|
Does decltype complicate the C++ spec/implementations very much? It seems like it would make typechecking and template expansion really difficult.
|
# ? Jan 11, 2019 02:37 |
|
|
# ? May 17, 2024 02:49 |
|
DONT THREAD ON ME posted:Does decltype complicate the C++ spec/implementations very much? It seems like it would make typechecking and template expansion really difficult. I don't think so? you could already use expressions in type definitions. it's not terribly different from sizeof in that regard, and it was already implemented by eg. gcc as an extension (named __typeof iirc). visual c++ didn't have it but through one of its own extensions and an incredibly convoluted contortion thereof (e: actually a compiler bug), that I never looked too hard into, you could implement it in a reasonably complete way (as done eg. by Boost.Typeof) hackbunny fucked around with this message at 03:06 on Jan 11, 2019 |
# ? Jan 11, 2019 03:03 |
|
yeah, the most complicated thing in c++ that most non-implementors are unaware of is that you can put almost any expression in a type (which you could always do because of array bounds and sizeof) and so there are all sorts of things like function template overloading that are dependent on arbitrary expression structure. this is why the itanium c++ abi has manglings for what’s almost a complete semantically-annotated reconstruction of the expression grammar (there are a handful of things that are hand-waved in the spec as “it’s unspecified if you overload function templates based on this”, like if the same lookup resolves to different overload sets, and the abi doesn’t mangle those differences)
|
# ? Jan 11, 2019 03:21 |
|
oh yeah of course, that makes sense.
|
# ? Jan 11, 2019 03:56 |
|
jit bull transpile posted:or just name it INCLUSIVE_MAX. that seems like the easiest part to solve unless I've totally misunderstood this but EXCLUSIVE_MAX imo: it's easier to write the enum like this code:
Than this: code:
and the later is also more error prone since you need to update the last line if you insert a new value after Bar either way you pretty much have to use enum class for this to avoid name collisions with the MAX value between multiple enums
|
# ? Jan 11, 2019 09:01 |
|
you could also make a traits class and specialize it for each enum. then you can fill it with whatever extra flags and data you need. if you forget to specialize it or leave something out, you'll get the same compile time errors so like code:
|
# ? Jan 11, 2019 21:33 |
|
on what world would “max” be exclusive
|
# ? Jan 12, 2019 03:58 |
|
in that an enum value called "max", as a distinct value after all the other values, is the max value. I've seen it in codebases before.
|
# ? Jan 12, 2019 04:20 |
|
then it should be called ceiling or top or something imo calling it max leads to confusion as it clearly isnt the maximum value....
|
# ? Jan 12, 2019 04:50 |
|
or “num” or “count”, yeah
|
# ? Jan 12, 2019 05:01 |
|
🎵 return of the MAX
|
# ? Jan 12, 2019 05:10 |
|
i definitely prefer "count" but yeah everybody has weird codebase wrinkles and its not my job to tell them theyre wrong
|
# ? Jan 12, 2019 05:18 |
|
you need to have someone to say "decision" so even if theyre wrong youre still all doing the exact same wrong thing make someone there say "this is the final enum" and then thats it then complain here
|
# ? Jan 12, 2019 05:32 |
|
so uh what's the difference between software development and software engineering? I've got my degree in computer and I still don't know
|
# ? Jan 16, 2019 07:49 |
|
software development is a descriptive term for the processes of making software software engineering is an aspirational term for the idea of applying sound engineering practices to making software almost nobody does actual software engineering but there are a lot of people who do perform convincingly authentic rites and incantations before letting the programmers do what they always do
|
# ? Jan 16, 2019 08:00 |
|
cool guess I'll feel free to apply for any of those too
|
# ? Jan 16, 2019 08:02 |
|
oh, as job titles they're interchangeable. different companies will use either "developer" or "engineer" for their programmers. means absolutely nothing there is no agreement about numeric ranks between companies. with named ranks, there's general agreement that "junior" means someone who needs a lot of guidance to get anything done and "senior" means someone who can provide that guidance, but both might correspond to multiple levels of payscale
|
# ? Jan 16, 2019 08:17 |
|
rjmccall posted:software development is a descriptive term for the processes of making software I've got a friend who programs for the military and he says he spends upwards of 80% of his time with uml diagrams and hundred page requirement/api docs, sounds loving miserable imo The other 20% being dealing with the grey beards who have full pension, effectively tenure, and haven't learned a new fact in 30 years and not about to start now
|
# ? Jan 16, 2019 09:37 |
|
rjmccall posted:software development is a descriptive term for the processes of making software depends on where you are as well of course, in some countries it is a defined diploma which requires a broader mix of subjects (e.g. courses in ethics, organizational theory and the philosophy of science)
|
# ? Jan 16, 2019 10:23 |
|
i worked for a company that called computer engineering grads "engineers" and everyone else "developers" (incl cs degree holders) it had zero influence on your status, salary, job duties, etc at the company though
|
# ? Jan 16, 2019 13:18 |
|
when I coded serial device drivers for medical equipment I was a “programmer”. now that I make lovely web apps I’m an “engineer”.
|
# ? Jan 16, 2019 13:43 |
|
brand engager posted:so uh what's the difference between software development and software engineering? I've got my degree in computer and I still don't know ive seen people use "developing" in a degrading way to imply just churning out rote code to accomplish something, and "engineering" conveys a striving for elegance or applying more academic / sophisticated concepts to something. that ignores the fact that 99% of developers are not doing anything new or exciting, but whatevs.
|
# ? Jan 16, 2019 13:46 |
|
I feel that an "engineer" should have some kind of professional liability for the quality/safety of their work. That's why I call myself a developer.
|
# ? Jan 16, 2019 15:30 |
|
Doom Mathematic posted:I feel that an "engineer" should have some kind of professional liability for the quality/safety of their work. That's why I call myself a developer.
|
# ? Jan 16, 2019 15:54 |
in latvian “software engineering” sounds very goofy so everyone is just a something programmer
|
|
# ? Jan 16, 2019 16:48 |
much rarer - something development specialist
|
|
# ? Jan 16, 2019 16:49 |
|
those are much better titles tbh
|
# ? Jan 16, 2019 16:58 |
|
cinci zoo sniper posted:much rarer - something development specialist
|
# ? Jan 16, 2019 17:23 |
mystes posted:"Development specialist" sounds like you specialize in developmental disorders. nah it's, for example, front-end development specialist
|
|
# ? Jan 16, 2019 17:24 |
|
"software developer" and "software engineer" mean the same thing as titles, but beware of jobs with the title "programmer". for some reason certain companies use that title to imply a lower-skilled and lower-paid position, someone who just codes what they're told to, possibly interchangeable with offshore
|
# ? Jan 16, 2019 17:39 |
|
mystes posted:"Development specialist" sounds like you specialize in developmental disorders. sounds like something that disappeared with digital cameras
|
# ? Jan 16, 2019 18:18 |
|
In some jurisdiction, the engineer title is regulated, even for software. This is the case in Quebec, where you have to study in a certified program, pass an exam that includes a lot of knowledge from regular engineering disciplines and civil responsibilities, complete an apprenticeship under an already-certified engineer, get professional insurance, etc. The OIC (order for all engineers) even won a lawsuit against microsoft for providing "Microsoft Certified Systems Engineer" certifications, which they saw as encroaching on their discipline. I'm a developer because doing otherwise is kind of risky. I'm sometimes tempting the devil with being a "systems architect" because actual architects don't seem to be too protective of that term in other disciplines.
|
# ? Jan 16, 2019 19:08 |
|
well architect isn't a real discipline
|
# ? Jan 16, 2019 19:32 |
|
for all you windows internals fans this is a p great breakdown of all of windows event handling mechanisms and a bunch of their quirks and limitations https://github.com/python-trio/trio/issues/52 quote:Windows has 3 incompatible families of event notifications APIs: IOCP, select/WSAPoll, and WaitForMultipleEvents-and-variants. They each have unique capabilities. This means: if you want to be able to react to all the different possible events that Windows can signal, then you must use all 3 of these. Needless to say, this creates a challenge for event loop design. There are a number of potentially viable ways to arrange these pieces; the question is which one we should use.
|
# ? Jan 16, 2019 19:40 |
|
TOPS-420 posted:for all you windows internals fans this is a p great breakdown of all of windows event handling mechanisms and a bunch of their quirks and limitations this is cool, thanks
|
# ? Jan 16, 2019 20:09 |
|
the first PE i worked with put sleep() in an ISR (its bad) and pitched a fit when i said to find another solution in the review he called it 'debouncing' i haven't been all that impressed with PE's
|
# ? Jan 16, 2019 23:15 |
|
Doom Mathematic posted:I feel that an "engineer" should have some kind of professional liability for the quality/safety of their work. That's why I call myself a developer. I'm a senior train driver (that's what we call network engineers here) and indeed I am required to pay for professional liability insurance - let me tell you it's loving expensive edit; i'm a consultant tho salaried people don't have to pay ofc
|
# ? Jan 16, 2019 23:25 |
|
TOPS-420 posted:for all you windows internals fans this is a p great breakdown of all of windows event handling mechanisms and a bunch of their quirks and limitations I wonder why microsoft never made an async I/O variant of select/poll so you could get writability/readability notifications on a iocp, instead of having to mix async and non-blocking styles. hell it could have been done with special flags to WSASend/WSARecv, to ask for a non-blocking send/receive. immediate completion would mean the operation completed without blocking, asynchronous completion with error WSAEWOULDBLOCK would mean "try again now", so it could be easily used in a loop and you wouldn't pay the cost of two system calls hackbunny fucked around with this message at 00:11 on Jan 17, 2019 |
# ? Jan 17, 2019 00:04 |
|
many places in the us consider "programmer", "software developer", and "software engineer" to be three ascending tiers of skill and pay grade. consultancies often like to give people "software engineer" as a title in order to justify charging clients more. it's basically just marketing puffery in those situations i prefer to be called "developer" because while i strive to do a good job and write good code i am not liable for software failures; sort of related to why i prefer not to work on systems which, should they fail, would endanger lives my company recently did a whole drawn out title normalization process and they decided to call us "software development engineers" which is loving ridiculous
|
# ? Jan 17, 2019 02:20 |
|
|
# ? May 17, 2024 02:49 |
|
JawnV6 posted:the first PE i worked with put sleep() in an ISR JawnV6 posted:he called it 'debouncing' ok I'm so mad but this fuckin owns actually
|
# ? Jan 17, 2019 02:54 |