Closed
Bug 99817
Opened 23 years ago
Closed 21 years ago
URL-based cache naming scheme
Categories
(Core :: Networking: Cache, enhancement, P5)
Tracking
()
RESOLVED
WONTFIX
Future
People
(Reporter: johnny.accot, Assigned: gordon)
Details
Since I have more disk space than I can actually use, I decided a couple weeks ago to set my disk cache size to ... 650MB. The idea is to save all pages that I'm loading and burn a CD when the cache quota is exceeded. In the past years I've often had that problem: I remember seeing an information in a web page, but I don't know where, I don't know how to find it again, and more importantly, when I find it in Google, the page is not here anymore! This is the motivation for saving all pages I see and archive them someday. Two problems appear when I want to do that: first my cache size seems to always stay under 25MB, not sure why or if it's a local problem; second it is really hard to browse though the archived data, and when the cache reaches 650MB, I'll have to recursively grep all files, which is awful. My idea was then that mozilla could (if I'm not the only one that needs that feature) let the user choose between the current naming scheme and a more human-readable one. For instance I have a couple GB of mirrors on my machine (laptop :-) ) and it is so easy to find information when I do a `wget -m <some url>'! So I thought I should suggest this scheme for the mozilla cache management: ~user/.mozilla/default/********.slt/Cache/www.mozilla.org/start/, etc... I'm not that much into development so I have no idea whether this would be really much slower than the current naming method. But for sure I'd rather have a slightly slower caching algorithm and have the archiving capability! Any idea?
Comment 1•23 years ago
|
||
Well-written. I couldn't find any dupes. This could be a big plus for evangelism. I have several friends who would switch from IE to Mozilla if it gained the ability to save everything you've ever seen on the web.
Comment 2•23 years ago
|
||
I think this would be a very good feature for Mozilla. It's just plain convenient if you want to mirror sites somewhere. Maybe it could be even easier code-wise, since if you save the filenames and change the modified date to the one reported by the server, you could do away with having an "index" file altogether. Offline browsing would also benefit.
Status: UNCONFIRMED → NEW
Ever confirmed: true
This sounds like the potential for some horrible bloat, IMHO. The cache code was already rewritten once. For browsing purposes, cache speed is the ONE main concern. For saving purposes, there are already suitable RFE's like for instance bug 11632 and bug 40873. Recommend wontfix.
Comment 4•23 years ago
|
||
However, bug 11632 (closest to what was suggested here) does not address the issue of automating the process for every page viewed. I would suggest keeping the primary browser cache how it currently is, but adding a preference to do this: Every time a file is removed from the cache, move the file to the user's web mirror folder instead of just deleting it.
Comment 5•23 years ago
|
||
Or if changing the cache to use this type of naming scheme, a built in function to "decode" all the cache files into a "domain" based system. I know that there are shell scripts to do this already, but having one built in would be great for the overall userbase
Comment 6•23 years ago
|
||
Some disk files may have more than one web file in it... (If not now, then ISTR that that will be the case soonish) As well, the cache has no concept of urls - it uses a "cache key", which is usually the url, but not always (eg forms POSTed to the same url). You could iterate through those, and you'd be right most of the time, I guess. Not everything is stored in the cache - nocache data, and so on.
Moving target milestone to FUTURE until we have a chance to triage this bug further.
Target Milestone: --- → Future
Comment 8•23 years ago
|
||
Excellent idea -- in fact I also thought of the same thing and I manually [UGH] implement it for ftp'd files. I think <mailto:d_yerrick@hotmail.com> has identified how to make it pallatable to folks who imagine the cache is somehow "secure." Let the user elect to save a set of pages using this name-replication, like certain other products have a "make available offline" function.
Summary: [RFE] URL-based cache naming scheme → URL-based cache naming scheme
This feature is not useful for most users and could incur a severe performance penalty. If such a feature WAS attempted, it would be better implemented at a semantic layer above the cache, perhaps http. Marking WONTFIX.
Status: ASSIGNED → RESOLVED
Closed: 21 years ago
Resolution: --- → WONTFIX
You need to log in
before you can comment on or make changes to this bug.
Description
•