Rebecca Menessec

What it would do is allow Windows 8 refuseniks (me) to TRIM a previously-used drive that's just had a fresh Windows install put on it. I can't find any references to Windows 7 helpfully TRIMming the content of a previously-used drive before installing to it, so presumably all the blocks that were in use before the new Windows install are still marked as "used" by the SSD. Yes, technically, I can boot from removable media with Linux and attempt to TRIM the drive wholesale, but that's a major headache for a variety of reasons, not the least of which is that many BIOSes don't really quite correctly support booting from USB flash drives; and the ones that do tend to be very quirky about just which drives they'll boot off. ...or I guess I can wait and see if Windows 8.2 is any better.

NEW ENTRY [Google Update*] LangSecRef=3024 DetectFile=%LocalAppData%\Google\Update\GoogleUpdate.exe Default=False FileKey1=%LocalAppData%\Google\Update\Download|*.*|RECURSE FileKey1=%LocalAppData%\Google\Update\Offline|*.*|RECURSE Warning=Not while Google Update is updating things. --- Not sure if this can be / needs to be expanded to detect Google Update installed in %PF%. Not clear how to install Google Update nor Chrome SxS--what I use--for all users to begin with. In fact, I wouldn't mind a pointer, if you can give one... ?

I haven't gotten any response at all from Piriform employees. None, zip, zilch. I'm trying to attract at least one of them, and get at least an acknowledgment that the next release will have fixes, or else get a request for more information to help them find the problem. CCleaner lists Firefox whether or not I use the CUSTOM directive, as I explained. Chromium is never displayed. CCleaner does not compact Firefox databases even though it's detected one of my profiles. Sorry, it's not one of the easy ones. I've been using, hacking up, or administrating Windows since 3.0. I use Process Hacker instead of Task Manager, by the way. It's a nice little replacement for SysInternals Process Explorer, and it's available in native 64-bit, unlike the SysInternals utils. Recommended. Not this, either. When CCleaner did compact Firefox databases, it compacted on every run, or else warned me there wouldn't be any compaction because Firefox was running. Now there's no compaction, ever. Sorry, I assumed forum moderators would be Piriform employees. I apologise for my own assumption.

Oh, by the way: this was broken in 2.34, too. It was broken in 2.35, which I merely pointed to for the Chromium announcement. 2.36? Yes, also broken. And? I can compact both Firefox and Chromium SQLite databases just fine with: SQLite. Yes, Firefox 3.x (and 4.0pre, which I've never even glanced at) all use SQLite 3.x databases. They haven't changed at all except in the details of how many databases and what those databases store. So now I'm using PowerShell, which I loathe, to pass out filenames to a Win64 build of SQLite 3.20, which can issue a VACUUM just fine. It's funny: I use Firefox *and* Chromium on Win32, Win64, and amd64 Linux (Gentoo and Ubuntu). SQLite compacts databases just fine in all four cases. Given the evidence that Firefox and Chromium are not in some way frakking around with their usage of SQLite, the evidence points to: CCleaner! Which, out of the two ways I know of to compact Firefox or Chromium databases, is the only thing consistently broken. Are you seeing the pattern, or is it just me?

Can the butchery of Shakespeare, please. What out of date application? See my previous response. And don't assume. Sorry, what? What's next year's application? Are you referring to Firefox? I said it doesn't compact database for ANY build of Firefox, including full release builds. Were you assuming I meant Firefox 4? I'm using Namoroka nightlies. Not Minefield. Or did you mean Chromium? There's precious little difference between a "stable" Chrome release, a "development" release, or the hourly-ish buildbot releases. And I said it hasn't been working for any version of Chromium I've tried. When I say "Chromium 6, 7, 8", I mean "any half-assed release of a browser from Google, whether it has a shorter name or no." Will the next contestant please read, then respond? In that order? Try to know something about web browsers, too. It's been a long-ass week already.

Of course. I linked to the announcement; not the version I have installed. My previous attempts to point out this problem ended up with people calling the syntax into question, so I linked to the only place Piriform appears to document it: one announcement. I was purposely not leaking information about my system(s) or account(s). Yes, mine looks similar.

I've posted several times without getting answers, so I'm hoping this title is more eye-catching Custom locations for Firefox don't work Custom locations for Chrome don't work Database compression for Firefox is broken Are there any plans to make this functionality work? Ever? ccleaner.ini looks like this, leaving out unnecessary entries: [Options] (App)Mozilla - Compact Databases=True CustomLocation1=FIREFOX|C:\Users\user\AppData\Roaming\Mozilla\Firefox\Profiles\profile.default CustomLocation2=CHROME|C:\Users\user\Applications\chromium\Profiles\profile\Default Ever since I started using Firefox nightlies (Namoroka), database compaction broke. Uninstalling the nightly build and reinstalling vanilla Fox made no difference. No prompts, no cleaning, nothing. CCleaner finds the Fox profile with or without the CustomLocation entry, but never performs compaction. Chromium profile cleaning's never worked at all, ever. Chromium 6, 7, and now 8: total loss. Yes, I have the correct syntax: http://www.piriform....5/ccleaner-v235 Is CCleaner defunct? Is bugfixing on hold until this mysterious 3.0 build is released? I have yet to get a reply from an actual Piriform employee that I know of. Do they even watch these forums?

Right, but every existing entry in ccleaner.ini uses = for assignment. Tried it without anyhow. No change. (The blog post about the new custom Chrome support used = in the examples.)

So, ccleaner.ini looks like-- well, a lot of things --but especially this: CustomLocation1=CHROME|C:\Users\user\AppData\Local\Chromium\User Data\Default CustomLocation2=FIREFOX|C:\Users\user\AppData\Roaming\Mozilla\Firefox\Profiles\5gt2o399.default I'm using the official Firefox nightlies and official Chromium hourlies, and that appears to have confused CCleaner. Ever since going nightly, CCleaner's refused to compact my Fox SQLite3 databases (but it performs other profile cleaning), so I finally added the Firefox entry. --still no compaction. I should note that Namoroka (the Fox nightly) is installed in %ProgramFiles%\Mozilla Firefox and is still called firefox.exe. This seems like some sort of bug, possibly. Now that there is theoretically support for custom Chrome locations, I said "Whoopee!" and added one. Chrome/Chromium/whatever doesn't appear anywhere in the sea o' checkboxes. I'd rather like compaction for the Chromium SQLite3 dbs, not to mention cache cleaning. I tried reversing the order-- still nothing. Have I missed something crucial in the syntax?

Right, I know about using the native Windows APIs. Problem is, I don't know much about those APIs.

Hm. No bad frag calculations here. (So far.)

Sorry, not what I was after. Windows since 95 (like DOS 6.0-6.22 before it) can transparently compress files without the aid of an archiving or compression utility. This is completely separate from recent Windows' support for Zip archives (located in ZIPFLDR.DLL), and also has nothing to do with the ancient Microsoft COMPRESS.EXE / EXPAND.EXE utilities, nor MAKECAB.EXE or anything else related to Cabinet archives. Assuming a system with compression available (you can check with "fsutil behavior query disablecompression" from at least Windows XP), you can use either COMPACT.EXE to set the necessary flag on one or more files or directories to begin compression, or the GUI: pull up the context menu for most any file / directory object in Explorer, select Properties -> Advanced -> Compress contents to save disk space. Compression generally takes at least a few seconds (longer for larger files), but at some point pulling up Properties again for the target will reveal a (hopefully useful) difference between "Size:" and "Size on disk:". Best of my knowledge, this still employs LZX compression, like nearly anything else Microsoft. More recent implementations of LZX actually yield surprisingly good compression (Windows NT6.x (>= Vista) .WIM disk archives), but NT6.1's implementation looks like it's still as terrible as Win95's or DOS': getting more than 2:1 compression on easily-squashed things like plaintext will have to wait for Microsoft to care about NTFS again (it's barely changed since NT 4.0) when (for instance) recent implementations of LZMA/LZMA2 can yield 15:1, 30:1, and sometimes even more. I'm not sure who to ask, so if you can find out about NTFS-compressed files, I'd appreciate it.

Oooohh, right. Silly question. I thought of a better one, though: does Defraggler handle NTFS compressed files in any way? I've not-installed, deleted and/or set compression for things like those pesky F/OSS apps that come with huge translation sets, Notepad++, and so on, plus just about everything on C: that ends in *.txt; *.htm[l], *.log; *.ini; *.edb; and so forth. Disk I/O being the bottleneck it is, and laptop drives being as (relatively) small as they are, it's a savings in both bytes and wall-clock-time (and even power: I'm saving the planet while increasing my battery life!) to squish highly-compressible, infrequently-modified files like this.[1] I used to know NTFS architecture better, but a vague memory is telling me that the contents of compressed files end up in the MFT (for no obvious reason...), where presumably Defraggler Dare Not Tread. --- [1] I have this great suggestion for CCleaner...

I love having a command-line version of Defraggler. Thanks! I'm firing it from Task Scheduler every so often, under Windows 7, targeted at C:, and forcing it to Background priority with PsExec (SysInternals PsTools). First, an important question: I'm wary, so I've got Task Scheduler set to kill the task (df64.exe) if it doesn't end after 4 hours (plenty of time). What will actually happen if the scheduler's forced to kill it? Will it flush current I/O and go away gracefully? I like my filesystems not-trashed. Suggestions: Consider a minimal-GUI mode for defraggler[64].exe: I'm allowing the task to run visible, whereupon TrayIt! is minimizing the window to systray. That way, I can glance at it for progress / errors if I have to. It's not very revealing, though; it would be nice to be able to launch the GUI app with commandline parameters (including /closeonfinish or similar). Alternately, consider a "verbose" mode for df[64].exe: Output % complete in one line only (instead of scrolling), with another line for the current file and a third line to report any errors. Possibly even a minimalist sort of textmode scrollbox for errors. PsExec works well enough for forcing the task priority low, but it would be nice if df[64].exe could handle this itself. Less complexity; no outside dependencies. Handling its own systray placement would be awesome, although I assume TrayIt! will continue to handle this semi-gracefully. Please add a single icon resource (the GUI app's would be fine) so I can pick it out in the systray. Not clear on whether df[64].exe would need to find its associated console window and change the icon-- been a while since I coded under Windows. Thanks for such a superb product!

I actually have a bash alias to do this on my Linux machines, both for Fox and Chromium. I'd prefer not to have to do it manually for Chromium under Windows, is why I was asking. Thanks, though. (Technically, this alias also works fine with Cygwin so long as the sqlite3 package is installed. I'd just prefer all of my crap cleaning in one place.)