Firefox Is A Memory Hog With Memory Freeing Issues

The “garbage collection” in all sorts of programs has gotten much worse over the years, in direct proportion to the increase in memory sizes in computers.

Programs, like browsers, that used to run in 64 MB machines Just Fine now get cranky with less than 1000 MB ( one GB ).

IMO, that’s just sloppy lazy programming.

Generally I’ve been able to avoid the issue. Often by adding large disk based swap areas on systems with smaller memory or adding real swap disk to systems running from an SD card. Yet there’s one place where it’s been a Royal PITA. The Mac on micro-SD card.

I’d been running on a 16 GB chip and the fat Macintosh software filled most of it, leaving just about 2 GB free for swap. I’d get slow browser performance, then it would eventually pop up a warning that I was running out of disk… So I’d quit and start over. At first, I’d do a general reboot. Then I found just exit and re-launch of FireFox would “fix it”.

Now partly this is due to Apple having a slightly bizarre way of doing swap. It goes to a dedicated name space in the file system, and incrementally adds chunks, but leaves them laying around if your system crashes. So one “bit of art” you get to learn is how to scrub that space sometimes. The other is that, expecting a SSD Solid State Disk, it is not bashful At All about writing large chunks. After all, speed isn’t an issue anymore, right?…. Except it is always an issue to waste speed or memory or reads or writes or anything else, really.

So a few days ago I figured I’d migrate to a larger mini-SD chip. That was it’s own adventure. Seems the Mac wants a GUID boot sector so you get to reformat the card and the button to make it GUID is hidden under some “advanced” or similar heading… but if you just do a dd copy onto the new chip, it works, but is only 16 GB with the rest of the space unreachable. Disk Utility doesn’t understand how to expand the partition and on Linux there may be a way, but Apple File Systems are an add on (that I’ve done) and gpartd was not really happy with it (for reasons I didn’t care to explore / fix). So the REAL magic sauce is that you can use Disk Utility to clone one chip onto the other, but doing it directly chip to chip failed with a “validation error”… perhaps related to speed / pauses in read / write? In any case, doing a “make image to disk” (AFTER booting from a hard disk… so both chips become non-system…) then “restoring” that image to the new chip worked.

Yay! I thought… My “out of swap” issues and slowing down will be over!

Nope.

Things did improve. I no longer run out of swap (having about 15 GB available for 2 GB of memory); but the incremental sloth remained.

As videos are watched, or pages opened and closed, or just general use of FireFox happens, the swap space used grows exponentially. I don’t know if this is a FireFox issue with a memory leak or “whatever” or (more likely IMHO) a Macintosh Bright Idea of increasing swap in increasing chunk size. Where before I hit “slow and out of memory warnings coming soon” I now just get ever increasing sloth to the point of unusable wait times (that I often wait through to hit “send” and then exit FireFox / re-enter…)

So what’s the deal?

Here’s the swap space:

When first launched and FireFox open:

EMs-MacBook-Air:~ chiefio$ ls -klh /private/var/vm
total 2162688
-rw——T 1 root wheel 2.0G Mar 26 09:11 sleepimage
-rw——- 1 root wheel 64M May 12 09:04 swapfile0

After some use and sloth has become large:

EMs-MacBook-Air:~ chiefio$ ls -klh /private/var/vm
swapfiles
total 4194304
-rw——T 1 root wheel 2.0G Mar 26 09:11 sleepimage
-rw——- 1 root wheel 64M May 12 09:04 swapfile0
-rw——- 1 root wheel 64M May 12 11:03 swapfile1
-rw——- 1 root wheel 128M May 12 11:03 swapfile2
-rw——- 1 root wheel 256M May 12 11:03 swapfile3
-rw——- 1 root wheel 512M May 12 11:03 swapfile4
-rw——- 1 root wheel 1.0G May 12 11:03 swapfile5

“sleepimage” is just what it sounds like. The Mac writes it so that when it “sleeps” it wakes up at the same place. At sleepytime, it updates / rewrites that 2 GB memory image (larger in larger memory machines)

Next, it starts adding “swapfiles”. First a 64, then another 64, then 128, then 256, then 512, then 1 GB, then my patience runs out before it can do a 2 GB, or it only does a total of 2 GB to match memory size – I’ve not bothered finding out which…

Note that there are 2 x 64 MB, but one is an older date stamp. Sometimes you will find multiples of these files with older dates where it didn’t get reused, just ignored, from some prior shutdown / crash. Sometimes this accumulates to larger numbers and sizes. Bad Apple! Rebooting usually clears these, but sometimes not for the bogus ones.

“Someday” I may get it loaded up to this point and just take lunch, see if it’s writing another 2 GB or not. But not soon.

Now really, just why is FireFox packratting 2 GB total of crap? Closing windows & tabs is not releasing the memory. It just grows. Bad memory management, IMHO. Always malloc-ing and never freeing… as a guess.

In Conclusion

So if you run FireFox and are in the habit of leaving it up and open for long times, and find yourself getting lots of swap activity or just having a slow response, try just quitting and restarting it. It may be that this only really “bites” on an SD card where writes are slow and expensive. BUT the behaviour driving it ought to be the same on other systems… modulo the exponential swap auto-allocation…

Subscribe to feed

Advertisements

About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in Tech Bits and tagged , , , . Bookmark the permalink.

33 Responses to Firefox Is A Memory Hog With Memory Freeing Issues

  1. Larry Ledwick says:

    I have been using the periodic close method for over a year due to those very issues. If you leave firefox open continuously on a system that does not get shut down often it slowly eats memory.

    Sometimes it would get so bad I had to use task manager to kill it as it would take forever to die if you just clicked the x box in the corner of the session.

    My current firefox session has 5 instances running using a total of 540476 K of memory, and that is with only one tab open.

    Brave is much much worse (chrome variant) it has 13 instances running on 3 tabs and over 800 meg in use.

    It helps if you set it to clear everything when you close the browser, so I run different things in different browser sessions and just periodically close and re-open the browsers. I no longer have the browsers lock up the system.

  2. E.M.Smith says:

    At first launch, Opera is using more memory than FireFox (why I first swapped to FF), but I’ve not tested it longer term… Maybe a nice project for today…

  3. E.M.Smith says:

    OK, I’m running from Opera now. We’ll see how memory usage grows with some video watching (I’m going to open the boat thread and watch a few. That usually causes F.F. to load up memory pretty quickly.)

    I’m also going to clear some notes while I’m here. So here’s the link to the page where I found how to clone your Mac Disk, except it didn’t work on the mini-SD card (that validation error) but it gets the idea across. So I had to boot off a hard disk image (that I already had on external USB, just don’t use it when not at a desk…it is a laptop after all… and a dangling USB hard disk in your lap is not a good thing…) and then just copy the SD card to a disk image, then restore the disk image to the new chip, AFTER formatting it as GUID…

    http://nyacomputing.com/how-to-create-a-bootable-clone-of-your-mac-hard-drive/

    (And I’m reminded why I’m not fond of Opera on the Mac… as when you go to paste in a URL it tries to concatenate the text with the last line of text despite blank lines and white space…)

    General directions on how to change startup disk in a Mac:

    https://support.apple.com/en-us/HT202796

    Directions on installing Debian on a Macbook. For use with my now freed up 16 GB mini-SD chip. I tried the Ubuntu install, but it didn’t like me. I think I was trying to do an install in place and it wanted to treat the SD chip as a CD / DVD drive and install into somewhere else, so I might need to do a two chip process. One to install from and one to install too… but sorting that is for later.

    https://wiki.debian.org/MacBook

    Some peculiar “magic sauce” about booting Macs from USB drives. It takes some work to get a linux compatible boot partition on a USB that is also Mac workable…

    https://apple.stackexchange.com/questions/103535/can-i-boot-linux-any-distro-from-usb-on-a-macbook-air-that-has-no-hard-drive

    A place where you can do internet speed tests:

    https://www.speakeasy.net/speedtest/

    The how to for Ubuntu on a Mac, that has “some assembly required” that I apparently didn’t get quite right yet…

    http://randomtutor.blogspot.ca/2013_02_01_archive.html

    and another version:

    https://help.ubuntu.com/community/How%20to%20install%20Ubuntu%20on%20MacBook%20using%20USB%20Stick

    And another:

    https://tutorials.ubuntu.com/tutorial/tutorial-create-a-usb-stick-on-macos#0

    So I’m going to sort through all those at some point and see if I can succeed at putting a “Linux on a chip” on the Mac. If for no other reason than to limit the size of swap and maybe force a bit more memory management action and less writing GB of SD card at a shot.

    So with those links now stored here, I can get on with opening lots of other stuff on Opera and maybe crashing it / losing history / whatever…

  4. E.M.Smith says:

    Well, the Opera test has reached a conclusion.

    It, too, has memory use issues. It handles the memory a little better in that it keeps running without as much lag / pause. Even to the point where the Mac adds ANOTHER GB chunk of swap:

    total 5242880
    -rw------T  1 root  wheel   2.0G Mar 26 09:11 sleepimage
    -rw-------  1 root  wheel    64M May 12 13:37 swapfile0
    -rw-------  1 root  wheel    64M May 12 15:56 swapfile1
    -rw-------  1 root  wheel   128M May 12 15:56 swapfile2
    -rw-------  1 root  wheel   256M May 12 15:56 swapfile3
    -rw-------  1 root  wheel   512M May 12 15:56 swapfile4
    -rw-------  1 root  wheel   1.0G May 12 15:56 swapfile5
    -rw-------  1 root  wheel   1.0G May 12 15:56 swapfile6
    

    So at this point we’ve got, basically, 3 GB of swap in use plus 2 GB of memory for 5 GB total. All from about 1/2 dozen low weight web pages and watching a couple of moves on YouTube.

    Then A Very Odd Thing happened. Web pages refused to load with DNS failure to find names / numbers… Even a direct nslookup in a terminal window failed. Ping to the network failed. Something in the network stack was hosed.

    I had to try to connect to a non-working connection, then reconnect to my WiFi so as to refresh the DHCP, then it all was working again. This was after an exit, reload of Opera didn’t cut it.

    So net-net is Opera is a bit more workable until such time as the amount of swap in use crashes networking (or potentially something else depending on just what got swapped out in any one case…) So very not good.

    But now I know. They Mac continues to try to add swap to crazy levels and eventually goes a bit nuts. All because the browser guys seem unable to let go of memory once used. (It could be the OS not letting it go, but that would be unlikely).

  5. Chris in Calgary says:

    Incremental sloth notwithstanding (must be a subspecies of tree sloth :) ) malloc is something that plagues all C/C++ programmers. I’ve seen otherwise well respected programmers crash programs because they forgot to free vectors of pointers. It’s just not something that humans do well.

    I vow to use self-managing types (e.g. unique_ptr) whenever possible, in the future.

  6. E.M.Smith says:

    @Chris in Calgary:

    I suspect C++ has made it worse as folks use some function via inheritance and don’t even know that the underlaying has a malloc problem; then don’t bother profiling and testing enough to see the bad inherited effect… Then so much development is now done on super over configured nearly gaming platform machines that they never see the effects of their decisions.

    IMHO, THE biggest problem with bad code comes from developers being given top of the line super machines with 16 cores and 20 GB of memory. They ought to be given more representative machines of what the general public actually has. 4 or 5 year old single core machines with 512 MB of memory. THEN you would get tight efficient code.

  7. Larry Ledwick says:

    Or at least insist that QA which does the testing on the code, runs it on representative hardware while running 5 other programs like a large spread sheet a word document and a couple other browser tabs open running typical auto play video crap.

  8. jim2 says:

    Seems it would be easy enough to create a QA code checker for non-deallocated memory.

  9. E.M.Smith says:

    @Jim2:

    IF people cared enough to bother.
    IF people were aware of the issue.

    Both of which tend not to arise when you are using a 16 core 12 GB machine…. with 24 GB of swap on SSD…

  10. E.M.Smith says:

    Oh, and worth mentioning is Standard Programming Practices…

    So these programs DO release the memory. They just do it in keeping with Standard Programming Practice of having a “close and exit” module. So all that memory they grab is cleaned up at the time the “close and exit” is called. A “code review” will find good Programming Practice of cleaning up after yourself and closing files, releasing memory.

    So your programmer is happily smug that they have done a Very Good Job of housekeeping.

    What they missed is that with extended use, their memory consumption rises without end. It is the reuse of memory that’s lacking. But the practice of judicious re-use of memory is seen as being old fashioned. Why waste your expensive programmer time on that? Memory is cheap and you follow good programming practices with a “close and exit” process, so “Ship it and move on”. That attitude tends to come from Management who are, after all, the ones setting the overall attitudes.

    So QA is pushed to “Finish the testing and ship!” and the programmer is pushed to “Stop screwing around and ship it, it works doesn’t it?”. ( I’ve been in all three roles, BTW… Manager, QA department, programmer.) There’s always a list of B, C, and D bugs you don’t get done when ship date arrives. Sometimes even an A list. You try to get all the critical bugs stamped out, and most of the important ones; plus add the most desired features. Nobody pushes for a “minor annoyance” or a “trivial performance” issue to be addressed.

    I tried to make it a habit to pick off at least ONE item from the lowest priority queue and do it / fix it every major cycle; but wasn’t always successful.

    It’s like your A, B, C list of Things To Do. Somehow there’s always some stuff on the C list that just never gets done. Clean the garage. Paint the bottom of the koi pond. ;-) So you’ve hit all the A list things (meet Standard Programming Practice and do garbage collection at the end) and that C item of “optimize memory usage on prolonged use on old decrepit machines” never gets brought up in status meetings …

  11. Bill S. says:

    I’ve been told to try Opera to use to Remote Desktop to my work computer (PC to PC).
    Anybody have any pros or cons regarding Opera? New PC at work is Windows 10. Have not been able to Remote Desktop since the “upgrade”.

    I’m also trying to figure out what brand of emergency generator to buy for my wine cooler and refrigerator in case of an earthquake – in case anybody wants to go back to that vein.

    BTW – You can get Kiawe wood chunks from the internet if you want to do authentic huli-huli chicken on your charcoal grill (mesquite works ok – have not tried manzanita yet). Best recipe at
    https://www.hawaiimagazine.com/blogs/hawaii…/make_huli_huli_chicken
    And yes – I am one of those who needs to learn how to insert a link into a comment :-(

  12. Bill S. says:

    huh. Looks like cut and paste works fine for the link???

  13. jim2 says:

    Opera has a VPN mode. I use it at times, but it is noticeably slower than non-VPN.

  14. jim2 says:

    I know most here don’t like MS, and I understand why, but in C# there exists a USING statement. With it you can use a resource, and after use, the USING statement disposes it.

    It seems C and C++ could be extended just a little bit with some helpful memory management features. (Flame away :)

  15. E.M.Smith says:

    @Bill S.:

    Yes, just paste it in without any other text touching it. So (link) would not be a live link as the ( and ) would deactivate it if touching… but ( link ) works as white space adjoins the link.

    Opera is an OK browser. It has some quirks with WordPress (so, for example, doesn’t like the ‘edit a comment’ script. That may be due to this version on the Mac being a bit old…) But I’ve generally liked it. Just like FireFox types a bit better. The “Opera Turbo” seems to cache pages for faster reads with compression ( I think at an Opera compute cluster) and can be helpful on non-dynamic pages over slow comms links.

    It’s one of the browsers I uses often, if not my first choice.

  16. cdquarles says:

    It seems that all major browsers “leak”. For me, before I ditched it, Firefox was among the worst. Safari being the absolute worst. I’ve used Opera a good bit, which for a time was my main browser. When Brave came along, I went that route. As Brave has followed the rest ‘feature’ wise, they’re becoming like the rest in resource usage.
    I am an outlier in that I pretty much always keep dozens of tabs open when using web browsers.

  17. llanfar says:

    @cdquarles I tend the same way. I used to use TabGroup Manager before Firefox Quantum, which let me keep it to a single window. Now I run 5 windows with 10-50+ tabs each – delineated by area of focus (database, office, programming, research, personal).

  18. Steven Fraser says:

    @EM:I see the same thing on Win 8.1 Firefox with process sizes. Even if I am doing nothing, the processes are busy doing something, and growing.

    Regardless of the use of free(), memory allocated by the process moves the break, and cannot be given back to the OS for re-use until the process is stopped.

    That said, I think the issue is in the way Java is configured to run by the browser. I will do a bit of digging to see what I can find. the symptom on my systerm that this is happening is slow mouse tracking. If I exit without ending the tab, it re-starts them when Firefox is started again.

  19. llanfar says:

    @steven The Java plug-in is now disabled by current Firefox…

  20. wyoskeptic says:

    I’ve noticed this with Firefox for some time now. One of the areas where it gets really noticeable is using Bing search in particular for images or videos. I use a Toshiba laptop w/windoze 7. I bought a new one w/8. I hated Win 8 and when it insisted I had to upgrade to 10, I dumped it, used a bought version of 7 and tried to install. But since the way 8 configures the boot section would not allow the 7 disk to overwrite the hard drive, I had to do a complete format and lost a lot of drivers. Finally got it usable but without internal wireless or even support for the onboard Lan network card. Able to use via Cell phone spike through usb port. Works well enough to keep it as a back up, but otherwise a total wasted investment. Have been playing with putting Linux Mint on it and using it that way. But I digress.
    I originally like FF because it booted quickly and did not bloat up. Now it is as bad as old Internet Explorer and its decedents. But it is usable by way of clearing history on shut down and restarting. But even with that it takes CCleaner to free up all the memory that is tagged and cached and not released. So, I shut it down frequently and I periodically run CCleaner.
    What I keep on board as a backup is Vivaldi. Using it for odd ball searches and things which bloat up Firefox. It works well enough and anyone used to Firefox shouldn’t have any probs other than finding the desired add-ons and the like. As I understand it, Vivaldi was started by the same one who was involved with Opera (thus the name Vivaldi). Frankly I have had good luck with it and once Firefox gets too bad to use, hopefully Vivaldi will be a good replacement.
    I have not seen a lot of press for Vivaldi, but while it seems a little clunky in some ways, one thing I have never experienced so far is any problem with its memory handling. The only thing I don’t care for is that most of its add-ons are aps from Chrome.The less I have anything to do with Google the better I like it.

  21. E.M.Smith says:

    @CGQ:

    I’ve been meaning to give Brave a try, but just not had time for the trial yet. Then the old Mac is limited to 10.7.5 so then you get the problem of new applications requiring newer OS releases…

    I had a hard enough time finding a FireFox that would run on it…

    I used to be one of those folks who would keep 50 tabs open for things. Now I cut / paste the links into a Document… unfortunately, then I rarely get back to them in the doc…

    @Steven F:

    I’ve noticed that even if you are not actively using it, FF takes up a fair amount of process time as something behind the active window. It really ought to halt when not active, IMHO, but maybe it has to stay active to prevent windows dropping connections.

    @Wyoskeptic:

    Well, I was almost ready to say Microsoft was becoming a reasonable OS when I had a 7 laptop. Even made some reluctantly positive statements about it. Then 8 came out and I refused to “go there”, then 10 came out (and the skipping of 9 for stupidity…) and 10 was such a suck-hole…

    If you tried to do it wrong you could not have had a worse path…

    Linux can be a bit slow on new drivers, but eventually they show up… I like to get about 4 year old hardware, sometimes for free after MS makes them near worthless, and then put Linux on them and have something faster than the newest stuff.

  22. philjourdan says:

    Actually, Windoze 9 was skipped for a good reason. It seems a lot of code would check for the version of “9” (due to 95 and 98) and if found, say sorry. So Windows in order to bypass that potential bug-a-boo, skipped windows 9.

    Of course like Mac OS9 (which was out for a mere months before Mac/Unix aka X came out), it does look mighty “suspicious”

  23. E.M.Smith says:

    @Phil:

    The rationale was good, but the necessity was due to bad programming. Instead of checking explicitly for 95 or 98, folks checked for “starts with 9”. Bad programmer….

    Then the MS response puts a permanent “glitch” in the name space. I’d have likely used an encoding of “9” internally that was not a plain 9 (think unicode character that looks like a 9 but isn’t) and named the release “9” but used “9” as the identifier char. Then it would not tickle the code checking for “starts with 9” in old programs; and only new applications developers would need to know that you check for the funny (unique!) unicode to detect that release. Smooth, elegant, and the only folks who see it would be skilled in the art and appreciate the hack. See:
    https://www.irongeek.com/homoglyph-attack-generator.php

    9	9 9
    
  24. philjourdan says:

    I agree it is poor programming. Unfortunately that was not Microsoft’s to control. However your solution is elegant. But I guess using a Unicode 9 would have confused some of the other lazy programmers that could not check for a 95 or 98 as well, and Microsoft decided it was not worth the hassle.

    Whatever the reason, 9 seems to be the unlucky number in OS land.

  25. kneel63 says:

    “Whatever the reason, 9 seems to be the unlucky number in OS land.”

    Wasn’t there an OS/9 or something like that?

  26. E.M.Smith says:

    https://en.wikipedia.org/wiki/Mac_OS_9

    Mac OS 9 is the ninth and final major release of Apple’s classic Mac OS operating system. Introduced on October 23, 1999, it was promoted by Apple as “The Best Internet Operating System Ever”, highlighting Sherlock 2’s Internet search capabilities, integration with Apple’s free online services known as iTools and improved Open Transport networking. While Mac OS 9 lacks protected memory and full pre-emptive multitasking, lasting improvements include the introduction of an automated Software Update engine and support for multiple users.

    Apple discontinued development of Mac OS 9 in 2001, transitioning all future development to Mac OS X. The final updates to Mac OS 9 addressed compatibility issues with Mac OS X while running in the Classic Environment and compatibility with Carbon applications. At the 2002 World Wide Developers Conference, Steve Jobs began his keynote address by staging a mock funeral for OS 9.

    If that isn’t the one you were thinking of, see:

    https://en.wikipedia.org/wiki/OS9

    OS9, OS-9, or OS 9 may refer to:

    Mac OS 9, an operating system for the Apple Macintosh
    iOS 9, the ninth version of the iOS operating system
    OS-9, a Unix-like real time operating system
    OS9 (gene), which encodes protein OS-9 in humans

    of which the Unix-like is the only real option:

    https://en.wikipedia.org/wiki/OS-9

    OS-9 is a family of real-time, process-based, multitasking, multi-user operating systems, developed in the 1980s, originally by Microware Systems Corporation for the Motorola 6809 microprocessor. It was purchased by Radisys Corp in 2001, and was purchased again in 2013 by its current owner Microware LP.

    The OS-9 family was popular for general-purpose computing and remains in use in commercial embedded systems and amongst hobbyists. Today, OS-9 is a product name used by both a Motorola 68000-series machine language OS and a portable (PowerPC, x86, ARM, MIPS, SH4, etc.) version written in C, originally known as OS-9000.

  27. Steve C says:

    I certainly agree re Firefox. In its youth it was fun to use, lightweight, customisable, quick and generally a head ahead. But in its old age, it’s become a bloated, sluggish, uncustomisable brute with a pile of ongoing “issues” like the above, most of which people have been discussing online forever but which still lurk, undealt-with, to trip the user. it was the promise of an experience more like the early Firefox that led to me trying PaleMoon a few years ago, and I haven’t looked back (well, except when I’m using the Tor browser. It always reminds me why I changed, alas.)

    Having said that, I must admit to being one of those who opens a browser, looks around a few relevant sites (rarely as many as half a dozen tabs open, except when the cricket coverage is on), then shuts it down ’til next time (except when the cricket coverage is on). It’s currently using 336.5 MiB, with just this page and my (local) “blue page” of mixed links open for now – although 300-odd megs does seem like a lot of memory for what it’s doing. Call it around 3KB for a nominal page of text, that’s the equivalent of well over 100,000 pages to keep track of me opening a couple of text-rich htm pages, which definitely seems a bit exuberant.

    The only browsing problem I’m looking at currently is that PM is now bringing up messages “offering” to disable NoScript for being buggy. Granted it breaks the odd over-scripted website, but I’d rather have broken sites which I can then fix as necessary than have to allow God-knows-what to run whenever I click a link. The PM website is getting quite disparaging about it (a bad sign), but NoScript has worked well for me for years. Anyone here having browser stability or other serious issues with NoScript (any browser, IIRC it’ll work with most), or has someone leant on PM? I’ve seen the web without NoScript at my local library, and had unpleasant flashbacks for days afterwards. Ugh! :-(

  28. llanfar says:

    @Steve NoScript 100% of the time here.

  29. jim2 says:

    I have a simple html page, a text file, with the links I visit most often. Make it my home page. Loads fast and very little typing required.

    H1 {color: green}
    HR {color: yellow}
    TD {font-weight:bold;
    align:left}

    TD#GenTD {text-align:left;
    font-weight:bolder;
    font-size:x-large;
    font-style:oblique;
    color: green;

    TD#NewsTD {text-align:center;
    font-weight:bolder;
    font-size:x-large;
    font-style:oblique;
    color:green;

    BODY {background-color:navy;
    width: 100%}

    Our Home Page

    Hot Links

    title1
    &nbsp
    title2
    &nbsp
    title3
    &nbsp
    title4
    &nbsp
    title5

    * W3 Schools
    &nbsp
    * Fox News
    &nbsp
    * Space Weather Charts
    &nbsp
    * First Trade
    &nbsp
    *Briefing.com

    * Thorium Energy
    &nbsp
    * Drudge
    &nbsp
    * Science Daily
    &nbsp
    * The Street
    &nbsp
    * Zacks

  30. jim2 says:

    Opps, didn’t intend for the links to do that.

  31. Larry Ledwick says:

    I use three browsers, fire fox, brave and pale moon, depending on what I am doing. Twitter runs by itself on brave (since it is pretty locked down) and that way twitter can only see what is going on in that browser, this forum I usually run on fire fox, and I use pale moon to do quick looks at things that brave refuses to open or refuses to play the videos, then close it. By doing that I have mostly eliminated system hangs from browsers going comatose.

  32. E.M.Smith says:

    @Larry:

    As I’m distributed across several OS types of several vintages I use different browsers in part based on what is available on a given platform. For the Linux boxes, as they are light weight ARM systems, I generally go with FireFox (it is ubiquitous and does actually work). Chrome has tended to either be unavailable or buggy as hell. (Actually Chromium as it isn’t official Google Chrome…) then many of the “variety browsers” (like Pale Moon and Brave and…) have not got ARM ports done or working well (yet).

    On the Chrombox I run Chrome. Mostly as that box is used only for media center so, well, what’s going to leak? That I watched Lady Gaga videos? YouTube already told Google that…

    On the Mac, due to OS type and age, I’ve basically got a choice of Old Safari (gak), FireFox (the hog) or Opera (reasonable). I use all 3 for different things… in different places.

    On the tablet (oh, yeah, need to replace the power cord and start using it again…) I like Pale Moon (it is actually there!) and a couple of others.

    So it goes…

  33. philjourdan says:

    @E/M/ – Yes; OS9 came out right after OS8 and was for all intents and purposes OS8. The only reason they released a “9” was so they could call X 10, when in reality Jobs created X when he was with the NeXt project (OS X is NeXt ported over to the Mac). They tried to convince the gullible that the X was for 10, not Unix. But that is what the X was for, not 10, but Unix.

Comments are closed.