Linus Torvalds On GCC 4.9: Pure & Utter Crap

[mlkj]

Back to Yotsuba with me

Wow, stop right there - when i tell people the truth, these basterds just ban me, but when some old men tells the same thing, its cool ?????

Welcome to every internet forum ever.

[bison]

Wow, stop right there - when i tell people the truth, these basterds just ban me, but when some old men tells the same thing, its cool ?????

The running definition of "old man" is usually the age of your dad. So of Linus is an old man, my guess is that you must be somewhere between 14 and 22.

[Guest]

Torvalds went on to say, "Note the contents of -136(%rbp). Seriously. That's an _immediate_constant_ that the compiler is spilling. Somebody needs to raise that as a gcc bug. Because it damn well is some seriously crazy shit."

Yeah and you know what they say on the forums. If you got a problem "Fix it yourself!"

Hurts the worst when it hits home.

Somebody is you buddy!

I personally like the 'fix it yourself' approach, but dependending on type of bug that isn't always trivial.
On the other hand, when Linus Torvalds reports a bug, it gets instantly fixed and backported everywhere.

Back to 'fix it yourself', there is also "don't touch my code" - people more often than not will not accept your patches,
unless you have 'street credit' with the project.

[bridgman]

The running definition of "old man" is usually the age of your dad. So of Linus is an old man, my guess is that you must be somewhere between 14 and 22.

Huh, that's changed. When I was a kid the working definition of an "old man" was the age of your grandfather...

[sdack]

I think we have to take linus hate words as love words... what is the project he got most hate in the past, at least recent years, yes gnome(shell).

What do we see on his last workplace video on his desktop u can guess 1 time

Maybe we should not wonder that he talkes shit like a random guy sitting in the basement, looking at his homeoffice it looks cracy "normal" and small. his creation powers more systems than windows but he works and lives like my dad who was a insurance agent. I think even my dads homeoffice (he worked self-employed) is bigger than linus work place.

Not meaning to argue with you. I just want to say that it is not really his creation any more. Very little code is his own. He only gets to choose what goes into the kernel and what does not. He often does not have a choice, i.e. when a patch adds support for new hardware or fixes a bug.

Every kernel developer trusts him to do the right thing and only because of this trust do large companies like Intel, IBM, AMD, ARM, and many more, contribute to the kernel code. As long as somebody signs off their patches will these companies care little about him or miss him when he leaves.

[gens]

to explain a detail, if any one misunderstood

movq $load_balance_mask, -136(%rbp) #, %sfp
...
movq -136(%rbp), %rax # %sfp, tcp_ptr__

first one puts an intermediate value onto the stack
second one loads it from the stack into a register
intermediate values are in the machine code itself
stack is in main memory (ram, cache)

this should be just

movq $load_balance_mask, %rax

since load_balance_mask is constant

[gens]

Not meaning to argue with you. I just want to say that it is not really his creation any more. Very little code is his own. He only gets to choose what goes into the kernel and what does not. He often does not have a choice, i.e. when a patch adds support for new hardware or fixes a bug.

Linus has every right to reject any patch
hes choice is to let patches in and that does not mean much of anything really

[sdack]

Linus has every right to reject any patch

This is true for companies where managers are given absolute powers, but not for open source communities. The moment that a leader of an open source project rejects a reasonable patch is also the moment where he distances himself from the project. In a company will you always be a part of the company, because of the contract you have signed and which defines who is a member of the company and regardless of your contributions to the projects within the company. An open source project defines itself by the respect of the developers for each other and an open source license. There is however no binding contract, no special law and therefore is there no right of any kind other than the rights of the country one is living in of course. Open source is a free software development and with it free of rights.

Linus does not have a right to reject or to accept a patch. All he has got is the choice of any human to be part of a free software development. The moment he chooses to reject reasonable patches does not mean he is exercising his rights. He would only choose to reject contributions to a free project, which eventually can only lead to his departure from the project, because others will include these patches into the source, given these are reasonable, and carry on without him.

Anyhow, not saying he needs to leave, because of another rage of his. Just killing your idols ...

[cb88]

Pretty sure my grandpa's definition of old man is around 95-100+

Also if you don't have any respect for older people that know more than you likely ever will your brain is worse off than the gcc compiler in question :P

But, then maybe at 26 I'm already too old for you to care what I think! Nice flameware btw.. glad to help...

[jrch2k8]

This is true for companies where managers are given absolute powers, but not for open source communities. The moment that a leader of an open source project rejects a reasonable patch is also the moment where he distances himself from the project. In a company will you always be a part of the company, because of the contract you have signed and which defines who is a member of the company and regardless of your contributions to the projects within the company. An open source project defines itself by the respect of the developers for each other and an open source license. There is however no binding contract, no special law and therefore is there no right of any kind other than the rights of the country one is living in of course. Open source is a free software development and with it free of rights.

Linus does not have a right to reject or to accept a patch. All he has got is the choice of any human to be part of a free software development. The moment he chooses to reject reasonable patches does not mean he is exercising his rights. He would only choose to reject contributions to a free project, which eventually can only lead to his departure from the project, because others will include these patches into the source, given these are reasonable, and carry on without him.

Anyhow, not saying he needs to leave, because of another rage of his. Just killing your idols ...

yet the kernel is one of the most active projects ever and in constant grow, additionally linus is not some random dude rejecting patches because he thinks you are too fat, he is a one of the greatest low level coders around(this days mostly manage tho) and when he rejects patches there is always a pretty damn good reason for it and probably a good story behind it(normally the context get "missed" in the headlines tho), don't believe me go and ask Greg hartmann.

your concept of free software and patches is plainly wrong, go to FSF and read i won't waste my time explaining the obvious.

Linus is not the head of the linux kernel project because of some sort of dictatorship conspiration theory or because he have some upper godly right for its creation, is for one very simple and straightforward reason, he is really damn good at it and every member and developer in the kernel arena respect his vast amount of knowledge and take very very seriously any comment or reason from him(rant or not) and the same can be said from greg, andrew, etc.(normally their rants don't reach headlines tho) and without people like them the kernel would be a mess of half done patches(again ask greg and andrew).

Next time instead of just barf things you interpret from a headline please go to LKML and get the context(assuming you can understand the issue, sometimes they discuss in ASM and is really hard to follow) and you will see his rant are 99% entirely justified.

In this case in particular is not the first issue with GCC letting pass ASM errors that give hell to kernel developers and well this wasn't the worst case either but i guess linus got fed up and exploded and if you had to debug it, im pretty sure you will be sending bomb letters to the GCC developers in an epic scale emo rage(as anyone will)

anyway same old same old, linus ranted then got rational again then developers rushed from everywhere and fix the issue with him then everyone is back to work again but a bit more careful this time and everyone back to their work