Well cried. Even the knowledgable proponents of GC (Boehm, Zorn) admit there's an overhead to it, but try to dismiss the overhead as negligible, using wording such as "typically competitive" and "only slightly higher". The quick translation of such claims into English is "yup, you got us, they are actually slower and use more memory, sorry".
I'm not sure I really see why the following (from the summary) is an _OR_:
Is Google trying to drive more people towards its own online suite of office applications?
Or
has it been stung into action by Steve Ballmer's recent comment that Microsoft Office faces stronger competition from StarOffice than it does Google Docs and Spreadsheet?"
I'd suggest that it was an _AND_ instead. _AND_ what you said too. The overlap of the 3 is pretty complete.
1) There's an email address or telephone number where you make your request for support. 2) Someone will answer that within 3, or sometimes 7, working days to indicate they've received your request for support.
Google have been dicking around with that form in the last week or so. They _removed_ the date range search functionality, for example. They put it back again when they realised they'd been brainless idiots. However, that bug illustrates that 10000 Ph.D.s can still be appear to be brainless idiots. One thing that annoyed me was the fact that they removed the message-id search. Given that message-ids are the PRIMARY KEY of Usenet, that demonstrates google have lost the plot when it comes to Usenet.
Feel free to bookmark, use, and suggest improvements to: http://theanna.org/goog.html http://theanna.org/goo.html (They're both practically the same, one's old-school 90s HTML, the other a little less-so; both will be maintained.)
Having seen some Theroux again after many years without, he also failed to live up to my memories (I remember him visiting the KKK or the US Nazi party what seems like over decades ago, that series I did enjoy). My biggest problem with the supersized NRA member's recent offerings was his horrifically unsubtle selective editing. TV Nation was acceptable, but you can't chop like that for a whole feature-length movie.
Likewise in Finland. Single-use random 4-digit ids. We've had them for 15 years or more. (So in the early 90s, Finnish banks were more security conscious than most modern-day US or UK banks.)
All my recent "fade to nothingness within a year" CDR's have been "high quality verbatim". I no longer think they deserve the adjective. Better than the old Traxdata gold, but even my own memory's better than those were. What was I talking about, again?
The third option is "something between telnetd and the kernel is broken - non-root users are able to do things that only root users should be able to do".
Or in the context of the earlier either/or question - "neither".
You misread or misunderstood. telnetd is not doing what it is supposed to do. telnetd's supposed to attempt to open a priveleged port (<1024), and supposed to fail if you do not have the required permissions to do so (i.e. be root). (Of course the way it's most likely to fail is by having the kernel itself reject this attempt.)
The user running telnetd was _not_ root. The telnetd executable was not suid root. It should not have been able to open the port. The acquiring of root permissions was not supposed to happen. This is a local root exploit. The phone's been rooted.
Yes, you should get in the habit of only using SSH. That way when the internet cafe has a keylogger trojan installed you can have your credentials stolen without even realising you might be at risk.
Nonsense. I've got hundreds of trojans on my system. Are you claiming that I've been rooted? A file is just a file. Rooting is privilege escalation. It's a category error to compare them.
Re:I haven't followed the whole Android business,
on
T-Mobile G1 Rooted
·
· Score: 1
Nonsense. Rooting is someone who shouldn't be root gaining root. It can be local root exploit or a remote root exploit, it matters not. What do you claim the popular geek vernacular for a local root exploit is, if it's not rooting?
My google-fu is fed up of the topic, after reaching so many dead ends; OSX introducing new terminology not helping matters at all. It's particularly fatigued given that all of the pro-Mac people I know, who said so many positive things in the past about OSX have, time after time, been unable to provide answers to simple questions. NFS 2.0 not talking to NFS 3.0 on my heterogenous debian network? 2 minutes on the IRC channel, and I've got a solution from whoever happened to be awake at the time. (Hoping that there would be OSX users were, like you, as helpful as the debian community I even went hunting for an OSX help IRC channel, but my google-fu couldn't even find one of those.)
Many, many thanks for that link, I will give that app a spin. However, the existance of it does not counter my original stance, that simple things ain't simple on OSX. Look at Dan Frakes own description, and see how many negative adjectives he uses when describing the underlying process. He's making my point far more eloquently than I can, because he's a Mac insider and expert. That whole page is oozing with "this trivial task is hard and potentially dangerous".
Well, in some ways that's progress. I pulled up NetInfo Manager, and pulled up the particular user whose name I wished to change, which happens to be the admin account. I was presented with the login name as the value in about 7 different fields! Neither MacHelp nor Spotlight could provide assistance on what most of those fields were, and as you mention the help for NetInfo Manager is nothing but 3 pages with bugger all on them (and that link to themselves, gee, that's clever).
MacHelp's pretty useless for everything, in fact. It can't tell me how to not automatically mount USB flash drives; it can't tell me how to only mount a USB flash drive read-only; it can't tell me how to stop Finder from automatically spewing files onto my USB flash drive when I open it. In fact it seems to not know anything about USB flash drives.
So I'm still where I was 2 years ago - merily sitting at my linux (and FreeBSD) boxes, working away knowing that I know how everything I encounter day-to-day works.
Who says that _I_ set up the account whose name I wish to change? I certainly didn't say that, as it would be false. Which is why I'm trying to change it...
On saner unixes, if you're changing UIDs, then make sure you followup with a chown -R on the home directory. (And then a find in case that user stuck stuff in/usr/local,/opt,/tmp, or wherever.) Probably ensuring no processes were running as that user would be prudent while doing this. I don't use any GUI apps in linux that attempt to be 'clever' (most that do end up being the most stupid), so I can't imagine what problems you might have encountered.
Anyway, The "Accounts" panel of System Preferences doesn't have "Advanced Options", and even when I unlock the panel the Short Name field remains greyed out - I can change the long name, but not the short name (and the latter is not simply the first word of the former). So, at least on 10.4, I still can't do what would be a 5 second job in linux.
And yes, I've done web searches, but not knowing the unfamiliar jargon that OSX uses (e.g. "NetInfo", "Short Name" etc.), I've previously found nothing of use. That's why I've asked 4 self-proclaimed mac gurus instead, and even when they looked at their manuals they still couldn't help. I guess that tally's up to 5 now. However, I've done another search, and it has confirmed my fears...... the only way to change username is to create a new user with the intended name, copy home directory contents over, chown them, and then trash the original. Of course - this changes the UID too, which is an unwanted side-effect. All in all, a 40-step process (plus 12 more to enable the root account) on the Apple support site: http://support.apple.com/kb/HT1428 .
And that's why I don't use the mac. Never before has the trivial been so unnecessarily complicated.
Slashdot Mirror
Well cried. Even the knowledgable proponents of GC (Boehm, Zorn) admit there's an overhead to it, but try to dismiss the overhead as negligible, using wording such as "typically competitive" and "only slightly higher". The quick translation of such claims into English is "yup, you got us, they are actually slower and use more memory, sorry".
I'm not sure I really see why the following (from the summary) is an _OR_:
Is Google trying to drive more people towards its own online suite of office applications?
Or
has it been stung into action by Steve Ballmer's recent comment that Microsoft Office faces stronger competition from StarOffice than it does Google Docs and Spreadsheet?"
I'd suggest that it was an _AND_ instead. _AND_ what you said too. The overlap of the 3 is pretty complete.
Loads of support:
1) There's an email address or telephone number where you make your request for support.
2) Someone will answer that within 3, or sometimes 7, working days to indicate they've received your request for support.
What more do you want?
Google have been dicking around with that form in the last week or so. They _removed_ the date range search functionality, for example. They put it back again when they realised they'd been brainless idiots. However, that bug illustrates that 10000 Ph.D.s can still be appear to be brainless idiots. One thing that annoyed me was the fact that they removed the message-id search. Given that message-ids are the PRIMARY KEY of Usenet, that demonstrates google have lost the plot when it comes to Usenet.
Feel free to bookmark, use, and suggest improvements to:
http://theanna.org/goog.html
http://theanna.org/goo.html
(They're both practically the same, one's old-school 90s HTML, the other a little less-so; both will be maintained.)
Having seen some Theroux again after many years without, he also failed to live up to my memories (I remember him visiting the KKK or the US Nazi party what seems like over decades ago, that series I did enjoy). My biggest problem with the supersized NRA member's recent offerings was his horrifically unsubtle selective editing. TV Nation was acceptable, but you can't chop like that for a whole feature-length movie.
Horror of horrors - a Michael Moore documentary about it!
(Don't get me wrong, I agree with many of MM's stances, but his documentary-making is awful.)
Permission requested for using the core of that anecdote as one of my usenet .sigs?
Me.
lrwxrwxrwx 1 root root 15 Aug 22 16:48 initrd.img -> boot/initrd.img
Likewise in Finland. Single-use random 4-digit ids. We've had them for 15 years or more. (So in the early 90s, Finnish banks were more security conscious than most modern-day US or UK banks.)
All my recent "fade to nothingness within a year" CDR's have been "high quality verbatim". I no longer think they deserve the adjective. Better than the old Traxdata gold, but even my own memory's better than those were. What was I talking about, again?
Instead of:
ps -ef | grep -i java | grep -v grep
just do:
ps -ef | grep -i j[a]va
which doesn't match the grep.
No!
dc -e'[[dc -e]P39P91PP93P[dx]P39PAP]dx'
So no-one's _ever_ written
if ((options == (__WCLONE|__WALL)) && (current->uid = 0))
in any kernel source code then?
I think I've got a guess what the problem might be:
/system/bin/telnetd
$ ls -l
-rwxr-xr-x root shell 9752 2008-09-13 01:13 telnetd
At 9752 bytes, I wonder if that's not the whole telnet, I wonder if it execs some other binary that _is_ setuid root.
Does anyone have one of these devices - can they find all setuid files?
Why do you not get teh whole point of this story?
The third option is "something between telnetd and the kernel is broken - non-root users are able to do things that only root users should be able to do".
Or in the context of the earlier either/or question - "neither".
You misread or misunderstood. telnetd is not doing what it is supposed to do. telnetd's supposed to attempt to open a priveleged port (<1024), and supposed to fail if you do not have the required permissions to do so (i.e. be root). (Of course the way it's most likely to fail is by having the kernel itself reject this attempt.)
The user running telnetd was _not_ root. The telnetd executable was not suid root. It should not have been able to open the port. The acquiring of root permissions was not supposed to happen. This is a local root exploit. The phone's been rooted.
Yes, you should get in the habit of only using SSH. That way when the internet cafe has a keylogger trojan installed you can have your credentials stolen without even realising you might be at risk.
Nonsense. I've got hundreds of trojans on my system. Are you claiming that I've been rooted?
A file is just a file. Rooting is privilege escalation. It's a category error to compare them.
Nonsense. Rooting is someone who shouldn't be root gaining root. It can be local root exploit or a remote root exploit, it matters not. What do you claim the popular geek vernacular for a local root exploit is, if it's not rooting?
FTFA:
/system/bin/telnetd
/usr/libexec/telnetd /usr/libexec/telnetd
/usr/libexec/telnetd /usr/libexec/telnetd
/usr/sbin/in.telnetd /usr/sbin/in.telnetd
$ ls -l
-rwxr-xr-x root shell 9752 2008-09-13 01:13 telnetd
See those last 3 characters in the perms? They say "you're wrong".
Android misconfigured? I don't think so:
OSX:
$ ls -al
-r-xr-xr-x 1 root wheel 53368 Sep 19 03:20
FreeBSD:
ls -al
-r-xr-xr-x 1 root wheel 78156 5 Mar 2004
SunOS:
ls -al
-r-xr-xr-x 1 root bin 28108 Mar 4 2003
Desktop Linux:
Fucknose, haven't installed it. Wanna bet against it being 555?
I'll chose to not believe that.
http://www.cityrating.com/citytemperature.asp?City=Chicago
http://www.cityrating.com/citytemperature.asp?City=Anchorage
You can drive a tram between those two curves, every month of the year.
That wasn't a military meeting, that was a press meeting. It was PR.
My google-fu is fed up of the topic, after reaching so many dead ends; OSX introducing new terminology not helping matters at all. It's particularly fatigued given that all of the pro-Mac people I know, who said so many positive things in the past about OSX have, time after time, been unable to provide answers to simple questions. NFS 2.0 not talking to NFS 3.0 on my heterogenous debian network? 2 minutes on the IRC channel, and I've got a solution from whoever happened to be awake at the time. (Hoping that there would be OSX users were, like you, as helpful as the debian community I even went hunting for an OSX help IRC channel, but my google-fu couldn't even find one of those.)
Many, many thanks for that link, I will give that app a spin. However, the existance of it does not counter my original stance, that simple things ain't simple on OSX. Look at Dan Frakes own description, and see how many negative adjectives he uses when describing the underlying process. He's making my point far more eloquently than I can, because he's a Mac insider and expert. That whole page is oozing with "this trivial task is hard and potentially dangerous".
Well, in some ways that's progress. I pulled up NetInfo Manager, and pulled up the particular user whose name I wished to change, which happens to be the admin account. I was presented with the login name as the value in about 7 different fields! Neither MacHelp nor Spotlight could provide assistance on what most of those fields were, and as you mention the help for NetInfo Manager is nothing but 3 pages with bugger all on them (and that link to themselves, gee, that's clever).
MacHelp's pretty useless for everything, in fact. It can't tell me how to not automatically mount USB flash drives; it can't tell me how to only mount a USB flash drive read-only; it can't tell me how to stop Finder from automatically spewing files onto my USB flash drive when I open it. In fact it seems to not know anything about USB flash drives.
So I'm still where I was 2 years ago - merily sitting at my linux (and FreeBSD) boxes, working away knowing that I know how everything I encounter day-to-day works.
Who says that _I_ set up the account whose name I wish to change? I certainly didn't say that, as it would be false. Which is why I'm trying to change it...
/usr/local, /opt, /tmp, or wherever.) Probably ensuring no processes were running as that user would be prudent while doing this. I don't use any GUI apps in linux that attempt to be 'clever' (most that do end up being the most stupid), so I can't imagine what problems you might have encountered.
... the only way to change username is to create a new user with the intended name, copy home directory contents over, chown them, and then trash the original. Of course - this changes the UID too, which is an unwanted side-effect. All in all, a 40-step process (plus 12 more to enable the root account) on the Apple support site: http://support.apple.com/kb/HT1428 .
On saner unixes, if you're changing UIDs, then make sure you followup with a chown -R on the home directory. (And then a find in case that user stuck stuff in
Anyway, The "Accounts" panel of System Preferences doesn't have "Advanced Options", and even when I unlock the panel the Short Name field remains greyed out - I can change the long name, but not the short name (and the latter is not simply the first word of the former). So, at least on 10.4, I still can't do what would be a 5 second job in linux.
And yes, I've done web searches, but not knowing the unfamiliar jargon that OSX uses (e.g. "NetInfo", "Short Name" etc.), I've previously found nothing of use. That's why I've asked 4 self-proclaimed mac gurus instead, and even when they looked at their manuals they still couldn't help. I guess that tally's up to 5 now.
However, I've done another search, and it has confirmed my fears...
And that's why I don't use the mac. Never before has the trivial been so unnecessarily complicated.