Re: Variables alignment
As far as I know compilers always align your variables unless you specify that you have other priorities, for example with the [packed pragma](http://forum.nim-lang.org///nim-lang.org/docs/manual.html#foreign-function-interface-packed-pragma). For stack variables I don't know it, but I also don't see any reason for compilers to not align variables.
Re: Variables alignment
Krux02, from what I've read about lockfree (namely Jeff Preshing's blog: [http://preshing.com](http://forum.nim-lang.org///preshing.com)/ ), one requirement for atomics in x86 (and other architectures) are aligned variables. And about the types anything from 1 to four bytes. Araq,I have already tried that, my problem is that I can't use it inside objects nor tuples, and while I've yet to try I think I can't use it on typedefs neither. I know MSVC always align,but how do I know gcc/clang may align them, from what I've tested gcc aligns global variables.
Re: zeroMem can't use?
Nim has implicit initialization ( [http://nim-lang.org/docs/manual.html#statements-and-expressions-var-statement](http://forum.nim-lang.org///nim-lang.org/docs/manual.html#statements-and-expressions-var-statement) ) so: var x : array[10,char] echo repr x # ['\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0'] Anyway, `alloc0` and the like return pointers, so you need to cast to `ptr array[]` : var inputs = cast[ptr array[10,int]](alloc0(sizeof(int) * 10)) inputs[0] = 1 inputs[1] = 2 inputs[2] = 3 for i in 0..9: echo inputs[i] echo repr inputs var test = alloc0(10) zeroMem(test, 10) echo repr cast[ptr array[10,char]](test) Produces 1 2 3 0 0 0 0 0 0 0 ref 0018F050 --> [1, 2, 3, 0, 0, 0, 0, 0, 0, 0] ref 00191050 --> ['\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0'] Or just use create: var ar = create array[10,char] echo repr ar # ref 0092F050 --> ['\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0'] Note that I don't think nim's garbage collector keeps track of pointers you get from `alloc` etc, so you may have to deallocate them yourself to avoid memory leaks
Re: Reflections on Nim and using it at work
Well, strictly speaking, these are not Posix.
>From the Linux man page:
ATTRIBUTES
For an explanation of the terms used in this section, see
attributes(7).
┌─┬───┬┐
│Interface│ Attribute │ Value │
├─┼───┼┤
│forkpty(), openpty() │ Thread safety │ MT-Safe locale │
├─┼───┼┤
│login_tty() │ Thread safety │ MT-Unsafe race:ttyname │
└─┴───┴┘
CONFORMING TO
These are BSD functions, present in glibc. They are not
standardized in POSIX.
So, not sure if the posix module is the correct place to add these, even though
they are present in glibc on linux.
I tried musl as well, with it the **{.passL: "-lutil".}** is no longer needed.
Re: Reflections on Nim and using it at work
Alright, this is what I got working for the posix stuff:
{.passL: "-lutil".}
var SIGWINCH* {.importc, header: "".}: cint
var TIOCGWINSZ* {.importc, header: "".}: cuint
var TIOCSWINSZ* {.importc, header: "".}: cuint
proc openpty (amaster, aslave : ptr int, mustBeNil : pointer, termp, winp :
ptr Termios) : cint {.importc, noDecl, header: "".}
proc forkpty (amaster: ptr int, mustBeNil : pointer, termp, winp : ptr
Termios) : cint {.importc, noDecl, header: "".}
I need to the put what will make **{.passL: "-lutil".}** happen in the importc
pragmas for openpty and forkpty, but I'm not sure how to do that. dynlib did
not do it.
With the above I was able to port the remaining C functions to Nim.
As far as **SigVal** not having **sival_int** an only having **sival_ptr**,
here was my workaround:
# for inserting:
var si_value:SigVal
si_value.sival_ptr = cast[pointer](myvar)
let rc = pid.sigqueue (SIGUSR2, si_value)
# for extracting:
proc handler (signal:cint, si: var SigInfo, context:pointer) {.noconv.} =
let myvar = cast[mytype](si.si_value.sival_ptr)
Where **mytype** objects are pointer sized or smaller, in my case I'm using an
enum type. I guess because of that, I don't really see the need to support
sival_int directly in Nim.
Should something like this be included in the PR?
Here is what is says currently:
SigVal* {.importc: "union sigval",
header: "", final, pure.} = object ## struct sigval
sival_ptr*: pointer ## pointer signal value;
## integer signal value not defined!
On a different note, why were these renamed?
* **cfmakeraw** to _cfMakeRaw_
* **tcsetattr** to _tcSetAttr_
* **tcgetattr** to _tcGetAttr_
Due to case insensitivity they worked without the camelcase. The reason I
initially thought they were missing was I could not find the in the
documentation.
readlink was not renamed to readLink, setgid to setGid, etc, so I don't really
see a point in renaming the aforementioned.
So, when I do a PR, I'll need to add the proper docstrings, for the new ones
above as well.
zeroMem can't use?
win10, nim0.14.2 # use alloc0 var inputs = cast[array[10,int]](alloc0(sizeof(int) * 10)) inputs[0] = 1 inputs[1] = 2 inputs[2] = 3 for i in 0..9: echo inputs[i] echo repr inputs var test = alloc0(10) zeroMem(test, 10) echo repr cast[array[10,char]](test) 1 2 3 6553000 0 6552640 6553000 1 7607136 1 [1, 2, 3, 6553000, 0, 6552640, 6553000, 1, 7607136, 1] ['P', '@', '\25', '\0', '\0', '\0', '\0', '\0', '\9', '\0'] Maybe I use the incorrect way?
Re: Improved Complex number string representation
Every language with complex number support seems to have its own representation of them. Python does 1 + 2j R does 0+1i Julia does 0.0 + 1.0im Nim seems to follow the same convention as C++ std::complex with (0.0, 1.0) I don't know if one of those is inherently any better than another.
var parameter
Can I assume that a var parameter will always be a pointer where I can use (addr varParm) as copymem destination. It seems to work but I can copy to tempory local variable and use normal assignment if there is no guarentee it will always work.
Re: Send data structures between threads?
@Araq This sounds like object mapping, in which data from a database is mapped into a set of objects. That's quite common, and makes working with data much easier.
Re: Send data structures between threads?
> So what do you do: do you tell these teams not to use refs? Or do you tell > them to go ahead and use refs and garbage collection, and deal with the > consequences? To be honest I would fire teams who model business taxonomies with C++ classes (OMG), keeping all the data in RAM (OMG) and not using a (noSQL / SQL) database (OMG). But in the context of Nim, I would likely live with refs, thread local GCs and when somebody queries my data, I would return a copy. Note that copying a query result doesn't mean to "keep full copies of the data around in RAM" which you sort of implied.
Re: Variables alignment
Use the `codegenDecl` pragma and to generate any kind of gcc/clang annotation you need. Compilers are usually good at alignment though.
Re: GTK 3.20 Nim wrapper
Thanks for reporting. I guess you are refering to the chess game? Lets look at line 70: if p0 != p1: window(widget.toplevel).title= "invalid move, ignored." As the compiler told us, there is a symbol window defined in module gtk3 as well as gdk3. So only two possibilities to try. But of course member field "title" told us that this is the gtk window, so fix is if p0 != p1: gtk3.window(widget.toplevel).title= "invalid move, ignored." You have to do 3 more similar fixes, sorry. Another fix would be to modify module import list like import gdk3 exept window Will fix file at github tomorrow -- I wonder why this error has not occured before... And yes, for that window() cast a proc would be fine, no template necessary. But it is some work to fix all that... If you are on Linux you may test the editor as well, it is now available in v 0.3. But you have to install nimsuggest first.
Re: GTK 3.20 Nim wrapper
I get an error when trying to compile board.nim(70, 26) Error: ambiguous call; both gtk3.window(obj: expr) and gdk3.window(object: expr) match for: (Widget) How to fix it?
Re: Send data structures between threads?
I used the word "database" in the generic sense. Basically, go to any good shopping website (eg, wayfair, or amazon, or google shopping, for example), and see what that website "knows" about coffee tables. If you take a little time and study the website, I think you'll realize that it's a surprisingly large amount of knowledge. There's a whole expert system in there about coffee tables - and another expert system about bedspreads, and so forth. At the very base of it all, is a taxonomy: Home Goods > Tools > Drills and Rotary Tools Some product categories fit into the taxonomy in multiple places: Apparel > Shirts > Football Jerseys Sports > Sports Memorabilia > Football Jerseys Then, for each category of products, you have data about which products in the category are most popular, data to drive a machine-learning classifier that classifies queries to determine if they belong in that category, etcetera, etcetera. Just take a little time studying one of these websites, and I think you'll come to appreciate how much is there. So when I say a "database" here, I don't mean that it's SQL. I'm just saying it's complicated data. It's plain old C++ classes, loaded into RAM - but it's legitimately _lots_ of C++ classes, and it takes up _lots_ of RAM. So imagine, now, that you've got a dozen teams: one is responsible for implementing a machine-learning classifier that differentiates queries about drills from queries about rotary hammers. Another is building a software system whose job is to estimate the popularity of products according to a dozen different metrics. Another is building a recommendation system that uses statistics to recommend "related products." And so forth. Now, you know that at some point, you're going to need to integrate all this software into a web-server, which will probably be multithreaded. So what do you do: do you tell these teams not to use refs? Or do you tell them to go ahead and use refs and garbage collection, and deal with the consequences?
