Closed Bug 23734 Opened 26 years ago Closed 26 years ago

Downloading files should not be added to the history.

Categories

(Core :: DOM: Navigation, defect, P3)

defect

Tracking

()

VERIFIED FIXED

People

(Reporter: bugzilla, Assigned: radha)

References

Details

(Whiteboard: [nsbeta2-][nsbeta3+][NEED INFO][PDTP3])

If you go to a site, select a file to download, go to another site and the presses "back" you are send to the file you just downloaded!
I wouldn't say it shouldn't be in the history _window_, but the back button is different of course. Perhaps downloads should spawn off and not actually have their own browser window.
Bill, take care of this post-Beta ...
Assignee: radha → law
Priority: P3 → P2
Target Milestone: M15
*** Bug 25608 has been marked as a duplicate of this bug. ***
Shouldn't this be fixed for beta?
I don't like this behavior too. I think that this should be fixed in beta.
OS: Windows 98 → All
Hardware: PC → All
*** Bug 30413 has been marked as a duplicate of this bug. ***
It should be in the history window.
*** Bug 30535 has been marked as a duplicate of this bug. ***
*** Bug 30833 has been marked as a duplicate of this bug. ***
Status: NEW → ASSIGNED
*** Bug 33358 has been marked as a duplicate of this bug. ***
*** Bug 34213 has been marked as a duplicate of this bug. ***
Move to M16 for now ...
Target Milestone: M15 → M16
Move to M20 for now. Should this be nominated for beta 2?
Target Milestone: M16 → M20
absolutely. this is _very_ bad/unexpected behavior... ..while "move back through all visited objects in the history" might make sense from a program-level point of view, when i click on "back" i mean "back to the last page i viewed", not "back to the gzip file i downloaded".
I'm going to try to pass this one off to Radha as it requires more knowledge of session history than of download. The code that handle unknown content is nsUnknownContentTypeHandler::HandleUnknownContentType. Either we need to fix the code to call this so it doesn't add the URL to session history, or, this code needs to remove it (at least that's my guess). This might well get fixed by SH changes that are pending (but Radha would know that, too).
Assignee: law → radha
Status: ASSIGNED → NEW
there's probably a dupe floating out there but in the meantime, I will nominate this bug for nsbeta2. As mentioned before, it makes for a horrible user experience to click back and amazingly start your 10Mb download over your slow connection again.
Keywords: nsbeta2
Whiteboard: [need info]
Claudius, I remember going through this question before in PDT, so I think this is a dupe, too. Could you find the dupe and let us know in this bug what its status is? Thanks.
I searched around a bit and found several dupes. Most of them led back to bug 29975 (M20, nsbeta2-'d on 5/25), none of the others were presented for PDT consideration. The statement of the problem there however didn't quite capture the full scope of the problem like this bug does however. radha and I are assigned to both so I'm actually going to dupe the other and let this stand. I vote for (+) btw.
*** Bug 29975 has been marked as a duplicate of this bug. ***
Declining beta2 nomination just like 29975
Whiteboard: [need info] → [nsbeta2-]
nominating for beta3
Status: NEW → ASSIGNED
Keywords: nsbeta3
Nav triage team: [nsbeta3+]
Keywords: relnote2
Whiteboard: [nsbeta2-] → [nsbeta2-][nsbeta3+]
*** Bug 33083 has been marked as a duplicate of this bug. ***
*** Bug 47028 has been marked as a duplicate of this bug. ***
*** Bug 47204 has been marked as a duplicate of this bug. ***
Priority: P2 → P1
*** Bug 48561 has been marked as a duplicate of this bug. ***
adding mostfreq kw as this has 11 dupes not counting the dupes of the dupes.
Keywords: mostfreq
i tried this today. What I see is, 0) visit ftp://sweetlou/products/client/6.0x/windows/32bit/x86/2000-08-16-10-M18/. This is added to SH. 1) Click on the link N6setup.exe. The url to load as per docshell and SH is, ftp://sweetlou/products/client/6.0x/windows/32bit/x86/2000-08-16-10-M18/N6Setup. exe. This gets added to SH. Then, the filedownload dialog appears and I save the file. 2) After the file is saved, I click back, I go to ftp://sweetlou/products/client/6.0x/windows/32bit/x86/2000-08-16-10-M18/ (the directory listing). The file download dialog doesn't appear now. 3) If I click forward, the file download dialog appears inresponse to loading the url, ftp://sweetlou/products/client/6.0x/windows/32bit/x86/2000-08-16-10-M18/N6setup. exe I think this behavior is OK. When you click back and forward, the download doesn't happen automatically with out the user's approval. Am I missing something here? Is there a better url where I can see the *real* problem?
steps 2 and 3 are what is different from 4.x. In 4.x SH completely ignores the file download and you don't run into any problems. In step 2 w/4.x the back button isn't even active. In step 3 the forward button isn't active. I think adding the download to SH just leaves you exposed for possible problems w/ no true benefit. I'll look around for a good testcase that easily demonstrates the downfalls.
Based on the comments by me and claudius, I think this isn't a horrible bug that it is being projected as. I will wait for another week for QA to come up with an example that will show how bad this is. Then work on fixing it.
Whiteboard: [nsbeta2-][nsbeta3+] → [nsbeta2-][nsbeta3+][NEED INFO]
I don't think the problem is in the re-appearance of the download window, but that the wrong window still shows up: the 'unknown content' window, even when you just downloaded a file with certainly *known* content (like .wav). Though, in my opinion, you seldom download and save the same file twice the same minute, so what's the point of putting these download events under the back button? I think it can be categorized as not horrible though quite annoying... Another little bug related to this is the behaviour of the mouse pointer and this nice spiral-like thing in my status bar (modern chrome). They keep telling me Mozilla is working, while I cancelled the re-download of the back-button-file. Pressing the stop button stops this.
PDT thinks this is a P3. Setting that priority and leaving [PDTP3] breadcrumb in status whiteboard.
Priority: P1 → P3
Whiteboard: [nsbeta2-][nsbeta3+][NEED INFO] → [nsbeta2-][nsbeta3+][NEED INFO][PDTP3]
Just adding some text why I reported this bug: 1) I when to http://ftp.mozilla.org/pub/mozilla/nightly/latest/ 2) clicked on the "mozilla-win32-installer.exe" link and downloaded the file. 3) Now I manually entered www.mozilla.org and went there. 4) After this I pressed Back and was presented with the download dialog again. This is not the case with both NS and IE.
Here's a decent test case to demonstrate the annoyingness of this bug where it occurs: Go to: http://www.freeciv.org/download.phtml Try to download, for example, the freeciv source code (under "Stable version"). You should get a dialog that says "This file has mime type application/octet- stream and cannot be viewed using Mozilla." You can click cancel to stop the download. Then click the link next to "Current Stable Version" that says "News!". YOu should see a text document load. Now click back, and it will pop up, interestingly enough, a *different* dialog that says "You are about to download a file of type application/octet-stream." (These two dialogs do the same thing, why have both of them? If it matters, I like the first one better.) Click cancel. That marks the download as "rendered" and so if you click back again, you'll get to the freeciv download page. I think, again, that it should be retained in the history window but definitely NOT the back button. One rarely needs to download the same file twice. Using windows build 2000083111.
Related bug: the URL in the address box should not change to the URL of the downloaded file. (I think this is really part of the same bug.)
I have a fix for this bug that's part of another bug I was working on. Pending approval and review from radha, I'll check it in.
I just checked in a fix for this.
Status: ASSIGNED → RESOLVED
Closed: 26 years ago
Resolution: --- → FIXED
this is fixed in the 200090608 builds. Marking VERIFIED
Status: RESOLVED → VERIFIED
Component: History: Session → Document Navigation
QA Contact: claudius → docshell
You need to log in before you can comment on or make changes to this bug.