But I need build instructions and test instructions and possibly a youtube video, written/made for a student, not for a programmer that already knows a number of things about firefox. That is the way I feel about most open-source projects. I don't want to contribute in huge quantities, but only bugfixes, in any area and not limited to any particular technology. Sadly, I see such build-instructions missing or the build-instructions are too complicated in major open-source projects that could use bug-fixers
You should try fixing some bugs in Sunbird, if Mozilla interests you but the hugeness of Firefox is intimidating. I was able to contribute code (granted, only two lines) to Sunbird that fixed a real live bug, and I was in high school at the time.
Here, let me click on the top link for "firefox build instructions" in google: simple firefox build [mozilla.org]. Looks pretty standard to me. Tests, if there are any, are usually automated or findable by a similar exercise.
The majority of us use Windows, and will therefore probably want to develop on that platform.
If you read the Windows section of the page you linked to, the very first line is "Building on 64-bit Windows does not seem to be supported."
If you read the rest, you get told about using Visual Studio Express Editions and Windows SDKs, but as anyone who's tried it will know, just finding and installing the right SDKs there can be tricky. (Microsoft's own web site had links to an o
The majority of us use Windows, and will therefore probably want to develop on that platform.
Right...
Seriously, if you think this is a "simple" build procedure that's going to get casual volunteers contributing small fixes, you're not part of the solution, you're part of the problem.
All that proprietary closed-source software required to build Open Source software (any software, really). Difficult to obtain, difficult to install and difficult to configure.
If there is a browser/extention (they run at browser level)/plugin(yes even a flash or adobe exploit) or other program vulnerability they can perminantly modify your firefox binary to execute whatever code they want. In addition to having your user account, where all your data is, completely owned, no OS has a particularly good record on preventing malicious binaries from getting root (ubuntu with sudo is particularly bad as it can just request permisions just after you grant another process root using sudo
Your post says "but also fixes the annoying slow-startup on Windows." which suggests that all Windows users were experiencing slow starts. That's not the case at all. It was only a small fraction of users affected by the now fixed issue.
And for the record, the security flaw was already fixed, even before it was lifted from our bug database and turned into a public exploit. It just takes a few days to get everything in order for a release to users.
It was user situation dependent. Firefox was reading all of a user's temp files to seed its RNG or something along those lines so if you had a lot of large temp files your startup time would be quite large.
Regardless, it still takes 5x Chrome's startup time with the fix so... peh.
I have never understood why people make such a big deal over Firefox startup times. It's a few seconds. On my two-year-old laptop, Firefox 3.5.1 starts in two seconds. Granted, Chrome starts in less than one second, but in absolute difference it's about a second.
OS dependent. They coded for the case where Windows CE/2000 did not have a certain call and they wanted to get good entropy for their RNG in NSS. https://bugzilla.mozilla.org/show_bug.cgi?id=501605 [mozilla.org]
From the link, it appears that files (probably having an excessive amount of files) in the IE cache was slowing down Firefox cache? Isn't the Firefox cache entirely separate? Does it look in the IE cache to try to be friendly and helpful, and if so, can that behavior be turned off?
NSS (Network Security Services) 3.12.3 is using IE temporary internet files to generate seeds. Sounds thoroughly stupid to me, as it means that if you never use Internet Explorer, your cryptographic seeds won't change. How about using the process list or something not Hard Drive dependent to generate the seeds instead?
On further study, it NSS DOES use process IDs and many, many other factors to generate the seeds. Searching the additional file locations ("C:\Documents and Settings\*user*\Local Settings\History", "C:\Documents and Settings\*user*\Local Settings\Temporary Internet Files", "C:\Documents and Settings\*user*\My Recent Documents", "C:\Documents and Settings\*user*\Temp\", "Recycle Bin", and "Network Neighborhood") were added because some older OSs (Win2k and WinCE) didn't have strong enough build-in pseudo-random number generators.
This patch changed NSS to use the built-in PRNG in Windows XP and up which uses "process ID and thread ID, the system clock, the system time, the system counter, memory status, free disk clusters, andthe hashed user environment block".
No less a personage than Brendan Eich says the whole issue with slow startup in the NSS module is snake oil that does nothing but "waste users' time at startup pretending to scrape entropy off the filesystem."
Becuase Gears uses low-level binary hooks (e.g. completely replacing the Firefox HTTP cache with its own) and presumably doesn't want to worry about your browser crashing due to a code change on the Firefox end?
"...fix a problem where Firefox on a Sparc platform would crash when visiting www.hp.com!"
Much like the memory leak to nowhere, It wasn't a problem - it was a feature!
by Anonymous Coward
on Friday July 17, @06:05AM (#28727661)
so can anyone tell me why Firefox felt like it had to scan my hard drive in the first place? i had it set to delete history on exit. why then did it feel like it had to go looking in *other* programs' folders for history files?
I mean, I've given up on scaling fonts lager on the fly (as opposed to zoom), but how about 'paste and go' for urls - like opera has had for years (and now chrome)
So what your saying is Microsoft could fix all of their problems by changing the color of the screen?
Microsoft tried that twice on the Xbox 360, and people continued to complain about the red ring of death (general hardware failure) and the green screen of death (E74 error).
Actually, the linux blue screen of death is blinking of 2 (or is it three?) of the keyboard leds. Though support for blue screen of death is coming, by the name of kernel mode-settting. It is pretty rare, though.
Lockups I have seen, too, in both linux and windows. Lots of cases is hardware problems, but your problem sounds like a driver issue. Using proprietary drivers, perhaps?
Lockups I have seen, too, in both linux and windows. Lots of cases is hardware problems, but your problem sounds like a driver issue. Using proprietary drivers, perhaps?
This is true. I've had my share of complete freezes under Linux. Ironically though, SSH access to the box still typically works and I can kill X if ctrl+alt+backspace doesn't work. It's rare to have a freeze that completely evicts all sense of response from the system (though I've had this happen before).
Interestingly, the last unusual behavior I had under Linux was when a video card blew 4 out of 7 or 8 capacitors. That was a real treat.
You can hardly call it a complete freeze if "only" X is frozen. Still pretty annoying but as you say you can usually recover by killing and restarting X.
> Still pretty annoying but as you say you can usually recover by killing and restarting X.
a) If you are a "Desktop Linux" user running actual Desktop applications, that means you lose most of your unsaved work (if there is a way to not lose the unsaved work, please let me know). b) If you use X as just a way to run screen/vi/emacs and browsers, then you are less affected.
Basically if I let my mom/uncle/aunt use "Desktop Linux" and X locks up, it's effectively as bad as a BSOD for them.
Saying X freezing is not a problem since you can usually recover by killing and restarting it is like saying that Windows 95 is stable as long as you regularly shutdown/exit to dos and type win to restart it[1].
[1] you could actually do that in the old days of Win 95:).
tee? Really? What the hell sort of DESKTOP APPLICATIONS produce all of their output on the terminal? OpenOffice? GIMP? KMail? GVim?
No, the only solution is the Jesus rule. Save your files. Save them early, save them often. Not just because the system is going to crash, but because you never know when the power will fail, lightning will strike, or a cow-orker will trip over your power cord.
This is exactly what I was going to suggest. SSH usually runs fine, or using CTRL+ALT+Backspace... if X crashes you won't be able to CTRL+ALT+FX I'm pretty sure.
I've only ever had Linux hard lock when I've been testing out early alpha stuff on a sandbox (used to be an old machine, now it's just a VM).
If you can't bring it back and you're not doing something stupid then it's probably hardware, so you might want to run some diags.
If you can't bring it back and you're not doing something stupid then it's probably hardware
As I said in the "Blue screen" post, I can't even use the "Magic SysRq key [wikipedia.org]". I've invested several days in solving this. I'm definitely not doing something stupid. It definitely isn't the hardware. It's a problem between ATI's drivers and the rest of the OS.
Scroll down to firefox-3.5. Stupidly, this package doesn't overwrite the firefox package, meaning that applications will still use 3.0 to open links. Even if you remove the firefox package, firefox-3.5 is still not used. Changing the webbrowser in preferred applications seems to work on some applications...
Anyway, in the end I just simlinked like so: ln -s/usr/bin/firefox-3.5/usr/bin/firefox, and everything worked great.
Scroll down to firefox-3.5. Stupidly, this package doesn't overwrite the firefox package, meaning that applications will still use 3.0 to open links. Even if you remove the firefox package, firefox-3.5 is still not used.
In Jaunty, this is because Firefox 3.0 remains the default version of Firefox (and the firefox package always points towards the default version of Firefox for that release). In Karmic, this is because the developers haven't switched the default from 3.0 to 3.5 yet, though they will soon [asoftsite.org].
Going by previous versions of firefox, shouldn't it be 3.5.0.1 rather than 3.5.1?
Mozilla decided to simplify that with Firefox 3 (note that the upcoming security release for Firefox 3 is 3.0.12, not 3.0.0.12). Exactly why they used four numbers in the first place is something I don't know, it seems it started with Firefox 1.5. I know that one advantage touted of XPCOM was the ability to easily make incremental updates, so maybe there was a plan for a Firefox 1.5.1 and 1.5.2 (with the final number for each being used for security updates). Of course that would've been complicated and silly, so it seems the plan was abandoned and the version number compacted.
I'd fix bugs and contribute quality code (Score:2, Informative)
Re:I'd fix bugs and contribute quality code (Score:5, Interesting)
You should try fixing some bugs in Sunbird, if Mozilla interests you but the hugeness of Firefox is intimidating. I was able to contribute code (granted, only two lines) to Sunbird that fixed a real live bug, and I was in high school at the time.
Parent
Re:I'd fix bugs and contribute quality code (Score:5, Informative)
Here, let me click on the top link for "firefox build instructions" in google: simple firefox build [mozilla.org]. Looks pretty standard to me. Tests, if there are any, are usually automated or findable by a similar exercise.
Parent
Re: (Score:3, Interesting)
That's cute, but missing the point.
The majority of us use Windows, and will therefore probably want to develop on that platform.
If you read the Windows section of the page you linked to, the very first line is "Building on 64-bit Windows does not seem to be supported."
If you read the rest, you get told about using Visual Studio Express Editions and Windows SDKs, but as anyone who's tried it will know, just finding and installing the right SDKs there can be tricky. (Microsoft's own web site had links to an o
Re: (Score:3, Insightful)
The majority of us use Windows, and will therefore probably want to develop on that platform.
Right...
Seriously, if you think this is a "simple" build procedure that's going to get casual volunteers contributing small fixes, you're not part of the solution, you're part of the problem.
All that proprietary closed-source software required to build Open Source software (any software, really). Difficult to obtain, difficult to install and difficult to configure.
It sounds like Windows is the problem. All of
Re: (Score:3, Interesting)
If there is a browser/extention (they run at browser level)/plugin(yes even a flash or adobe exploit) or other program vulnerability they can perminantly modify your firefox binary to execute whatever code they want. In addition to having your user account, where all your data is, completely owned, no OS has a particularly good record on preventing malicious binaries from getting root (ubuntu with sudo is particularly bad as it can just request permisions just after you grant another process root using sudo
slow start for _some_ (Score:5, Informative)
Re:slow start for _some_ (Score:4, Funny)
slow start for _some_. Miniature Type-R stickers for others.
Parent
Re: (Score:2)
Was it OS dependent, or hardware dependent?
I had the issue in winxp 32bit sp3.
Re: (Score:2)
Was it OS dependent, or hardware dependent?
I had the issue in winxp 32bit sp3.
Beats me - but I don't have it on that OS. It still takes 2 seconds to start.
Re: (Score:2, Informative)
It was user situation dependent. Firefox was reading all of a user's temp files to seed its RNG or something along those lines so if you had a lot of large temp files your startup time would be quite large.
Regardless, it still takes 5x Chrome's startup time with the fix so... peh.
Re: (Score:3, Insightful)
Re:slow start for _some_ (Score:5, Informative)
Parent
Re: (Score:3, Interesting)
Re:slow start for _some_ (Score:5, Interesting)
NSS (Network Security Services) 3.12.3 is using IE temporary internet files to generate seeds. Sounds thoroughly stupid to me, as it means that if you never use Internet Explorer, your cryptographic seeds won't change. How about using the process list or something not Hard Drive dependent to generate the seeds instead?
Parent
Re:slow start for _some_ (Score:5, Informative)
On further study, it NSS DOES use process IDs and many, many other factors to generate the seeds. Searching the additional file locations ("C:\Documents and Settings\*user*\Local Settings\History", "C:\Documents and Settings\*user*\Local Settings\Temporary Internet Files", "C:\Documents and Settings\*user*\My Recent Documents", "C:\Documents and Settings\*user*\Temp\", "Recycle Bin", and "Network Neighborhood") were added because some older OSs (Win2k and WinCE) didn't have strong enough build-in pseudo-random number generators.
This patch changed NSS to use the built-in PRNG in Windows XP and up which uses "process ID and thread ID, the system clock, the system time, the system counter, memory status, free disk clusters, andthe hashed user environment block".
Parent
Re: (Score:3, Informative)
Re: (Score:2)
Firefox has a healthy eating mode?
Good. (Score:5, Insightful)
HP.com (Score:2)
they also had to fix a problem where Firefox on a Sparc platform would crash when visiting www.hp.com!"
Anyone that sees a downside to not accessing hp.com must not use NoScript.
Holy negatives Batman! (Score:3, Funny)
"Now correct me if I'm incorrect, but was I told it's untrue that people in Springfield have no faith? Was I not misinformed?"
Google Gears disabled again?! (Score:5, Insightful)
Kind of offtopic, but by upgrading to FF 3.5.1, Google Gears is again disabled. Why did Google allowed it to be compatible with only 3.5.0?!
Re: (Score:3, Informative)
Becuase Gears uses low-level binary hooks (e.g. completely replacing the Firefox HTTP cache with its own) and presumably doesn't want to worry about your browser crashing due to a code change on the Firefox end?
And from Unix wars... (Score:2, Funny)
Re: (Score:2)
Isn't that OS X?
BSD isn't dead heck it has overtaken Linux by strides and is a serious contender to windows. Just just downplay the BSD roots of the OS.
problem? (Score:3, Funny)
Firefox 3.5.1 released (Score:3, Insightful)
so can anyone tell me why Firefox felt like it had to scan my hard drive in the first place? i had it set to delete history on exit. why then did it feel like it had to go looking in *other* programs' folders for history files?
gpg: Note: This key has expired! (Score:3, Interesting)
gpg: Signature made 07/15/09 19:56:19 using DSA key ID 17785FE8
gpg: Good signature from "Mozilla Software Releases <releases@mozilla.org>"
gpg: Note: This key has expired!
Primary key fingerprint: 8D6F 1BA4 A340 4DDB 3F2F D080 7447 4499 8123 47DD
Subkey fingerprint: 3338 E6BA FF10 3B3D A6A9 E424 B57B 5484 1778 5FE8
Does it finally have paste and go? (Score:4, Interesting)
I mean, I've given up on scaling fonts lager on the fly (as opposed to zoom), but how about 'paste and go' for urls - like opera has had for years (and now chrome)
Re: (Score:2, Insightful)
I have yet to see a single blue screen on Linux.
FOSS isn't perfect, it's just a whole lot better than one of the competitors.
and I enjoy my FOSS haven very much, thank you.
Re:FROSTY PISS (Score:5, Funny)
Parent
Re:FROSTY PISS (Score:5, Funny)
Make it black and hope people just think they accidently turned their computer off.
Parent
Re:FROSTY PISS (Score:5, Funny)
Obama-mode
Parent
Green and red indicators of death (Score:4, Funny)
So what your saying is Microsoft could fix all of their problems by changing the color of the screen?
Microsoft tried that twice on the Xbox 360, and people continued to complain about the red ring of death (general hardware failure) and the green screen of death (E74 error).
Parent
Re:Blue screen (Score:5, Informative)
Actually, the linux blue screen of death is blinking of 2 (or is it three?) of the keyboard leds. Though support for blue screen of death is coming, by the name of kernel mode-settting. It is pretty rare, though.
Lockups I have seen, too, in both linux and windows. Lots of cases is hardware problems, but your problem sounds like a driver issue. Using proprietary drivers, perhaps?
Parent
Re:Blue screen (Score:5, Interesting)
This is true. I've had my share of complete freezes under Linux. Ironically though, SSH access to the box still typically works and I can kill X if ctrl+alt+backspace doesn't work. It's rare to have a freeze that completely evicts all sense of response from the system (though I've had this happen before).
Interestingly, the last unusual behavior I had under Linux was when a video card blew 4 out of 7 or 8 capacitors. That was a real treat.
Parent
Re: (Score:2, Insightful)
You can hardly call it a complete freeze if "only" X is frozen. Still pretty annoying but as you say you can usually recover by killing and restarting X.
Re:Blue screen (Score:5, Insightful)
a) If you are a "Desktop Linux" user running actual Desktop applications, that means you lose most of your unsaved work (if there is a way to not lose the unsaved work, please let me know).
b) If you use X as just a way to run screen/vi/emacs and browsers, then you are less affected.
Basically if I let my mom/uncle/aunt use "Desktop Linux" and X locks up, it's effectively as bad as a BSOD for them.
Saying X freezing is not a problem since you can usually recover by killing and restarting it is like saying that Windows 95 is stable as long as you regularly shutdown/exit to dos and type win to restart it[1].
[1] you could actually do that in the old days of Win 95
Parent
Re:Blue screen (Score:4, Insightful)
tee? Really? What the hell sort of DESKTOP APPLICATIONS produce all of their output on the terminal? OpenOffice? GIMP? KMail? GVim?
No, the only solution is the Jesus rule. Save your files. Save them early, save them often. Not just because the system is going to crash, but because you never know when the power will fail, lightning will strike, or a cow-orker will trip over your power cord.
Parent
Re:Blue screen (Score:5, Insightful)
Ironically though, SSH access to the box still typically works...
That is not ironic: it is good design...
Parent
Re: (Score:2)
This is exactly what I was going to suggest. SSH usually runs fine, or using CTRL+ALT+Backspace... if X crashes you won't be able to CTRL+ALT+FX I'm pretty sure.
I've only ever had Linux hard lock when I've been testing out early alpha stuff on a sandbox (used to be an old machine, now it's just a VM).
If you can't bring it back and you're not doing something stupid then it's probably hardware, so you might want to run some diags.
Re: (Score:2, Interesting)
As I said in the "Blue screen" post, I can't even use the "Magic SysRq key [wikipedia.org]". I've invested several days in solving this. I'm definitely not doing something stupid. It definitely isn't the hardware. It's a problem between ATI's drivers and the rest of the OS.
Re:Someone tell Canonical. (Score:5, Informative)
I installed it ages ago:
aptitude install firefox-3.5
http://packages.ubuntu.com/search?searchon=names&keywords=firefox-3.5 [ubuntu.com]
Parent
Re:Someone tell it to Canonical. (Score:4, Informative)
Just add the fta repository & install "firefox-3.5". They even link to a mozilla daily build repository if that's your thing.
Parent
Re: (Score:3)
So - who got brave, and installed FF 3.6? Am I that brave, or am I not? Hmmmm........
Re: (Score:2)
Re: (Score:2, Informative)
Scroll down to firefox-3.5. Stupidly, this package doesn't overwrite the firefox package, meaning that applications will still use 3.0 to open links. Even if you remove the firefox package, firefox-3.5 is still not used. Changing the webbrowser in preferred applications seems to work on some applications...
Anyway, in the end I just simlinked like so: ln -s /usr/bin/firefox-3.5 /usr/bin/firefox, and everything worked great.
Re: (Score:3, Informative)
Ubuntu uses update-alternatives to select between different packages providing the same functionality
to see which browsers are installed:
/usr/bin/firefox-3.5
update-alternatives --list x-www-browser
to select firefox-3.5:
update-alternatives --set x-www-browser
Re: (Score:2)
Scroll down to firefox-3.5. Stupidly, this package doesn't overwrite the firefox package, meaning that applications will still use 3.0 to open links. Even if you remove the firefox package, firefox-3.5 is still not used.
In Jaunty, this is because Firefox 3.0 remains the default version of Firefox (and the firefox package always points towards the default version of Firefox for that release). In Karmic, this is because the developers haven't switched the default from 3.0 to 3.5 yet, though they will soon [asoftsite.org].
Re:version numbers (Score:5, Informative)
Going by previous versions of firefox, shouldn't it be 3.5.0.1 rather than 3.5.1?
Mozilla decided to simplify that with Firefox 3 (note that the upcoming security release for Firefox 3 is 3.0.12, not 3.0.0.12). Exactly why they used four numbers in the first place is something I don't know, it seems it started with Firefox 1.5. I know that one advantage touted of XPCOM was the ability to easily make incremental updates, so maybe there was a plan for a Firefox 1.5.1 and 1.5.2 (with the final number for each being used for security updates). Of course that would've been complicated and silly, so it seems the plan was abandoned and the version number compacted.
Parent