Cover photo
Verified name
Linus Torvalds
Works at Linux Foundation
Attended University of Helsinki
Lives in Portland, OR


Linus Torvalds

Shared publicly  - 
My email spam situation seems to be all good.

But just to annoy me, we've had two calls today from the IRS phone scam (google "IRS phone scam" to get some news on it).

I wonder why those things are so hard to crack down on.  Phone companies really don't even check ID when you get a phone number? They sure wanted to get my credit history when I opened a line, but some low-life scammers who try to fool people can get one?

Anyway, if anybody wants to talk with the "federal agent" about their tax situation and how you want to give them money over the phone, the number is (276) 258-6019. 

Scammers like that piss me off. 

[ Edit: the number is apparently disconnected now. So don't call it - at some point that number will get re-used, and some poor sap will wonder why he gets all these people offering him their credit card numbers ]
Keshab Sethy's profile photoJason Williams's profile photoMarwan Burelle's profile photoEric De La Cruz Lugo's profile photo
Add a comment...

Linus Torvalds

Shared publicly  - 
Much better now.

Of the 100+ messages caught as spam over-night, only two were false positives (and I reported them). My email is getting back to normal.
Alessandro Di Giacomantonio's profile photoDipesh Kumar's profile photoSJ Waters Fitzgerald's profile photoANTONIO MANUEL MEDINA MILTIMORE's profile photo
Hmmm, normal email. What's that? :P
Add a comment...

Linus Torvalds

Shared publicly  - 
Dear Google Mail Team,

  I've said very nice things about your spam filter in the past, but I'm afraid I am going to have to take it all back. I'm currently going through the spam for the last week, and have gone through about a third of it.

Something you did recently has been an unmitigated disaster. Of the roughly 1000 spam threads I've gone through so far, right now 228 threads were incorrectly marked as spam.

That's not the 0.1% false positive rate you tried to make such a big deal about last week. No. That's over 20% of my spambox being real emails with patches and pull requests.  Almost a quarter!

I don't know how to even describe the level of brokenness in those kinds of spam numbers. There were a few pages of email (I've got it set up so it shows me 50 threads per page) where more than half of the "spam" wasn't.

Quite frankly, that sucks. It's not acceptable. Whatever you started doing a few days ago is completely and utterly broken.

It's actually at the point where I'm noticing missing messages in the email conversations I see, because gmail has been marking emails in the middle of the conversation as spam. Things that people replied to and that contained patches and problem descriptions.

They didn't try to sell me a bigger penis or tell me about how somebody is cheating on me. Really.

You dun goofed. Badly. Get your shit together, because a 20% error rate for spam detection is making your spam filter useless.

[ Edit: looks like it started four days ago. As of July 13, it looks like a big swath of lkml has been marked as spam for me. ]

[ Edit 2: final numbers: out of around 3000 spam threads, I had to mark 1190 threads as "not spam". So the numbers actually got worse: about 30% of my spam-box wasn't actually spam. It started around 1pm on Monday, July 13th. The problem really is that clear, that I can tell pretty much when it started ]

[ Edit 3: it wasn't just patches, and it's not just lkml. There were things like Junio's recent git v2.5.0-rc2 announcement etc. The new gmail spam filter hates any mailing list emails, apparently. In the time I wrote the last note, I got seven more emails marked as spam, two of which weren't. ]
Amy Leal's profile photoMario Gundić's profile photohussein elalamy (‫العلمي‬‎)'s profile photoBernhard Breytenbach's profile photo
+Dave Neary
although NLP == difficult.  That said, if the 'understanding' part of NLP were easy(ier), then a lot of the backenedness needed by traditional spam detection engines wouldn't be needed.  I imagine that the 'getting the understanding part' rather better is mainly why Google has an email offering, i.e., it's not purely about targetting ads.  Get that right, and the world of search comes with it.
Add a comment...

Linus Torvalds

Shared publicly  - 
The git people had their official 10-year anniversary earlier in the year, but today is another anniversary for the project.

Today is exactly ten years since I asked +Junio C Hamano to be the +Git maintainer. And what a maintainer he has been.
Diego Salinas's profile photoDaniel Reynaga's profile photoZenia Zen's profile photoMurphy Tsai's profile photo
And ten-years ago yesterday, Jul 25, was when you heard my "Yeah, I'll do that", which was buried in the noise and sitting in your mailbox for about a week ;-)
Add a comment...

Linus Torvalds

Shared publicly  - 
Ugh. I despise systems that allow others to register your email for anything.

GitHub - today I'm looking at you. 

For some reason (probably just courtesy) people end up giving me write access to their github repositories. These people probably don't even realize that that makes GitHub spam me with their repo data, their commit notifications etc. I assume they do it as a gesture of "hey, I trust Linus with my repo", knowing full well that I won't actually do anything, but doing it as a funny courtesy.  I can't really blame them for the gesture.

But I can blame GitHub for turning that gesture into a source of spam.

Sure, the emails are easy to delete (and I do - sorry guys, I get too much email as-is), but it's an example of a deep flaw in the system. Letting others sign you up for things is really very annoying.

I shouldn't need to unsubscribe to things I don't need filling up my inbox: you should check at registration point whether the email is willing and able to be contacted.

I'd much rather get  the occasional email saying "So-and-so has invited you to watch his repo xyzzy, please follow this link to activate, or ignore this email if you aren't interested". That kind of verification email I do not mind at all, because it shows that the service understands about this whole opt-in and email verification issue. It would also just be a once-per-repo thing.

Of course, GitHub is by no means the only offender. Lots of places are perfectly fine taking any random email address and registering that email address to their mailing lists etc, with no verification of "do you actually own that email" at any point.  

I generally mark all such email as spam. Because if you have a mailing list that allows others to sign up as me, you aren't a mailing list - you're an internet menace and a spam provider. If you are running a mailing list, please check that your subscription process has that kind of verification stage. Ok?
John Partain's profile photoNeo Smith's profile photoNemanja Maksimovic's profile photoJim Robinson's profile photo
+Linus Torvalds and especially this thing . Truly I had a good laugh while reading this.
Add a comment...

Linus Torvalds

Shared publicly  - 
I really like the updates on what went wrong.  With updates like this, even the failures are entertaining.

[ Side note: and how they name their barges by Iain M Banks culture ship names.. If you haven't read the novels, you're missing out. ]
Some of you may have been following our recent attempts to vertically land the first stage of our Falcon 9 rocket back on Earth.
Aamir Malik's profile photoNiranjan Patil's profile photoJosé Miguel Sarmiento Palma's profile photoPaul M Edwards's profile photo
+Rani Ahmad Even though, as far as I know, Linus tries to stay as distant as possible from the desktopwars, I do agree something has to be done, and maybe it's "intervention" time for the desktop teams that are creating the disastrous tablet alike desktops for non tablet devices.
Add a comment...

Linus Torvalds

Shared publicly  - 
I'm not a fan of traditional AI (rule building and LISP/prolog etc), but am just waiting for neural nets to take over. 

This is a great example, and training on the Linux kernel sources just explains so much.

This is wonderful; neural network setup producing wonderful results; one example is training it on wikipedia and it ends up producing valid XML pseudo articles.  Another is training it on Linux kernel source, and it ends up producing C code that looks appealingly complex until you try and figure it out (even with bogus comments).
7 comments on original post
fleuria lee's profile photoDmitry Serebrennikov's profile photoPaolo Cavazza's profile photoAnusorn Chaikaew's profile photo
+Robert Reppel  this is the article I was mentioning yesterday.
Add a comment...

Linus Torvalds

Shared publicly  - 
I don't have the hi-res in-the-air pictures yet, but here's me looking cool.

Although apparently +Jim Zemlin was laughing so hard when I did the thumbs-up sign that the camera shake makes that picture a bit blurry. Thanks, Jim.
Rob Fielding's profile photoBruno Aiub's profile photoVitaliy Pakshaev's profile photoPiotr Golonka's profile photo
+Linus Torvalds Are you a pilot?
Add a comment...

Linus Torvalds

Shared publicly  - 
Not many people know this about me, but in between kernel pulls I relax by doing fighter jet rides.
ranvir jogi's profile photoCaio Peixoto's profile photoBernard Halas's profile photoBruno Aiub's profile photo
Add a comment...

Linus Torvalds

Shared publicly  - 
Not even halfway through the merge window, but I suspect that by the end of day today I will have merged more commits for 4.2 than we had during all of the 4.1 release.

4.0 was pretty small (by our modern standards, which is to say it was still a lot of changes), 4.1 was about average, and it looks like 4.2 may end up being the biggest release (in number of commits, at least) we've ever had.

So much for the summer slowdown (and no, it's not the Aussies and other southern hemisphere people picking up the slack).

[ Update: I don't think I'll beat the 4.1 numbers today. Partly because Greg's pull requests were smaller than I expected, but partly because I'm slowing down my merges because I think I just hit a jbd2 bug introduced in this merge window ]
Thomas Fischer's profile photovis wu (visual)'s profile photoGeorge Tonelli's profile photoVictor Mazariego's profile photo
A lot of people are making a lot of noise about 4.2 rc1 on my stream today. Do you think they have the stomach to handle that much code in a release candidate? They're getting nervous … :0)
Add a comment...

Linus Torvalds

Shared publicly  - 
So Google photos seems to want to make odd videos of the random movie clips I uploaded from last week.

And apparently, with dramatic music, some color tinting, and by making the cuts be frequent enough, you can make even my blurry fish butt videos entertaining.

There is a shark in there. And Daniela, who got certified last year and did very well as a dive buddy. But the real star is definitely Google photos.
Alexandro Colorado's profile photojohnson anthraper's profile photoSimos Katsiaris's profile photoSuresh Patel's profile photo
Damn. I love this google photo feature. 
Add a comment...

Linus Torvalds

Shared publicly  - 
I may have mentioned before that C++ isn't my favorite language. But for various reasons, it's what subsurface is written in these days. And it seems to trigger an annoying gdb bug.

Are there any gdb people out here that have a solution to this "feature" of gdb:

 - a C++ program that uses idiomatic C (because it started out that way):

    struct test test;

 - compile as C++, run under debugger, try to show the value of 'test':

    (gdb) p test
    Attempt to use a type name as an expression

Yeah, gdb, you're badly confused. I want the variable test, not the type test. And yes, it's confusion caused by silly C++ "improvements"  to C syntax, but still..

How do I disambiguate this to gdb?
Rush Damage's profile photoShashishekhar S's profile photoMax Reig's profile photoТарас Шевченко's profile photo
I love python.
Add a comment...
Creator of Linux and git
  • University of Helsinki
Basic Information
SW Engineer
  • Linux Foundation
    SW Engineer, present
  • OSDL
  • Transmeta Corp
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Portland, OR
Helsinki, Finland - Bay Area, CA
I've been to Fish 'n fins twice now, once during off-season (June), and once during high season (November). Very professional. They clearly get very crowded during high season, but seem to successfully juggle having multiple boats out. Good diving, good people.
Public - a year ago
reviewed a year ago
1 review