"Document Expired" Internal Error Page

[niteshade]

Is there a way to disable Palemoon's "Document Expired" internal error page in this situation? Steps to reproduce:

1) Navigate to http://www.trackinfo.com/dog-search.jsp
2) Input into form and submit: Fast Sally
3) Click race result link: 2012-03-27 N9
4) Click browser Back button

I get a Palemoon internal error page with a "try again" button. Clicking that button produces a confirmation popup to resend the data. Palemoon versions prior to 11 didn't generate the internal error page, just the confirmation popup. I'd like to avoid the extra click if possible. Since there are several SSL elements on the search page I made the browser.sessionstore.privacy_level config change mentioned here: http://forum.palemoon.org/viewtopic.php ... 2770#p2770 and restarted Palemoon with no effect.

Thanks.

[satrow]

Same thing happens with Firefox 11.

[megaman]

I came across the same situation (FX 14 and PM 11), but I get sent back to the "Coming Soon" page and never saw the confirmation pop-up.

[Moonchild]

A little related snipped on Mozillazine deals with this:

Agreed that this new feature is annoying, considering you have to hit two retry buttons to resubmit.

For POST results that don't need to be refreshed, you could ask the web developer to add a

Code: Select all

Cache-Control: max-age=600

header to the document. This allows the browser to cache the results of the POST form (for up to 10 minutes - hence the 600 seconds value), which means you can just hit Back and see the same results again (no resubmit).

There are more things that can be done server-side like cache-control:private etc. Since the data entered (a search term) is not sensitive, the people running the site should either:

• Add a cache-control statement to their server setup, OR
• Provide a "back" button on the "coming soon" page or another way to navigate back to the search results, to not break navigation of the site

The reason for this "double click" is bug #451250 that got implemented in the version 10 code base. You are dealing with two different warnings from two different routines here. The first (page has expired) results from the server explicitly telling the browser not to cache data. The second (resubmit) results from the fact that a previously called-up page required submission of form data. The two don't necessarily have to happen at the same time, I think.

The reason this information screen was implemented is because when paging back, the old method would not allow you to page back quickly past the "results" page to the "form" page on sites, since it would give you an error pop-up and people would have to click "cancel" on the popup. So there are valid reasons to go either way with this, but it's a borderline case either way.

It should be fairly easy to revert for me, BUT I have to make sure it's the right thing to do. How many people actually run into this problem? Is it really a browser problem or rather a misconfigured server or poor page design causing this inconvenience? Is having the browser cater to server misconfigurations a good thing to do?...

[niteshade]

It should be fairly easy to revert for me, BUT I have to make sure it's the right thing to do. How many people actually run into this problem? Is it really a browser problem or rather a misconfigured server or poor page design causing this inconvenience? Is having the browser cater to server misconfigurations a good thing to do?...

Thanks for the analysis and explanation. I'd vote for an about:config pref to work around the issue if that's possible. I think that would give control back to the user and not leave him at the mercy of every site admin on the WWW. I don't know that there's a perfect solution but from my experience the way the pre-10.x code handled it provided a smoother browsing experience than the 2-click solution we have today.

[Moonchild]

I've actually given this some more thought, and you are correct. The "document expired" page would almost always be tied to data submitted in a form, so there really shouldn't be a reason apart from people wanting to page back twice instead of resubmitting the form data (which would happen more often) for this page to be displayed in case of a "navigation dead end" on a website. Either way it's very much a convenience problem. Unfortunately catering to this kind of "optional" issue for borderline convenience problems would stack up real quick since Mozilla forces Pale Moon development to start over with every "major" release and reimplementing all the changes that are required to transform Firefox into Pale Moon. The longer the list, the more time and effort it will take for Pale Moon to be released.

I'll think about it. That's all I can promise.

[niteshade]

I'll think about it. That's all I can promise.

That's all I can ask.