Difference between revisions of "Firefox Save Image As Save Per Site Annoyance"
(→Further Testing) |
(→sessionstore.js) |
||
(4 intermediate revisions by one user not shown) | |||
Line 80: | Line 80: | ||
related terms: browser.download.lastDir.savePerSite image save as | related terms: browser.download.lastDir.savePerSite image save as | ||
+ | |||
+ | === sessionstore.js === | ||
+ | |||
+ | ''Note that browser.download.lastDir.savePerSite is set to false in Firefox version 15.0 for this test.'' | ||
+ | |||
+ | In a test, after several image saves from different sites, a subsiquent image save suddenly requested a path different from the previous saves. The site was 'www.starpulse.com' which seemed to trigger the Firefox site memory bug. Using the windows XP search tool, searching for files containing the text string 'starpulse' resulted in a match on sessionstore.js file. Windows XP path: | ||
+ | |||
+ | C:\Documents and Settings\<username>\Application Data\Mozilla\Firefox\Profiles\uoi53d54.default\sessionstore.js | ||
+ | |||
+ | This is a text readable file you can open in windows notepad. Within is a rich history tracking your online activity. The search engine query string that was clicked with lead to the search result on starpulse.com and additional data are all present in the file. What's NOT in the file is the local path to which was 'remembered' by Firefox. | ||
+ | |||
+ | Conclusion: Although this file contains suspiciously relevant data, it doesn't seem to relate to the file save path. | ||
+ | |||
+ | === untested hypothesis: retention prior to parameter change === | ||
+ | |||
+ | Since it seems that only certain sites are impacted by the glitch, and more are not than are, it is possible that prior to the parameter browser.download.lastDir.savePerSite being set to false, the site to path values that were set are being retained after the value of the parameter has been set to false. | ||
+ | |||
+ | On a fresh installation of Windows, with a fresh installation of Firefox, and having browser.download.lastDir.savePerSite set to false immediately prior to any files or images have been downloaded, perhaps the glitch would never manifest. | ||
+ | |||
+ | If it could be known where these associations are stored, that file could be erased, or the data cleared. It is known that it has nothing to do with clearing the download history nor does it have to do with clearing the browser cache. These associations seem to be kept 'somewhere' and are not plain text. | ||
== References and Resources == | == References and Resources == | ||
related: | related: | ||
− | * [ | + | * [http://support.mozilla.org/en-US/questions/887248?page=2 Mentions clearing Site Preferences] |
− | * [ | + | * [http://tifflife.wordpress.com/tag/browser-download-lastdir-savepersite/ Explains the Bug Very Well and Includes Details on How the URI Data is Saved] |
Latest revision as of 19:41, 9 September 2012
Contents
Description of Issue
In Firefox 7 and newer versions, the save path of files and images that you download is "remembered" based on the web site address. Firefox remembers the last saved download directory specifically for that site, so that when you return to that site, Firefox will associate it with the remembered download path for files and images.
So, if you save five images in a row, from five different sites, each of which are sites you have downloaded images from in the past, for each image Firefox will attempt to download to a different path. Confused? You should be, and most find this idiotic feature both confusing and extremely annoying.
Firefox basically remembers the save location on a per-site basis.
The best explanation for this feature comes from a blog called Tiff Life, quoted below:
When you have one link like “i32.site.com” then another on “m22.site.com” that these pictures are being fetched from, guess what? The next time you get a picture from either of those servers, it’s going to remember the last directory you saved to for that server. Let’s say we’re browsing something about kittens, we save a bunch of pictures. Then later, maybe we’re looking at pictures of rabbits. We have our folders very organized like “Pictures/Animals/Cats/Kittens”, “Pictures/Animals/Rabbits.” So we start saving the rabbit pictures and then suddenly we realize that some of them (not even all of them, just some) are going to the Kittens folder. What’s worse is that some are going to maybe Giraffes, or Whales, Sharks, Lions, or who knows what!
citation: credit to TiffvZ for the example above!
Most people organize files by other means than what site those files were downloaded from. That is why this new feature seems so utterly counter productive. Moreover, it was added without much announcement, and no fix was available until several releases later.
Moreover, the issue is so difficult to articulate that it took time for enough angry users to report it using similar enough descriptions to eventually clear through the fog of the problems it was causing leading to the root culprit.
Official Fix
A "hidden" configuration setting was added to Firefox 11 to address this feature / bug.
This fix from the developers of Firefox requires the user to have some technical competency since it is not just an option under one of the regular menus. The user must edit the Firefox configuration parameters in about:config .
1. In the browser URL bar type the following:
about:config
2. If prompted with the warning about voiding the warranty, choose to continue
3. Right-click in the listing and select New > Boolean from the context menu and enter the following:
browser.download.lastDir.savePerSite
4. Set the value for the new preference key to:
false
The preference key is CASE SENSITIVE: browser.download.lastDir.savePerSite with capital letters D, P, S.
Note that browser.download.lastDir.savePerSite controls whether the directory preselected in the file picker for saving a file download is being remembered on a per-website (host) base. If set to true, the data is stored as content preference.
- reference: browser.download.lastDir.savePerSite
values:
- true (default) - The last used directory for the website (host) serving the file for download will be preselected in the file picker. If no download directory for the current website has been stored, browser.download.lastDir will be used.
- false - The last used directory for any download (stored in browser.download.lastDir) will be the preselected directory in the file picker.
This has been discussed on Firefox Help on the thread, Save image doesn't remember last location saved for Firefox 7, page 3 mentions browser.download.lastDir.savePerSite.
Implementation Background
This behavior was introduced in FF7, associating a downloaded image with a folder on your harddrive.
FF7 Workaround: Apply the trick once a while when you want FF to use the general/last download directory and forget about all these useless directories it memorized. It is easy to apply : Select Tools | Option and choose Privacy tab, Click on "clear your recent history", Time range to clear : choose Everything (last option in the listbox), Click on Details (if necessary), **Uncheck** all options **except** Site Preferences (thus only Site Preferences is checked at the end).
result: inconclusive
source: Save Image As keeps changing directories
FF8 Workaround: Apparently the issue was addressed by the addition of an option to disable this behavior.
* Tools, Options, and Privacy Tab * uncheck "remember download history"
You can also uncheck "remember my browsing history" for privacy purposes.
result: inconclusive
Further Testing
Testing after applying the browser.download.lastDir.savePerSite fix indicates that the save path memory behavior still exists for certain web sites.
It seems the issue is still not resolved. this bug is persistent. with value set to false, ff version 14.0.1, sometimes the save directory changes (is not the previous one used) and seems related to the TLD URL + the host, particularly noticeable with blogspot.com image saves.
- www.domainexample.com will save differently than www.anothersite.com because the domain is relevant.
- a.domainexample.com will save differently than b.domainexample.com because the host is relevant.
- a.domainexample.com/cats will save in the same path as a.domainexample.com/dogs because site paths right of the slash are irrelevant.
update: testing concludes this glitch relates directly to "Site Preferences" browsing history. In an experiment, going to "Clear Recent History" and selecting only "Browsing and Download History" causes next save on stick site to be the default /download directory, which at first seemed related, until further testing revealed that only clearing "Site Preferences" had a specific and direct result; this will cause the "save as" to download to the last save as directory, even for TLD domains with the memory effect. If the user manually cleared only the "Site Preferences" each time, then the image save directory for specific sites glitch will not manifest. (In the experiment the browser was closed and opened again each time after clearing.)
browser.download.lastDir.savePerSite and set the value to false DOES NOT RESOLVE THIS BUG! It might for some installations, but does not for our lab computers. Tested, glitch remains unresolved. ATTENTION FIREFOX DEVELOPERS! THIS NEEDS FIXED!
savePerSite URI is stored in a file named content-prefs.sqlite
related terms: browser.download.lastDir.savePerSite image save as
sessionstore.js
Note that browser.download.lastDir.savePerSite is set to false in Firefox version 15.0 for this test.
In a test, after several image saves from different sites, a subsiquent image save suddenly requested a path different from the previous saves. The site was 'www.starpulse.com' which seemed to trigger the Firefox site memory bug. Using the windows XP search tool, searching for files containing the text string 'starpulse' resulted in a match on sessionstore.js file. Windows XP path:
C:\Documents and Settings\<username>\Application Data\Mozilla\Firefox\Profiles\uoi53d54.default\sessionstore.js
This is a text readable file you can open in windows notepad. Within is a rich history tracking your online activity. The search engine query string that was clicked with lead to the search result on starpulse.com and additional data are all present in the file. What's NOT in the file is the local path to which was 'remembered' by Firefox.
Conclusion: Although this file contains suspiciously relevant data, it doesn't seem to relate to the file save path.
untested hypothesis: retention prior to parameter change
Since it seems that only certain sites are impacted by the glitch, and more are not than are, it is possible that prior to the parameter browser.download.lastDir.savePerSite being set to false, the site to path values that were set are being retained after the value of the parameter has been set to false.
On a fresh installation of Windows, with a fresh installation of Firefox, and having browser.download.lastDir.savePerSite set to false immediately prior to any files or images have been downloaded, perhaps the glitch would never manifest.
If it could be known where these associations are stored, that file could be erased, or the data cleared. It is known that it has nothing to do with clearing the download history nor does it have to do with clearing the browser cache. These associations seem to be kept 'somewhere' and are not plain text.
References and Resources
related:
- Mentions clearing Site Preferences
- Explains the Bug Very Well and Includes Details on How the URI Data is Saved