Recently posted comments

Previous 90 comments.

GNOME 3.32 released

Posted Mar 16, 2019 16:56 UTC (Sat) by DOT (subscriber, #58786)
In reply to: GNOME 3.32 released by sbakker
Parent article: GNOME 3.32 released

Why does the session depend on the Shell, actually? Couldn't it simply restart and continue compositing?

The one thing I miss from X is the ability to reconfigure monitors programmatically with a tool like xrandr.

Downloading packages

Posted Mar 16, 2019 14:25 UTC (Sat) by ms (subscriber, #41272)
In reply to: Downloading packages by corbet
Parent article: Federated blogging with WriteFreely

No problem. Glad I could help!

Downloading packages

Posted Mar 16, 2019 14:23 UTC (Sat) by corbet (editor, #1)
In reply to: Downloading packages by ms
Parent article: Federated blogging with WriteFreely

OK, clearly I'm pretty ignorant about how Go packaging works, and it's not as bad as I had feared. Thanks for enlightening me.

Downloading packages

Posted Mar 16, 2019 14:12 UTC (Sat) by ms (subscriber, #41272)
In reply to: Downloading packages by corbet
Parent article: Federated blogging with WriteFreely

"go get" inspects the go.mod and go.sum files in the root of the project.

https://github.com/writeas/writefreely/blob/master/go.mod
https://github.com/writeas/writefreely/blob/master/go.sum

There's been an awful lot of thinking that has gone into this mechanism, and whilst it's not quite finished in Go 1.12, it's pretty close these days. There's a lot of writing about it all at https://research.swtch.com/vgo

Announcing the release of sway 1.0

Posted Mar 16, 2019 14:09 UTC (Sat) by mathewcohle (guest, #118622)
In reply to: Announcing the release of sway 1.0 by jdulaney
Parent article: Announcing the release of sway 1.0

I'm using https://github.com/bugaevc/wl-clipboard as xclip replacement - haven't found any problem yet.

Announcing the release of sway 1.0

Posted Mar 16, 2019 14:07 UTC (Sat) by mathewcohle (guest, #118622)
Parent article: Announcing the release of sway 1.0

Works surprisingly well so far! Glad to have reason to switch to Wayland.

Downloading packages

Posted Mar 16, 2019 13:56 UTC (Sat) by corbet (editor, #1)
In reply to: Federated blogging with WriteFreely by ms
Parent article: Federated blogging with WriteFreely

From my reading, it's using a bunch of "go get" commands, with no version pinning in sight anywhere. Are those version strings hidden somewhere that I'm unaware of?

Haller: WireGuard in NetworkManager

Posted Mar 16, 2019 11:43 UTC (Sat) by georgm (subscriber, #19574)
Parent article: Haller: WireGuard in NetworkManager

There hasn't been a request for inclusion of wireguard into the kernel in the last couple of months. What is the state here?

Static site generators

Posted Mar 16, 2019 11:05 UTC (Sat) by smurf (subscriber, #17840)
In reply to: Static site generators by andrewsh
Parent article: Federated blogging with WriteFreely

So is Nikola.

Some of the static-site systems even have a Web editor, which sounds strange but happens to be quite convenient.

Five new stable kernels

Posted Mar 16, 2019 11:05 UTC (Sat) by jg71 (guest, #67102)
In reply to: Five new stable kernels by alanjwylie
Parent article: Five new stable kernels

I was about to report it myself (avid user of 4.19), so thanks are in order to Alec Ari for already doing so.

Federated blogging with WriteFreely

Posted Mar 16, 2019 10:32 UTC (Sat) by ms (subscriber, #41272)
Parent article: Federated blogging with WriteFreely

Re building and "downloading several dozen packages". The project is at least using go.mod and so the downloads are fully pinned to git commits/tags and checksummed. So it's not quite the wild west of "let's just download the latest version of foo and hope for the best". Personally, I much prefer pointers to vendoring all your deps, but I know opinions differ on that.

They're using the built-in Go webserver which as far as I know has an excellent security record - I would have no worries about exposing it directly to the internet; I have used it a lot commercially/professionally.

The template language *is* documented. It's the standard Go templating system. https://golang.org/pkg/text/template/ https://golang.org/pkg/html/template/
That said, I agree about documenting which resources/methods/vars are available to the templates. The template entry point is Execute or ExecuteTemplate, and the last arg to that is always the "receiver" for the template, but eg https://github.com/writeas/writefreely/blob/cb1bd37f64abc... is unhelpful given it comes in as the empty interface. Generally I consider use of the empty interface in Go as a code smell... https://github.com/writeas/writefreely/blob/32e99d00415c6... is somewhat more discoverable. OTOH, depends on the intended audience - I've been working professionally in Go for over 5 years now so I have no qualms digging through this.

Just FTR, I've no association with WriteFreely and have never heard of it or looked at the code base before this morning.

Fixing programmers

Posted Mar 16, 2019 10:30 UTC (Sat) by roc (subscriber, #30627)
In reply to: Fixing programmers by HelloWorld
Parent article: Cook: security things in Linux v5.0

Many pots these days have coatings that burn off, ruining the pot, if it overheats.

Fixing programmers

Posted Mar 16, 2019 9:29 UTC (Sat) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by martinfick
Parent article: Cook: security things in Linux v5.0

> Unfortunately, I have yet to see a stove that comes with a timeout on every burner and doesn't continue to burn your food, your pot, and potentially your house if you forget to turn it off.
Pots are made of metal and hence won't burn, and if the food burns, there will be smoke which will trigger the smoke detector, which is mandatory where I live.

Fixing programmers

Posted Mar 16, 2019 9:25 UTC (Sat) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by scientes
Parent article: Cook: security things in Linux v5.0

How does this break distcc?

Fixing programmers

Posted Mar 16, 2019 9:22 UTC (Sat) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by jrigg
Parent article: Cook: security things in Linux v5.0

> This could be considered a feature.
Yes, by other stick-in-the-muds.

Federated blogging with WriteFreely

Posted Mar 16, 2019 9:17 UTC (Sat) by xanni (subscriber, #361)
Parent article: Federated blogging with WriteFreely

I tried to subscribe to either your RSS feed or your ActivityPub feed with Hubzilla, but unfortunately neither worked.

Static site generators

Posted Mar 16, 2019 8:37 UTC (Sat) by pbonzini (subscriber, #60935)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

The QEMU website uses Jekyll, which also powers GitHub pages. One advantage for us was that contributions are done with the usual mailing list workflow.

A couple years ago I wrote a tutorial at https://opensource.com/article/17/4/getting-started-jekyll based on my experience building qemu.org.

The creation of the io.latency block I/O controller

Posted Mar 16, 2019 7:35 UTC (Sat) by josefbacik (subscriber, #90083)
In reply to: The creation of the io.latency block I/O controller by martin.langhoff
Parent article: The creation of the io.latency block I/O controller

The 1s timer is only armed while there’s IO happening. No IO, no periodic timer.

SUSE completes its management transition

Posted Mar 16, 2019 5:38 UTC (Sat) by alison (subscriber, #63752)
Parent article: SUSE completes its management transition

I was impressed by the quality of the talks in the OpenSUSE track at Southern California Linux Expo last weekend.

https://www.socallinuxexpo.org/scale/17x/schedule/friday

I've never used OpenSUSE beyond Open Build System, but the quality of technical contribution there obviously remains high.

Security quotes of the week

Posted Mar 16, 2019 5:37 UTC (Sat) by sasha (guest, #16070)
Parent article: Security quotes of the week

I guess I am not the only Russian reader here, so here is the link to the Russian text from the Medusa: https://meduza.io/feature/2019/03/01/k-mobilnomu-izbirate...
It took some time from me to find it out...

submitted

Posted Mar 16, 2019 5:16 UTC (Sat) by scientes (guest, #83068)
In reply to: Cook: security things in Linux v5.0 by jani
Parent article: Cook: security things in Linux v5.0

> Granted, I would have preferred __attribute__ ((fallthrough)) with some syntactic sugar macro.

https://lkml.org/lkml/2019/3/15/792

Because it is needed by distcc.

Fixing programmers

Posted Mar 16, 2019 5:09 UTC (Sat) by scientes (guest, #83068)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

It is also a bad idea to parse comments. It breaks distcc for example, so I submitted a patch fixing this stupidness

https://lkml.org/lkml/2019/3/15/792

Democracy...

Posted Mar 16, 2019 2:51 UTC (Sat) by Garak (guest, #99377)
In reply to: Democracy... by Garak
Parent article: Security quotes of the week

Or rather, the unattributed cliche I was grasping for was perhaps- Democracy- It will eventually arrive at the optimal solution, after exhausting all the alternatives.

The creation of the io.latency block I/O controller

Posted Mar 16, 2019 2:49 UTC (Sat) by martin.langhoff (guest, #61417)
Parent article: The creation of the io.latency block I/O controller

How does the 1s tick play with power management? The progress towards tickless systems makes for visibly better battery runtime, and better VM behavior on the server side...

Static site generators

Posted Mar 16, 2019 2:38 UTC (Sat) by zlynx (guest, #2285)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

Static sites can have comments. You just arrange things to generate a new static page which includes the comment after a comment is posted.

That takes the same amount of time as generating a dynamic page and can be reused by every visitor until a new comment is added.

Democracy...

Posted Mar 16, 2019 2:22 UTC (Sat) by Garak (guest, #99377)
Parent article: Security quotes of the week

The problem is that almost no policymakers are discussing this policy issue from a technologically informed perspective, and very few technologists truly understand the policy contours of the debate. The result is both sides consistently talking past each other, and policy proposals -- that occasionally become law -- that are technological disasters.

Democracy. Gotta Love It. It doesn't get you a great solution to every problem, or even a good one. Just better than all the alternatives.

standard plug and play security convenience tradeoff thing

Posted Mar 16, 2019 1:51 UTC (Sat) by Garak (guest, #99377)
In reply to: The Thunderclap vulnerabilities by nivedita76
Parent article: The Thunderclap vulnerabilities

Not for at least 5 years,-

https://arstechnica.com/information-technology/2013/12/inside-the-nsas-leaked-catalog-of-surveillance-magic/

the NSA has built and deployed its own USB cables at target locations—complete with spy hardware and radio transceiver packed inside."

I also recall a few articles from those key early Snowden days where one of the oddly louder complaints from the government about Snowden's leak related to them having to replace a large number of cables.

The COTTONMOUTH series of implants are USB devices that provide a covert wireless bridge into a target network. They can be integrated into any USB plug, so check your mouse.

The Linux Foundation's CommunityBridge platform

Posted Mar 16, 2019 1:29 UTC (Sat) by em-bee (guest, #117037)
In reply to: The Linux Foundation's CommunityBridge platform by keithp
Parent article: The Linux Foundation's CommunityBridge platform

"prefer to use single-point-of-failure, proprietary software rather than FOSS for their infrastructure" sounds a bit snarky ;-)

greetings, eMBee.

Jekyll as a static site generator for blogging

Posted Mar 16, 2019 1:01 UTC (Sat) by michaelkjohnson (subscriber, #41438)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

I recently moved to using jekyll as the static site generator for my blog — I just git commit and push, and shortly thereafter it's updated. I like being able to track my blog history in Git. I've been thinking about running Git on my android phone to push blog entries when I'm away from my computer, but somehow it doesn't seem sufficiently important. ☺

Static site generators

Posted Mar 16, 2019 0:31 UTC (Sat) by ay (guest, #79347)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

You can still have comments in static site generators by hooking up to Discourse. That's what many people do with GitHub pages or gitlab hosting.

A static site generator makes a lot of sense to me since it's easy to host (even for free) and nothing of value is lost versus a CMS type system.

Development quotes of the week

Posted Mar 16, 2019 0:20 UTC (Sat) by rahulsundaram (subscriber, #21946)
In reply to: Development quotes of the week by flussence
Parent article: Development quotes of the week

I noticed that Pulp (which is the open source content management component) of Red Hat Satellite simply moved on to using PostgreSQL instead of MongoDB + PostgreSQL and Red Hat has dropped MongoDB completely. No forks seem to have emerged. It may very well be that the license change would get more people to look at traditional databases which have grown support for NoSQL type features. That is probably a good thing.

Development quotes of the week

Posted Mar 16, 2019 0:12 UTC (Sat) by flussence (guest, #85566)
In reply to: Development quotes of the week by mjthayer
Parent article: Development quotes of the week

PostgreSQL's native JSON support has pretty decent performance (a lot of ActivityPub server software relies on it, for example). In the face of that competition MongoDB didn't have a whole lot going for it even before the SSPL.

Python dictionary "addition" and "subtraction"

Posted Mar 15, 2019 23:56 UTC (Fri) by quietbritishjim (subscriber, #114117)
In reply to: Python dictionary "addition" and "subtraction" by jani
Parent article: Python dictionary "addition" and "subtraction"

The values in Counter don't need to be integers, or even numbers. Both the constructor and the update method can take either iterables or mappings, and the iterable versions put integer counts in while the mapping versions just directly use the values (with the + operator, in the case of update). There's a note in the documentation [1] (scroll down to grey box) saying the implementation of each method makes minimal assumptions on the value type.

[1] https://docs.python.org/3/library/collections.html#collec...

Fixing programmers

Posted Mar 15, 2019 23:15 UTC (Fri) by mjg59 (subscriber, #23239)
In reply to: Fixing programmers by madscientist
Parent article: Cook: security things in Linux v5.0

C# blocks implicit fallthrough and requires an explicit goto statement referencing another switch label. goto might be considered harmful in general, but its use there seems consistent with the way it's generally used in the kernel.

Static site generators

Posted Mar 15, 2019 22:43 UTC (Fri) by me@jasonclinton.com (subscriber, #52701)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

I use https://github.com/getzola/zola .

Fixing programmers

Posted Mar 15, 2019 22:42 UTC (Fri) by madscientist (subscriber, #16861)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

> They make clear that someone was using lint (or similar). Not particularly interesting and
> technically useless as the default case is - well - the default case. Whether or not this is
> correct is a different conversation, but that's something which can only be determined by
> looking at the code. There's no more reason to assume that two cases with fallthrough
> are correct or incorrect than two cases seprated by a break;. I've actually written code
> where I forgot to omit the break in the past :-).

I'm not going to get drawn too far into this conversation because it's ridiculous, but I'll say that the above is at best missing the point.

A fallthrough comment is not trying to tell you whether or not the code will fall through. Obviously it will, there's no break! The comment is telling you that someone thought about it and INTENDED the fallthrough to happen.

If lint or the compiler warns someone about a missing break and they want a break, they'll obviously just add a break. If they get a warning and they intended to fallthrough, they will add a fallthrough comment. If they didn't add either one then you don't know whether they thought about it and intended the fallthrough, or they just forgot.

Of course, even with the comment the fallthrough behavior could be a bug. Code has bugs. But the comment is not telling you the code is bug-free, it's telling you that it's intentional, rather than an omission. That's still a big benefit.

Static site generators

Posted Mar 15, 2019 22:42 UTC (Fri) by andrewsh (subscriber, #71043)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

I use Pelican, it’s written in Python, which is (for me) a huge benefit over Jekyll and its derivatives, since it’s written in a programming language I can actually understand and write in.

Static site generators

Posted Mar 15, 2019 22:35 UTC (Fri) by rahulsundaram (subscriber, #21946)
In reply to: Static site generators by corbet
Parent article: Federated blogging with WriteFreely

Nikola, Hugo etc seem to do everything that you appear to care about in WriteFreely. They are also far more healthier projects with a larger set of contributors and users. They are not federated and there is no activitypub support etc but it looks like all you need is a static site anyway.

Static site generators

Posted Mar 15, 2019 22:29 UTC (Fri) by corbet (editor, #1)
In reply to: Federated blogging with WriteFreely by me@jasonclinton.com
Parent article: Federated blogging with WriteFreely

I've not looked at static site generators, probably should. Can you name the one you're using?

The lack of comments leaves me rather less than upset in general...the last think I need is another comment area to manage!

Federated blogging with WriteFreely

Posted Mar 15, 2019 22:20 UTC (Fri) by me@jasonclinton.com (subscriber, #52701)
Parent article: Federated blogging with WriteFreely

Have you considered a static site generator? They seem to be all the rage and I'm quite happy with the one that I'm using. Most of them allow authorship in Markdown and the images that are embedded in the posts are uploaded with the rest of the static content so there is no image hosting problem, per se. No comment section, of course. But some offer integration with Discuss or other options. Or, since the comment sections on the internet are always difficult to manage, you can avoid tcomment section entirely.

Fixing programmers

Posted Mar 15, 2019 21:44 UTC (Fri) by rweikusat2 (subscriber, #117920)
In reply to: Fixing programmers by madscientist
Parent article: Cook: security things in Linux v5.0

> But that's not the language we have, so FALLTHROUGH-type comments that are recognized by the compiler are great things and
> everyone should use them. Even if they're weren't recognized by the compiler, they're recognized by the next programmer to come > along and read the code, and make clear that you didn't forget something, so they should still always be used.

They make clear that someone was using lint (or similar). Not particularly interesting and technically useless as the default case is - well - the default case. Whether or not this is correct is a different conversation, but that's something which can only be determined by looking at the code. There's no more reason to assume that two cases with fallthrough are correct or incorrect than two cases seprated by a break;. I've actually written code where I forgot to omit the break in the past :-).

Fixing programmers

Posted Mar 15, 2019 21:36 UTC (Fri) by madscientist (subscriber, #16861)
In reply to: Fixing programmers by myxie
Parent article: Cook: security things in Linux v5.0

Continue and return work the same inside a switch statement as outside a switch statement. They are not special.

The main problem is K&R chose the wrong "default" behavior. Usually you DON'T want to continue, so that should have been the default (that is, the current case is automatically terminated by the next case or the end of the statement), and there should be a special keyword used only when you DO want to fall through.

The one exception is empty cases ("case a: case b: case c: foo(); break;") A special dispensation could have been made that the keyword is not needed to fall through unless there are other statements in the case: empty cases don't need the keyword.

Of course, you need a different keyword. They could have used "continue", but both "break" and "continue" are already useful from within switch statements (to break out of or continue an enclosing loop) so using either one of them as meaningful inside a switch was (IMO) not a good decision. I've definitely run into situations where I wanted to break out of a loop from within a switch statement and having "break" overloaded is annoying.

A "fallthrough" or "nobreak" keyword would have been fine. Or if you REALLY didn't want to add a keyword you could have played tricks with "goto", maybe, allowing "goto case X;" where "case X:" existed somewhere in the current switch statement, plus "goto default;"

But that's not the language we have, so FALLTHROUGH-type comments that are recognized by the compiler are great things and everyone should use them. Even if they're weren't recognized by the compiler, they're recognized by the next programmer to come along and read the code, and make clear that you didn't forget something, so they should still always be used.

Federated blogging with WriteFreely

Posted Mar 15, 2019 21:26 UTC (Fri) by unixbhaskar (guest, #44758)
Parent article: Federated blogging with WriteFreely

Looks interesting Jon!

Fixing programmers

Posted Mar 15, 2019 21:15 UTC (Fri) by mpr22 (subscriber, #60784)
In reply to: Fixing programmers by myxie
Parent article: Cook: security things in Linux v5.0

continue inside a switch has no special semantics - if you're inside a loop construct then it interacts with the loop, otherwise it's an error.

Fixing programmers

Posted Mar 15, 2019 20:46 UTC (Fri) by myxie (subscriber, #127909)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

If only the original designers of C had thought of using a different keyword to indicate continuing to the next case... but that would have meant another keyword to add to the language... or would it have? :-)

I'm almost scared to look up the semantics of continue inside a switch statement.

Python dictionary "addition" and "subtraction"

Posted Mar 15, 2019 19:00 UTC (Fri) by k8to (guest, #15413)
In reply to: Python dictionary "addition" and "subtraction" by NRArnot
Parent article: Python dictionary "addition" and "subtraction"

Sometimes I think language projects just try to incorporate change because solving problems means making changes in the minds of most people, while accepting the status quo as fine can feel like not solving a problem.

I find the methods quite clear and provide good hints about what sort of data type we're working with here anyway for code that doesn't bother to provide that information in an easily accessible way. Even with the relatively constrained set of semantics for built-in types in basic python, I often find production code where the developers haven't really understood the provided toolkit. I'm not a fan of adding semantic synonyms that offer no specific utility when in practice it's clear that programmers are otten already operating beyond their understanding of their tools.

Fixing programmers

Posted Mar 15, 2019 18:51 UTC (Fri) by rgmoore (✭ supporter ✭, #75)
In reply to: Fixing programmers by anselm
Parent article: Cook: security things in Linux v5.0

To my mind, requiring explicit break statements lets you do useful¹ things that would be impossible if there was an implicit break in front of every case

You could fix that by allowing an explicit fallthrough when that's what you want. Unfortunately, as you say it's too late to change the syntax now, so the best we can do is use the compiler to catch cases where the break is left out accidentally.

Fixing programmers

Posted Mar 15, 2019 17:52 UTC (Fri) by anselm (subscriber, #2796)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

Mistyping == as = is another, extremely rare error and some things will always have to be found and fixed via code rewiev and/ or testing.

You want code review and testing, but you also want reasonable compiler warnings. If a developer is made aware of a “==” vs. “=” typo by a compiler warning while they're writing the code in the first place, the issue doesn't even come up in code review or testing (where it would be more expensive, in terms of developer time, to detect and fix). This is what in security circles we call “defense in depth”.

Algorithmic errors are far more common than any kind of syntax misuse.

Yes, but the syntax problems still exist and are often easier to detect and fix. This is like saying garbage in the street is not important because there are millions of children starving in Africa.

Debian project leader candidates emerge

Posted Mar 15, 2019 17:51 UTC (Fri) by amacater (subscriber, #790)
In reply to: Debian project leader candidates emerge by mgb
Parent article: Debian project leader candidates emerge

TINC [There is no cabal]

GNOME 3.32 released

Posted Mar 15, 2019 17:47 UTC (Fri) by quotemstr (subscriber, #45331)
In reply to: GNOME 3.32 released by kmare
Parent article: GNOME 3.32 released

It's also a hack. AIUI, it's just a large fractional scaling multiplier (with who-knows-what cost perf) with a second downsampling pass. I've never understood why fractional scaling is so hard for the Gnome people to pull off when Unity was able to do it without any real problems.

Five new stable kernels

Posted Mar 15, 2019 17:37 UTC (Fri) by zdzichu (subscriber, #17118)
In reply to: Five new stable kernels by tsdgeos
Parent article: Five new stable kernels

Isn't that a duplicate of bug in Xorg itself? https://bugs.freedesktop.org/show_bug.cgi?id=109806#c24

"Not a bug in the kernel, xorg's modesetting driver needs to set all connectors/crtc's directly when using atomic. The legacy path disabled crtc B for you if you stole all its connectors for a different crtc. In the atomic case you need to disable it yourself.

This is a bug in x.org's modesetting driver."

Xorg is slowly rotting away in Wayland's world.

Debian project leader candidates emerge

Posted Mar 15, 2019 17:14 UTC (Fri) by willy (subscriber, #9762)
In reply to: Debian project leader candidates emerge by mgb
Parent article: Debian project leader candidates emerge

I think it's fairly clear that NOTA would handily defeat this candidate.

Debian project leader candidates emerge

Posted Mar 15, 2019 16:56 UTC (Fri) by cjwatson (subscriber, #7322)
In reply to: Debian project leader candidates emerge by mgb
Parent article: Debian project leader candidates emerge

This sort of thing is analogous to problems in statutory interpretation. One traditional canon in that field is called "expressio unius est exclusio alterius", and if applied here would imply that the mention of one category of permitted nominations would ordinarily exclude others. One might also try to work out what the drafters intended (which should be fairly straightforward since they could just be asked directly). Or there are various other principles that one might apply.

It's also worth mentioning that if anyone were allowed to nominate themselves, then there would be no purpose in specifically listing it under the powers of individual developers (see 3.1(3)).

In any case, if you have a dispute about the interpretation of the constitution, you're free to ask the Secretary (under 7.1(3)).

Debian project leader candidates emerge

Posted Mar 15, 2019 16:37 UTC (Fri) by corbet (editor, #1)
In reply to: Debian project leader candidates emerge by mgb
Parent article: Debian project leader candidates emerge

I don't really have a horse in this race, but I think that's debatable. The constitution says "For the first week any Developer may nominate themselves as a candidate Project Leader". It doesn't say "any interested person". The person you are referring to is not a developer, and thus does not appear to qualify.

(For the curious, this self-nomination is what's being discussed here.)

Five new stable kernels

Posted Mar 15, 2019 16:02 UTC (Fri) by tsdgeos (guest, #69685)
Parent article: Five new stable kernels

Sadly the regression that makes Linux 5.0 give a black screen for a whole class of laptops hasn't been fixed yet :(

https://bugs.freedesktop.org/show_bug.cgi?id=109959

Fixing programmers

Posted Mar 15, 2019 15:56 UTC (Fri) by rweikusat2 (subscriber, #117920)
In reply to: Fixing programmers by Cyberax
Parent article: Cook: security things in Linux v5.0

"Switch is fine except that it's not"?

I agree that the fallthrough is mostly useless and that all the time people and computers wasted by typing or processing

break;

statements could have been put to better uses. But overloading the syntax supposed to be used for inline documentation so that yet more useless code has to be added is not the answer, it's the problem taken to the next power.

Debian project leader candidates emerge

Posted Mar 15, 2019 15:53 UTC (Fri) by mgb (guest, #3226)
Parent article: Debian project leader candidates emerge

The Debian Constitution provides that Debian Developers may nominate themselves. There is no prohibition against other nominations. There is no requirement that the DPL be a Debian Developer.

Under the Debian Constitution there is only one valid candidate, the nomination period has closed, and no more candidates may be nominated. Those who would like to see a second election have the option to vote None Of The Above in the first.

The refusal by the cabal to allow Debian Developers to vote for the only valid nominated candidate is entirely unconstitutional.

Fixing programmers

Posted Mar 15, 2019 15:51 UTC (Fri) by rweikusat2 (subscriber, #117920)
In reply to: Fixing programmers by anselm
Parent article: Cook: security things in Linux v5.0

Linux makes heavy use of gcc extensions already, so, their would be little, additional damage done here. Further, my opinion on this is that C would benefit more from proper multiway conditional than from other, invasive additions like threading support, IOW, if someone would implement support for something like this in a compiler I could use, I'd certainly be using it and expect that the standardized language would eventually catch up.

BTW, I've come to dislike if (a = b) warnings as well: Mistyping == as = is another, extremely rare error and some things will always have to be found and fixed via code rewiev and/ or testing. Algorithmic errors are far more common than any kind of syntax misuse.

Fixing programmers

Posted Mar 15, 2019 15:28 UTC (Fri) by jrigg (guest, #30848)
In reply to: Fixing programmers by HelloWorld
Parent article: Cook: security things in Linux v5.0

> the C committee is run by a bunch of stick-in-the-muds.
This could be considered a feature.

GNOME 3.32 released

Posted Mar 15, 2019 15:24 UTC (Fri) by rahulsundaram (subscriber, #21946)
In reply to: GNOME 3.32 released by mathstuf
Parent article: GNOME 3.32 released

> Why is it required for Wayland specifically? Or is it just that X has had it "forever" anyways?

The latter. Note that it is not Wayland that should have the accessibility features. It is the individual Desktop Environments or Window Managers

GNOME 3.32 released

Posted Mar 15, 2019 14:43 UTC (Fri) by mathstuf (subscriber, #69389)
In reply to: GNOME 3.32 released by yodermk
Parent article: GNOME 3.32 released

Why is it required for Wayland specifically? Or is it just that X has had it "forever" anyways?

GNOME 3.32 released

Posted Mar 15, 2019 14:00 UTC (Fri) by yodermk (subscriber, #3803)
Parent article: GNOME 3.32 released

But does it have the "mouse locator" accessibility feature working in Wayland yet? You know, where you press Ctrl and a little animated squiggly grows around the cursor. I brought that up years ago, filed in Bugzilla, it was closed and merged with another issue about re-writing some layer, and AFAIK there has been no action on it.

IMHO essential for Wayland usability.

Fixing programmers

Posted Mar 15, 2019 13:33 UTC (Fri) by anselm (subscriber, #2796)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

What would solve it would be to add another syntactic construct providing an actual multiway conditional, where the code blocks associated with the cases are independent of each other and at most one will be executed.

That would help. But programs using this construct would no longer be C programs. If sticking to a popular and well-understood language standard is of value to you (which makes sense in a large community like the Linux kernel crowd), then this is something not to be thrown away lightly.

OTOH, code that uses the “metasyntactical sort-of language extension” is valid C whether the compiler pays attention to it or not. But if you have a compiler that does pay attention, it's probably a good idea to exploit that, much like we exploit our compiler's paying attention to all sorts of other little things that might otherwise slip us by (think “if (foo = bar) …”).

Like a surgical scalpel, the C programming language is conceptually pretty simple but using it still requires considerable care and attention to detail. I haven't done lots of C programming recently but when I do I'm humble enough to appreciate all the assistance I can get.

Fixing programmers

Posted Mar 15, 2019 13:19 UTC (Fri) by anselm (subscriber, #2796)
In reply to: Fixing programmers by jschrod
Parent article: Cook: security things in Linux v5.0

The Unix guys worked shoulder-to-shoulder with the C guys

Not just that. The Unix guys basically were the C guys.

I wonder what all the fuss is about. The C language has its problems but as far as I'm concerned break is not one of the bigger ones. It's more of an idiosyncrasy. To my mind, requiring explicit break statements lets you do useful¹ things that would be impossible if there was an implicit break in front of every case (think Duff's Device), and that, together with the fact that this style of switch is fairly straightforward and minimal to implement in a C compiler based on other stuff you already have, may have been enticing to the designers of the language.

Chances are that if you were designing the C programming language from scratch today you would do various things differently, probably including switch statements. But for better or worse, we're stuck with C – at least the C switch statement – as it has been for almost 50 years, so we must remember to put a break in all the appropriate places. If we can get the compiler to remind us to do that if we forgot, or else to tell it explicitly that we didn't want one, that can only be a good thing – ace programmers of rweikusat2's calibre will of course always do the Right Thing™ from the get-go and should therefore not ever be annoyed by superfluous compiler warnings, but people like you and I (and incidentally like jschrod's, my beard is also fairly white by now) may appreciate an occasional heads-up.

¹ Useful, that is, if you think of C as a high-level portable assembly language.

Python dictionary "addition" and "subtraction"

Posted Mar 15, 2019 11:29 UTC (Fri) by jani (subscriber, #74547)
In reply to: Python dictionary "addition" and "subtraction" by NYKevin
Parent article: Python dictionary "addition" and "subtraction"

Doesn't that require the dict values to be integers? The proposal above only expects them to have + and - operators.

Cook: security things in Linux v5.0

Posted Mar 15, 2019 11:24 UTC (Fri) by jani (subscriber, #74547)
In reply to: Cook: security things in Linux v5.0 by rweikusat2
Parent article: Cook: security things in Linux v5.0

> the language is not that complicated, mind you

I find that more than a little condescending.

Even so, it's not about complexity of the language per se. It's about the complexity of using the language in a way to avoid common mistakes. The problem is in the language, not the people.

> inventing one bazillion /*** OMFG !! Not like Pascal !!! ***/ pseudosyntactic tags?

Granted, I would have preferred __attribute__ ((fallthrough)) with some syntactic sugar macro. But this is one. What are the other (bazillion - 1)? Or even your top three pet peeves?

GNOME 3.32 released

Posted Mar 15, 2019 11:11 UTC (Fri) by sbakker (subscriber, #58443)
In reply to: GNOME 3.32 released by jhoblitt
Parent article: GNOME 3.32 released

Really? For me it was one of the reasons to switch to Wayland. I had different monitors of various sizes and resolutions (1 external 24", 1 external 27", and a 13" laptop display), and Wayland allowed me to scale individual monitors.

My main problem with Wayland for now is that if the GNOME Shell crashes, my whole session is gone. Under X.org, it just restarts the shell.

Fixing programmers

Posted Mar 15, 2019 10:28 UTC (Fri) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by epa
Parent article: Cook: security things in Linux v5.0

> Pushing it upstream to the ISO C standard would be better still of course.
Or people could just switch to C++ where is this already standardized, along with dozens of other useful features.
https://en.cppreference.com/w/cpp/language/attributes/fal...

Fixing programmers

Posted Mar 15, 2019 9:24 UTC (Fri) by halla (subscriber, #14185)
In reply to: Fixing programmers by jem
Parent article: Cook: security things in Linux v5.0

I think it's older than guis. Wasn't this already the default in things like db2 on the Apple 2? Or am I backprojecting... I cannot remember anything else than tab, shift-tab for moving between fields, enter for submitting and esc for cancelling.

Development quotes of the week

Posted Mar 15, 2019 9:20 UTC (Fri) by mjthayer (guest, #39183)
Parent article: Development quotes of the week

Regarding the SSPL - do the target users have any reasonable alternative to MongoDB? If not I could imagine it working for MongoDB unless someone thinks it worth their time to fork the pre-SSPL version. But that leads to the more interesting question is - is there anything that the MongoDB team can offer large users which would cost them more if they were to do it themselves in-house? Like prioritising certain features or fixes during development?

Five new stable kernels

Posted Mar 15, 2019 9:20 UTC (Fri) by alanjwylie (subscriber, #4794)
In reply to: Five new stable kernels by unixbhaskar
Parent article: Five new stable kernels

Most people will never see this error because 1) they are running on Intel processors, rather than AMD and 2) they have never tweaked their kernel config to disable options that aren't applicable to their use case. In particular, it involves going first to "General Setup", enabling "CONFIG_EXPERT", then going to to "Processor type and features", selecting "Supported processor vendors" and in that sub-menu disabling "Support Intel Processors".

Five new stable kernels

Posted Mar 15, 2019 9:12 UTC (Fri) by nivedita76 (subscriber, #121790)
In reply to: Five new stable kernels by unixbhaskar
Parent article: Five new stable kernels

Your question is very confusing. Not sure exactly what you were asking but that config flag is not mandatory, presumably someone introduced a bug with this release.

Fixing programmers

Posted Mar 15, 2019 8:53 UTC (Fri) by epa (subscriber, #39769)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

OK, so I can only speak for my personal experience here. When writing switch statements I sometimes forget the 'break;' at the end of one case. This happens by accident rather more frequently than the times when I deliberately want to use fallthrough. So for me, it saves time to have a warning or error from the compiler when implicit fallthrough is used -- it saves my time in headscratching and debugging when the code doesn't work, and may in some cases stop buggy code from being used in production systems. This has nothing to do with "disagreeing" with the C standard. There are plenty of other cases where the standard allows a particular construct but I prefer to get a warning, because it is empirically more likely to indicate a mistake on my part rather than a deliberate choice, and so the warning helps me be a more productive programmer. YMMV.

A language extension is a great idea. GCC effectively provides a GCC-specific extension to the C language here -- or perhaps you could call it a restriction, since the code allowed with -Wimplicit-fallthrough is a subset of that without the warning. Either way, it's useful in practice to many programmers, and Linux has started making use of it. Pushing it upstream to the ISO C standard would be better still of course.

Fixing programmers

Posted Mar 15, 2019 7:45 UTC (Fri) by jem (subscriber, #24231)
In reply to: Fixing programmers by Wol
Parent article: Cook: security things in Linux v5.0

> Case in point - who was the idiot that decided that in Windows "tab" should step between fields, and "return" should send the completed form back to the program. It took a MAHOUSIVE training effort to fix that programming blunder, and people STILL get it wrong on a regular basis.

That convention is older than Windows, possibly older than the Apple Mac GUI. The real idiot was the one who later decided to implement this functionality differently on marginal platforms like Acorn's RISC OS.

Fixing programmers

Posted Mar 15, 2019 7:11 UTC (Fri) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

> Assuming this problem exists (cf next paragraph), adding a metasyntactical sort-of language extension requiring an explicit something in both cases doesn't solve it.
Actually it does.

Five new stable kernels

Posted Mar 15, 2019 3:54 UTC (Fri) by unixbhaskar (guest, #44758)
In reply to: Five new stable kernels by alanjwylie
Parent article: Five new stable kernels

Are those mandatory configuration flags to get the kernel compiled successfully?? But, why it would get not set by default?? Does it require manual intervention to set it on compile time??

It might not sound rational ...but still ..I just did compile (No VM, all in the physical partition) for 6 different distros and all of them went well.

Anyway, probably I am not able to capture the point, honestly.

The creation of the io.latency block I/O controller

Posted Mar 15, 2019 3:41 UTC (Fri) by unixbhaskar (guest, #44758)
Parent article: The creation of the io.latency block I/O controller

Well, sounds good. Many many moons ago, I used to handle throttled web server, specifically runs Apache, and I had encountered quite a few oom. Due to my lack of understanding and precautions, I failed to get over it. Looks like something going to assist people like me in a big way. Thank you and the entire fellas for the hard work.

GNOME 3.32 released

Posted Mar 15, 2019 3:04 UTC (Fri) by jhoblitt (subscriber, #77733)
In reply to: GNOME 3.32 released by kmare
Parent article: GNOME 3.32 released

Which is a bummer as I've never had a good Xwayland experience with multiple displays.

Emacs and Magit

Posted Mar 15, 2019 2:00 UTC (Fri) by bb010g (guest, #130946)
In reply to: Emacs and Magit by ceplm
Parent article: Emacs and Magit

I'd also highly recommend checking out gina.vim. https://github.com/lambdalisue/gina.vim

Fixing programmers

Posted Mar 15, 2019 1:47 UTC (Fri) by jschrod (subscriber, #1646)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

Ah, the "programming C is only for the math elite argument". Splendid.

As an old Unix grey beard (well, not quite literally, by now my beard is white), I'll have to say that demanding that all kernel programmers are "proficient in math" and "have to learn C" is not sufficient. Some decades ago, I was able to read binary OS/370 code, and did so everyday, not needing disassemblers. Nowadays, not so much any more; but that it is clearly the reason why I'm too old to be a kernel developer. But, since I still remember this time at the start of the 80s, it's not too far away, these kids today should get off the lawn. Obviously, this kind of my then-proficiency should be the standard for all current kernel developers -- they should stop writing that kind of code in C (a.k.a. abstract assembler of PDP-11 hardware) and should start using better abstract assembler for current hardware system abstractions. Since they are these geniuses, this should be no problem at all; if there is no such system, they can invent it. The Unix guys worked shoulder-to-shoulder with the C guys, and did it too, don't they. Insistence on C is so 70s, bring on new assemblers!

IOW: An argument that one should "fix programmers" is as silly as my paragraph above.

FTR: I've known Dennis Ritchie personally, and I'm proud that he took the time sharing several evenings with me. He was the first person who would have called these arguments "you have to learn C properly" not appropriate, since he knew the deficiency of his own creation quite well.

Announcing the release of sway 1.0

Posted Mar 15, 2019 0:01 UTC (Fri) by jdulaney (subscriber, #83672)
Parent article: Announcing the release of sway 1.0

I wonder if copy/paste works yet.

Fixing programmers

Posted Mar 14, 2019 23:03 UTC (Thu) by Cyberax (✭ supporter ✭, #52523)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

> What would solve it would be to add another syntactic construct providing an actual multiway conditional, where the code blocks associated with the cases are independent of each other and at most one will be executed.
Why? Switch is fine for this. In 99% of cases it's NOT used as computed goto, so implicit fallthrough is actively harmful.

That's why other more modern languages (C#, Go, ...) have ditched that stupid semantics.

And yes, it's stupid. So stupid that the very first linter for C had the "forgotten break" inspection.

GNOME 3.32 released

Posted Mar 14, 2019 22:27 UTC (Thu) by jkingweb (subscriber, #113039)
In reply to: GNOME 3.32 released by josh
Parent article: GNOME 3.32 released

Some of us have been asking for it for as long as hours! I was setting up a new computer with a 12" 2400x1600 display today with GNOME and thought to myself "200% is very readable but I with I could scale it to 150% or even just 175%." What timing.

Fixing programmers

Posted Mar 14, 2019 21:59 UTC (Thu) by NAR (subscriber, #1313)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

Take a look at this road crossing. The 2011 October state shows something like the "C switch statement". It's clearly defined in the highway code that the cyclists coming from the north ("behind") have priority, the cars should give way to them. Yet many cyclists were hit in this very road crossing - apparently the drivers "just can't learn" and apply the highway code. So the road crossing was updated. The highway code was not changed (the semantics of the C switch statement did not change), the cars still need to yield to the cyclists. It's the radius of the curve that was changed, so the drivers have to break more, otherwise they can't turn in, and at a slower speed they have more time, more chance to notice the cyclist (the path of the cyclist was also changed, now there's no 250 m straight line before the crossing, only 20 m, so they don't arrive to the crossing at 30 km/h speed). Effectively a "look around carefully" comment was introduced to this "C switch statement".

Accidentally leaving out a break statement in the Linux kernel might not lead to actual accidents (but who knows, Linux is running on cars nowadays), but it's still better to make it safer for those who "just can't learn", just like it's better to make that road crossing safer.

Fixing programmers

Posted Mar 14, 2019 21:19 UTC (Thu) by rweikusat2 (subscriber, #117920)
In reply to: Fixing programmers by HelloWorld
Parent article: Cook: security things in Linux v5.0

What's so wonderful about always trying to misunderstand something?

It is claimed that the semantics of switch, specifically, that it provides a way to jump to a certain location in a block but no automatic exit from there before the code tagged with the next case-label starts, are problematic. Assuming this problem exists (cf next paragraph), adding a metasyntactical sort-of language extension requiring an explicit something in both cases doesn't solve it. It just punishes people who need the fallthrough semantics. What would solve it would be to add another syntactic construct providing an actual multiway conditional, where the code blocks associated with the cases are independent of each other and at most one will be executed.

As to the "problematic", are we perhaps totemically repeating some "But that's not how I think it should be done!"-complaints someone initially made at least 40 years ago which since keep being repeated out of habit? According to the numbers in the text, about 99% of the warnings were false positives. Comparing the number of case-labels in a current -stable tree with the number of missing breaks indicates that this aspect of the code was correct in about 99.99% of all cases. This doesn't look like a "common" problem to me.

GNOME 3.32 released

Posted Mar 14, 2019 20:51 UTC (Thu) by kmare (guest, #113077)
In reply to: GNOME 3.32 released by josh
Parent article: GNOME 3.32 released

I agree. Just as a reminder though, it is Wayland only.

Five new stable kernels

Posted Mar 14, 2019 20:27 UTC (Thu) by alanjwylie (subscriber, #4794)
Parent article: Five new stable kernels

There's a build failure with AMD-only builds, I've seen it myself, there is also this report by Alec Ari

https://lkml.org/lkml/2019/3/13/1113

GNOME 3.32 released

Posted Mar 14, 2019 19:51 UTC (Thu) by josh (subscriber, #17465)
Parent article: GNOME 3.32 released

Fractional scaling is *huge*; people have been asking for that for years, and it's critical for certain resolutions like 2560x1440.

Cook: security things in Linux v5.0

Posted Mar 14, 2019 18:49 UTC (Thu) by tdz (subscriber, #58733)
In reply to: Cook: security things in Linux v5.0 by rweikusat2
Parent article: Cook: security things in Linux v5.0

Time for coloring some bike sheds!

I don't see anything wrong with this. Of all the changes from the linked article, this is the least controversial. This has no runtime/compile overhead or any form of weird semantics. It doesn't obfuscate the code base. This change has zero negative impact. If anything, I'd ask for using the related gcc attribute instead of a comment.

Fixing programmers

Posted Mar 14, 2019 18:31 UTC (Thu) by HelloWorld (guest, #56129)
In reply to: Fixing programmers by rweikusat2
Parent article: Cook: security things in Linux v5.0

> This "reminder" that a very vocal group of people has been disagreeing with pretty much any C design decision since 197x isn't useful. A much better idea would be to push for a language extension with the desired semantics, assuming some set of desired semantics can ever be agreed on.
How is a "fallthrough comment" at the end of a case branch not a language extension, assuming that its presence is enforced by the compiler?

Besides, this has been standardized in C++17. The only reason it's not in C is that the C committee is run by a bunch of stick-in-the-muds.

Next 90 comments.