the button stays at OK so you can tweak the options and try again. If it does return a success code (zero), the button changes to Done. The same applies to the backup dialogues. Regards, Dave
That makes sense. But it still might be helpful to add an additional safeguard (like a pop-up message, for example). I've seen cases where the log returns only one error--something insignificant--and the user decides they can live with it. So they click OK, thinking they're exiting the screen, moving on. But instead they may have inadvertently corrupted their database. Dave Page wrote: > Pierce Tyler wrote: >> This may be more of a feature request/suggestion than a bug. But here's >> the issue: After finishing a restore using the Windows version (tested >> in 1.6.3, but was also present in 1.4.x), if you click the "OK" button >> when you're done, the restore process runs a 2d time--which is not >> desirable. I've used the program enough to know I should select "CANCEL" >> instead. But the distinction between OK and CANCEL (or "DONE", which I >> believe I've seen in some versions) is not totally clear. The bottom >> line is it's far too easy for a user to click OK thinking they're doing >> the right thing, when in fact it's not the right thing to do. > > If the restore doesn't return a success code the button stays at OK so > you can tweak the options and try again. If it does return a success > code (zero), the button changes to Done. The same applies to the backup > dialogues. > > Regards, Dave > -- Pierce Tyler Dir. of Product Support OpenMFG, LLC 757.461.3022, ext. 103 www.openmfg.com
I agree with Pierce, instead of OK buttom we can have another message like "start restore" or just "restore". I don't remember how much times inadvertently press OK buttom to close this screen. (Yeah, I was thinking that this occur only with me :P ) This is a little change, but will help a lot. TIA. Pierce Tyler wrote: > That makes sense. But it still might be helpful to add an additional > safeguard (like a pop-up message, for example). I've seen cases where > the log returns only one error--something insignificant--and the user > decides they can live with it. So they click OK, thinking they're > exiting the screen, moving on. But instead they may have inadvertently > corrupted their database. > > > Dave Page wrote: >> Pierce Tyler wrote: >>> This may be more of a feature request/suggestion than a bug. But here's >>> the issue: After finishing a restore using the Windows version (tested >>> in 1.6.3, but was also present in 1.4.x), if you click the "OK" button >>> when you're done, the restore process runs a 2d time--which is not >>> desirable. I've used the program enough to know I should select >>> "CANCEL" >>> instead. But the distinction between OK and CANCEL (or "DONE", which I >>> believe I've seen in some versions) is not totally clear. The bottom >>> line is it's far too easy for a user to click OK thinking they're doing >>> the right thing, when in fact it's not the right thing to do. >> >> If the restore doesn't return a success code the button stays at OK so >> you can tweak the options and try again. If it does return a success >> code (zero), the button changes to Done. The same applies to the backup >> dialogues. >> >> Regards, Dave >> > -- Luiz K. Matsumura Plan IT Tecnologia Informática Ltda.
Yep I agree, with pg_restore, warnings also generate a non-zero return code, but the restore does complete successfully, so this is mis-leading. An example is "language <xxx> already exists" - pg_restore just throws a warning to let you know this, but the restore is successful. Perhaps what might work better is a message box - if you've already performed a restore but it returned non-zero, and OK is pressed again in the same dialogue box, throw up a message saying something like "you've already attempted a restore, are you sure you want to run it again?" Andy. Luiz K. Matsumura wrote: > I agree with Pierce, instead of OK buttom we can have another message > like "start restore" or just "restore". > I don't remember how much times inadvertently press OK buttom to close > this screen. (Yeah, I was thinking that this occur only with me :P ) > This is a little change, but will help a lot. > > TIA. > > > Pierce Tyler wrote: >> That makes sense. But it still might be helpful to add an additional >> safeguard (like a pop-up message, for example). I've seen cases where >> the log returns only one error--something insignificant--and the user >> decides they can live with it. So they click OK, thinking they're >> exiting the screen, moving on. But instead they may have >> inadvertently corrupted their database. >> >> >> Dave Page wrote: >>> Pierce Tyler wrote: >>>> This may be more of a feature request/suggestion than a bug. But >>>> here's >>>> the issue: After finishing a restore using the Windows version (tested >>>> in 1.6.3, but was also present in 1.4.x), if you click the "OK" button >>>> when you're done, the restore process runs a 2d time--which is not >>>> desirable. I've used the program enough to know I should select >>>> "CANCEL" >>>> instead. But the distinction between OK and CANCEL (or "DONE", which I >>>> believe I've seen in some versions) is not totally clear. The bottom >>>> line is it's far too easy for a user to click OK thinking they're >>>> doing >>>> the right thing, when in fact it's not the right thing to do. >>> >>> If the restore doesn't return a success code the button stays at OK so >>> you can tweak the options and try again. If it does return a success >>> code (zero), the button changes to Done. The same applies to the backup >>> dialogues. >>> >>> Regards, Dave >>> >> >
Andy Shellam wrote: > Yep I agree, with pg_restore, warnings also generate a non-zero return > code, but the restore does complete successfully, so this is mis-leading. > An example is "language <xxx> already exists" - pg_restore just throws a > warning to let you know this, but the restore is successful. > > Perhaps what might work better is a message box - if you've already > performed a restore but it returned non-zero, and OK is pressed again in > the same dialogue box, throw up a message saying something like "you've > already attempted a restore, are you sure you want to run it again?" Now thats a solution I'm happy with - I've added code to do this, if a restore (or backup) is attempted from the same dialog invocation, with the same filename as previously used. Thanks, Dave
That's great Dave, I'm sure it'll please many folks!! Thanks Andy. Dave Page wrote: > Andy Shellam wrote: > >> Yep I agree, with pg_restore, warnings also generate a non-zero return >> code, but the restore does complete successfully, so this is mis-leading. >> An example is "language <xxx> already exists" - pg_restore just throws a >> warning to let you know this, but the restore is successful. >> >> Perhaps what might work better is a message box - if you've already >> performed a restore but it returned non-zero, and OK is pressed again in >> the same dialogue box, throw up a message saying something like "you've >> already attempted a restore, are you sure you want to run it again?" >> > > Now thats a solution I'm happy with - I've added code to do this, if a > restore (or backup) is attempted from the same dialog invocation, with > the same filename as previously used. > > > Thanks, Dave > > !DSPAM:37,466eaba6343941575920893! > > >