WTF Bites
-
@Bulb ... and miss an opportunity to post here? Nah, too easy.
You wouldn't miss that opportunity. You still did climb that pile of s trying to compile or convert that Fortran code before it turned out you don't need it all that much…
Though I could make it a XML file where arrays are stored with one node per element, and at least one attribute per element storing the type, position in the array etc.
Apropos XML, colleague is replacing a massively shitty XML parser with a sane(ish) one. Nevertheless the new one still does not have any useful conversions in it's
date_time
type (e.g. to/fromstd::chrono::time_point
).
-
@remi dunno, seems to miss the self-describing property of XML still. you'll want to be able to full express array semantics in a system-independent fashion in some sort of
prologuePROLOG element, or perhaps included resource. some of this, for sure, is taken care of by XSD, but for interoperability you need a fully specified declaration and at least some assertions about implementation (0 vs 1 based addressing, for instance).FTFTT
-
… not that
std::chrono::time_point
had many useful methods itself. You still end up converting it to bad oldtime_t
to be able to print it.
-
Apropos chrono, https://en.cppreference.com/w/cpp/chrono/duration says:
std::chrono::days (since C++20) duration</signed integer type of at least 25 bits/, std::ratio<86400>>
std::chrono::weeks (since C++20) duration</signed integer type of at least 22 bits/, std::ratio<604800>>
std::chrono::months (since C++20) duration</signed integer type of at least 20 bits/, std::ratio<2629746>>
std::chrono::years (since C++20) duration</signed integer type of at least 17 bits/, std::ratio<31556952>>Yeah,
month = 2629746 s = 30 d + 10 h + 29 min + 6 s
and
year = 31556952 s = 365 d + 5 h + 49 min + 12 s
are extremely useful conversions. We all know what
date manipulation is, but there is ISO-8601 to describe it and legal requirements to demand it, so we can sure all imagine the fun when hapless junior code monkeys start calculating dates like
something + 2 * month
orother + years / 2
and the bugs from production because it will happen to round correctly on the day QA tests it.
-
@Bulb upvoted for amazing triforce.
-
with a 10-foot pole
You'd think that all those 10-foot Poles, Poland would have an unbeatable basketball team.
-
@HardwareGeek well, we are volleyball world champions...
-
@Gąska I thought that was Bulgaria.
-
-
@Zerosquare ... s everywhere
-
Though the real memory hog is all those high-res images for the xxxxxxxxdpi displays the OEMs like to show off with.
Things might be improving on that front - I think the S21 has a lower resolution than its predecessors.
-
@Zerosquare said in WTF Bites:
"If you wish to do work like this, I suggest you find a different community to run your experiments on, you are not welcome here."
-
@Zerosquare said in WTF Bites:
University duo thought it would be cool to sneak bad code into Linux as an experiment. Of course, it absolutely backfired
Really? Or is that a plot by the Linux community to prove that they can find out such contributions?
Let's ask some conspiration contributors!
-
LLVM can cross-compile to Windows all right, but every language has a library of compiler intrinsics, and needs to link the system libraries that may have their quirks.
Windows is tricky for several fiddly reasons.
- It uses a different object file format to everything else. Just because.
- SEH, because Fuck You that's why.
- It uses a different set of calling conventions.
- You really really have to exactly match the system class layouts for a number of things to stand a chance of working at all. And that means you've also got to match the basic type widths and alignments because otherwise you can't use the system headers and you will go insane.
There's a bunch of other little things, but they're mostly minor by comparison.
-
Good opportunity to throw not just the fortran code but also
f2c
into the trash were they belong and just rewrite it.If it's something intensively numeric, it's probably worth keeping the Fortran code. Alas. Numeric algorithms are intensely difficult to get right, and it is the one area where Fortran has an advantage over everything else (it's where the actual numerical analysts worked for many years).
-
it's probably worth keeping the Fortran code
Nah, with @topspin on this one: . Code that people with an actual understanding of numerics worked on is a small minority, Fortran or not.
-
@cvi I'm actually with @dkf here. If it was intensively numeric code (which we do have quite a lot), I would want to keep it in Fortran as much as possible, because even if the people who wrote it didn't understand all the numerical subtleties, at least they understood what they were trying to achieve. If we now pick up that code and rewrite it in C++ (or anything else), we don't have the same understanding of the physical/mathematical problem they were solving, so even assuming we get the numerical implementation right, we can't be sure we haven't introduced some subtle effects that will mess up things later.
Yes, it means the whole thing is fragile, and yes that's a bad thing. But it's the nature of a lot of highly complex scientific code, and the point here is that it's a fragility that we have learnt to live with in, sometimes for more than 20 years (and some of the code I'm thinking of is probably close to 30 years, if not more, and still being used).
But on this specific case, I'll remind everyone that the code is actually just reading a serialised binary format.
-
LLVM can cross-compile to Windows all right, but every language has a library of compiler intrinsics, and needs to link the system libraries that may have their quirks.
Windows is tricky for several fiddly reasons.
Those are mostly backend concerns though, and the backend already handles them.
The language-specific part basically just needs a library to call for built ins and the standard library.- It uses a different object file format to everything else. Just because.
Well, it predates the Unix one.
- SEH, because Fuck You that's why.
SEH is one of the few advantages of the Windows ABI. They allow actually handling faults; Unix signals effectively don't.
- It uses a different set of calling conventions.
The IR is still high level enough that the front-end does not need to care.
- You really really have to exactly match the system class layouts for a number of things to stand a chance of working at all. And that means you've also got to match the basic type widths and alignments because otherwise you can't use the system headers and you will go insane.
The runtime library for Fortran probably still needs to be written in C to correctly bind to the standard C library of the platform, and compiling that is already supported.
There's a bunch of other little things, but they're mostly minor by comparison.
But which of them actually affect the front-end?
-
They allow actually handling faults; Unix signals effectively don't.
This explains the stability and fault-tolerance of a Windows installation.
-
-
But which of them actually affect the front-end?
The compiler driver front-end? Why would you give much of a fuck about that? You always drive those things by script or cmake or GUI or something like that because nobody wants to remember all the details of building a large project.
(All the matters I raised show up in actual code. Therefore the compiler needs to care. You might be able to ignore them most of the time… but the compiler doesn't have that luxury.)
-
@Zerosquare said in WTF Bites:
except this is new. I missed that this was Marlinspike. She's hilarious.
-
@Zerosquare d in some other dedicated thread.
But, fuck, if you told me when I was a kid that - in the future - we'd be uploading booby trapped files to our star trek communicator-like devices that all of us carry around daily... Hmm. Actually, I'd have been quite excited about that future. Less enthusiastic now, though.
-
The compiler driver front-end?
Front-end means the part of the compiler that deals with specific language. Parses the language and generates IR. In contrast to back-end, which is the part that takes the IR and produces machine code for the target platform. Which is why most of the quirks of the platform only concern the back-end.
-
@Bulb he said
compiler driver
notcompiler
! (are there in fact any plumbers here tonight?)
-
@Gribnit But I didn't.
-
This download may take a while, seeing as the progress has gone into the negatives by quite a lot.
-
@Atazhaia That's what happens when you let a slut drive your Förd. You probably shouldn't let her drive your Saab, either.
-
@Atazhaia Looks like they've put a bad condition in their för lopp.
-
för lopp.
får loop would sound closer to the english, I think. It's also more nonsensical, so an overall win.
-
för lopp.
får loop would sound closer the english, I think. It's also more nonsensical, so an overall win.
agreed, but you're wrong. it's too subtle.
-
för lopp.
får loop would sound closer the english, I think. It's also more nonsensical, so an overall win.
agreed, but you're wrong. it's too subtle.
Fár öüt!
-
I found this short story in a YouTube comment, demonstrating again that
a) front-line support is not always , and
b) some customers are even stupider than you would believe possible.
-
@HardwareGeek Any sufficiently advanced technology is indistinguishable from magic. For some people, electricity is sufficiently advanced.
-
@HardwareGeek One could make a male-male extension chord, take a power bar, plug it into itself, then connect it to the wall socket through one of its sockets and that extension chord. Not sure if it would work as a prank on anybody.
-
@HardwareGeek Not sure if it would work
Well, I hope your fuses are fast-switching.
-
@HardwareGeek Not sure if it would work
Well, I hope your fuses are fast-switching.
With an E-type bar like we have here you can't short it. With an F-type like you have it's 50/50 that you do…
-
@Zerosquare d in some other dedicated thread.
But, fuck, if you told me when I was a kid that - in the future - we'd be uploading booby trapped files to our star trek communicator-like devices that all of us carry around daily... Hmm. Actually, I'd have been quite excited about that future. Less enthusiastic now, though.
That reminds me how I always found these common Trek tropes (overriding security access, disabling alien ship systems, non-responsive controls without any override) as painfully unrealistic and stupid. Looks like this is actually how the future works!
-
Microsoft released an update to the Remote Desktop Mac client "fixing" an error, which actually manages to cause the error for people who weren't previously experiencing it.
Hilariously enough, the change notes for 10.6.4 indicate that they fixed an issue that was causing the client to return a 0x907 error code, but it is this version that caused me to experience this issue for the first time, preventing me from doing work.
-
@loopback0 said in WTF Bites:
Microsoft released an update to the Remote Desktop Mac client "fixing" an error, which actually manages to cause the error for people who weren't previously experiencing it.
Hilariously enough, the change notes for 10.6.4 indicate that they fixed an issue that was causing the client to return a 0x907 error code, but it is this version that caused me to experience this issue for the first time, preventing me from doing work.
They just encountered a minor temporal issue where the patch notes for the next release accidentally got added to this one
-
@HardwareGeek Any sufficiently advanced technology is indistinguishable from magic. For some people, electricity is sufficiently advanced.
And we call these people hardware engineers.
-
I have three TDWTF tabs open (one main reading tab, and two dedicated to the current D&D campaign in the Games category). Suddenly all 3 windows loaded with the start of a completely different thread (NodeBB for older browsers), with no history to get back to the threads I had been reading in them.
-
@HardwareGeek said in WTF Bites:
I have three TDWTF tabs open (one main reading tab, and two dedicated to the current D&D campaign in the Games category). Suddenly all 3 windows loaded with the start of a completely different thread (NodeBB for older browsers), with no history to get back to the threads I had been reading in them.
I think it was the Jeffing of that thread that triggered it.
-
-
@HardwareGeek said in WTF Bites:
I have three TDWTF tabs open (one main reading tab, and two dedicated to the current D&D campaign in the Games category). Suddenly all 3 windows loaded with the start of a completely different thread (NodeBB for older browsers), with no history to get back to the threads I had been reading in them.
Sorry, not sorry.
-
DNS propagation
I added both records at the same time.
Fake edit: The second one has now gone through, but the time between one and the other was still a lot longer than it should have been
-
same time maybe half a minute apart
The concept of simultaneity is subjective.
-
@error Oddly enough, I checked again and the www one had disappeared, but then reappeared again after another little bit. Anyway everything is verkink now
-
-
The concept of simultaneity is subjective.
But throwing the whole lot into orbit around a black hole is a bit of an unusual way to handle DNS record updates.