I Moved to Linux and It’s Even Better Than I Expected
-
Emmm... I think Win2k installs on WinNT folder by default. Why don't you check for it too?
btw, yet another good reason for not having system drive installed on C:
-
I grew up in a household which only had Mac Classic computers
https://www.youtube.com/watch?v=VKHFZBUTA4k
That said: DOS was a stupid operating system. Even before the IBM PC was released, Cromemco was shipping the Unix-like Cromix on its higher-spec S-100 Z-80 machines.
Given the 8088's ability to use segment registers to achieve what Cromemco's Z-80 boxes did with bank switching, the original IBM PC hardware should have been technically capable of supporting a Cromix port. Shipping it with yet another shit-grade CP/M clone might have been expedient, but it was technically very poor.
-
I don't think there is any way to limit the size of specific folders
Apparently, at least [url=https://technet.microsoft.com/en-us/library/cc754810(v=WS.10).aspx]Windows Server 2003 and 2008 can do that[/url]:Introduced with the Windows Server® 2003 R2 operating system, File Server Resource Manager is a suite of tools in the Windows Server® 2008 operating system that enables administrators to place storage limits on volumes and folders
No idea if FSRM is also in more recent versions, and/or if it still has that ability.
-
Because
format c:
probably.
-
btw, yet another good reason for not having system drive installed on C:
Yeah, 'cause
%SystemDrive%
is really unheard of.
-
How do you continue to run correctly if malloc() starts failing?
Depends greatly on what the app is doing.
If your app is keeping caches of data, it can blow away some cache in order to free up RAM for that malloc(). If there are multiple threads running, it could signal some to stop, possibly freeing the memory used by them. If there are other less important processes running (as a part of a multi-process system), it could tell some to terminate. It could gracefully quit and restart. Or at minimum, it could log a critical error before terminating.
Any of the above would be better than ignoring the error and later crashing when the NULL pointer is dereferenced.
what to do if there isn't enough memory available to report failure.
You pre-allocate some memory at startup, which can be used by the no-memory-handling code. It doesn't take much to send a message to the syslog and maybe pop up an alert on the desktop.
-
I grew up in a household which only had Mac Classic computers. What the fuck is wrong with your assholes?
You grew up around Apple, and you think we're assholes?
Wow, that Apple reality distortion field is way more powerful than I thought.
Yeah, 'cause %SystemDrive% is really unheard of.
Even virus writers have "WORKS ON MY MACHINE" syndrome.
-
@Lorne_Kates said:
You grew up around Apple, and you think we're assholes?
How does that make any sense?
-
@cheong said:
btw, yet another good reason for not having system drive installed on C:
Yeah, 'cause
%SystemDrive%
is really unheard of.
Because even legitimate installers from well known vendor hardcoded C: for installation destination and D: as source drive at the time...
-
@Lorne_Kates said:
Even virus writers have "WORKS ON MY MACHINE" syndrome.
For that claim, I'll add comment that even Visual Studio 2008 SP1 installer will not detect your installation unless you install it in C:.That's nothing too surprising.
-
Because even legitimate installers from well known vendor hardcoded C: for installation destination and D: as source drive at the time...
-
@cheong said:
Because even legitimate installers from well known vendor hardcoded C: for installation destination and D: as source drive at the time...
Come on. Nobody would ever ship a product with [url=http://arstechnica.com/gaming/2007/12/eve-online-trinity-borks-windows-deletes-boot-ini/]that kind of error[/url]...
-
@rc4 said:
@cheong said:
Because even legitimate installers from well known vendor hardcoded C: for installation destination and D: as source drive at the time...
Come on. Nobody would ever ship a product with [url=http://arstechnica.com/gaming/2007/12/eve-online-trinity-borks-windows-deletes-boot-ini/]that kind of error[/url]...
Reminds me of the time that one of the previous employees of my group developed an installation script for some software. When our lead sysadmin (who is still here, relating this story) ran it on his system to check it before he deployed it company-wide via SCCM, it deleted most of the files on his HDD.
-
And I thought software that need a batch file with "subst" command to load is bad enough...
-
Aborting on malloc() fail isn't a sin for most code.
It's a situation that people rarely test, where there is a lot of effort to do correctly, and a waste to do it half-assed.On Linux, malloc will never fail, because it just reserve the address space. The failure will happens when the memory is actually used, with a page fault. So, for Linux-only code it's always a waste.
-
On Linux, malloc will never fail
That depends on your overcommit settings. So: no, you cannot assume that.
-
@fbmac said:
On Linux, malloc will never fail
That depends on your overcommit settings. So: no, you cannot assume that.
also assuming that
@fbmac said:
On Linux, malloc will never fail
That depends on your overcommit settings. So: no, you cannot assume that.
there is address space available to reserve.
-
When did you last unintentionally try to allocate more than, say, 250TB?
-
When did you last unintentionally try to allocate more than, say, 250TB?
never, still it's an assumption that is not always true.
maybe you're running discourse and it decided to shit its pants all over your process address space?
-
When did you last unintentionally try to allocate more than, say, 250TB?
Even if you don't, and even if you ignore the fact that some server admins disable overcommit altogether:
According to its documentation, the kernel uses heuristics to determine whether an allocation is OK or not in its default setting. [source] Assuming that malloc() never returns NULL is just plain wrong, even on Linux.
-
or did they finally let that bit of backwards compat go?
Unlikely. Things like this tend to linger around - Intel didn't get rid of the A20 switch until goddamn Haswell came around.
-
Wait, 95 posts and nobody's
'shoppedgimped in a football? Jesus, this forum sucks.FTFY
-
Believe it or not, I actually went to college on a football scholarship.
-
@Lorne_Kates said:
When I started working, I got into pentiums, and it all gets hazy from there.
I've seen what pentium addiction does to people. I had to stage an intervention once, it wasn't pretty.
-
@Lorne_Kates said:
When I started working, I got into pentiums, and it all gets hazy from there.
I've seen what pentium addiction does to people. I had to stage an intervention once, it wasn't pretty.
I have a pentium addiction. One of my workstations is a dual-socket system with a pair of dual-core Pentium 4's. P4 x 2 x 2 = 16 pentiums! I think it's HyperThreaded too so it kind of looks like 32 pentiums.
-
@tar said:
@Lorne_Kates said:
When I started working, I got into pentiums, and it all gets hazy from there.
I've seen what pentium addiction does to people. I had to stage an intervention once, it wasn't pretty.
I have a pentium addiction. One of my workstations is a dual-socket system with a pair of dual-core Pentium 4's. P4 x 2 x 2 = 16 pentiums! I think it's HyperThreaded too so it kind of looks like 32 pentiums.
i see it's time to stage another intervention......
-
-
P4 x
2 x 24,195,835 / 3,145,727 = 5.3349562756 pentiums
I’m sure you meant to say that.
-
The famous FP bug rearing up again?