126621 - Search within Cached History pages

Status: RESOLVED WONTFIX

(Core Graveyard :: History: Global, enhancement)

Description

[james wilkes]

Would it be possible to provide the ability to search within the _content_ of
cached pages that I've recently visited? I've just been trying to find a page I
looked at some point in the last 2 weeks and had to use grep in Cache directory.
It would be a most useful feature for finding things you forgot to bookmark.

The current history search is useful, but this would add a lot (eg. "where did I
see that code example of using JUnit?")

Comment 1

[Asa Dotzler [:asa]]

type about:cache into the addressbar. Select the cache you want to browse. Use
find in page to search. 

Comment 2

[Asa Dotzler [:asa]]

from reporter via email:

That's not quite what I meant. Using find in page on the about:cache pages would
only search in the urls stored. I meant searching in the _contents_ of the
chached pages, which can be done currently with grep (on *nix or Find on
windows). It's a common problem for me and I would think a lot of others. You've
read a load of sites and want to find a particular page again. The url's are
usually pretty meaningless (half the time you can't even remember the domain of
the site)

Seems like a pretty useful feature to me rather than feature creep.

I'm finding it happens to me less now that I'm getting more into browsing with
tabs and keeping a lot more pages open.

Anyway, I leave the decisions to you. Mozilla's great already. Keep it up.

Cheers,
j 

I don't believe that this is desired by "a lot of others". It seems to me to be
a niche feature that a very small minority of users will ever use. I'll let the
module owner decide whether or not this should be implemented. 

Comment 3

[Jesse Ruderman]

See also bug 114409, Search in the content of the bookmarked pages.

Comment 4

[Ken Takusagawa]

Omniweb http://www.omnigroup.com/applications/omniweb/ has this feature.

Comment 5

[Ken Takusagawa]

This bug is similar to bug 154114, in which the reporter (as an addendum)
asks for the ability to search permanent history.
This bug may be duplicated by bug 162543
-- I can't tell; its description's too vague.  It is related
to bug 81756 where I make a comment of how IE implements this feature.

And oh yeah. IE6 implements this feature.

Comment 6

[Andrew Hagen]

*** Bug 199487 has been marked as a duplicate of this bug. ***

Comment 7

[kevinar18]

I would like to see this feature in Firebird as well.  However, I'm not sure if
I need to file a separate report in Firebird for this feature.  Could someone
perhaps let me know if I need to?

Comment 8

[[e-mail]]

The solution proposed in bug 154115 looks cleaner IMO: the cache contains an
arbitrary subset of recent pages and searching it will be quite time-consuming.
The proposal in bug 154115 seems to be to just update a search index for each
page seen that allows the user to later retrieve the URL based on the text
content, even if the page itself is no longer in the cache. 

Comment 9

[[e-mail]]

Correction: bug 154114 as mentioned in comment #5 

Comment 10

[Jesse Ruderman]

*** Bug 230741 has been marked as a duplicate of this bug. ***

Comment 11

[u58943]

must say this is a feature ive greatly missed since trying to use firebird after
using IE.  used it quite often in IE to search for pages i know ive visited and
can only remember a bit of the content of those pages rather than the name or url.

extremely useful feature and to be quite honest thats what ive always expected
the "search" when using the history sidebar to actually do.  the way IE does it
is perfect imo and should be done the same in firebird.

of course with the firebird cache being trashed every 5 minutes despite the
browser not having crashed (i.e. simply shutting down windows without cloding
firebird first, and even sometimes just closing firebird is enough to wipe the
cache), theres not usually much in it anyway at the moment (it still knows the
urls but the cached content you want to search in is simply not there anymore,
and for that matter neither are the favicons but thats another matter entirely).

so i say implement it like IE has.  at least then if the trashing of the cache
ever gets sorted properly (and imo it should have been sorted before now, hell
even a 'pro' option to never trash the cache would be of great use to a lot of
us) we will be able to search the cached content as you would expect to be able to.

Comment 12

[Rishi Chopra]

Worth noting is that bug 23373 requests an extension for the browser history
feature and bug 208795 also comments on the current history search functionality.

Comment 13

[Rishi Chopra]

Sorry, that should have been bug 233773

Comment 14

[Rishi Chopra]

One problem with the suggestion of poster #8 ([e-mail]) is that
many pages nowadays are rendered dynamically; hence if the state of a server's
database changes, rendering and then subsequently searching will yield different
results.

If the goal is to keep a accurate record of one's footsteps accross the web,
you'd want to have the state of the page when it was viewed, not when it is
rendered.

Comment 15

[u58943]

That's besides the point, dynamically or not (not sure what u mean by 
dynamically to be honest, even something like a constantly changing forum has 
the same urls for each post), it still enables a user who remembers at least a 
few of the words from the site they visited, to find the page or at least the 
site it was on.

if IE can do this perfectly well, and very effectively, theres no reason 
firefox cant.  then again i said the same about the firefox/mozilla cache 
compared to IE and we all know how bad that is and will continue to be until 
its rewritten properly.

Comment 16

[Dac Chartrand]

Check out suggestion #1 in the following post:

http://slashdot.org/comments.pl?sid=102100&cid=8704328

And this follow up gives some hints on how a programmer would be able to
implement it in Mozilla:

http://slashdot.org/comments.pl?sid=102100&cid=8704772

Ideally, have mozilla record the referrer for every domain, and if it came from
a search engine it stores the query you sent. This would later enable you to
find that site by keyword in the history. Also, the ability to easily turn this
feature off and on to avoid future cache purges for pr0n.

Comment 17

[Rishi Chopra]

Dac's idea would be a moot point with code snapshots, timestamps, and complete
archive history (e.g. a cahce entry for every page rendered)...

Ewok seemed to miss my point; even if the URL is the same, the database that
processes the query embedded as part of the URL may have changed.  What I was
suggesting was code 'snapshots' of pages visited.

At any rate, I suppose it's up to Blake and the other developers to decide how
much of the needed functionality is included to the Mozilla engine, and how much
is pushed to the periphery to be enabled by plugin dvelopers.

Here are some more thoughts on history archiving:

http://forums.mozillazine.org/viewtopic.php?t=64908&highlight=

Comment 18

[eyal gruss]

i currently use the Cachesearch add-on for this functionality

Comment 19

[Marco Bonardo [:mak]]

Not going to happen any time soon. Activity Stream is already investigating indexing and it may appear there, it's unlikely to ever be part of plain history, so I don't think there's any value in keeping this open forever.