EPTconst char ca[] = { 'h', 'e', 'y' };should assign 4 bytes and place h e y \0 in them.
The output would be:
h
e
y
The Big Programming Thread - Page 774
| Forum Index > General Forum |
Thread Rules 1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution. 2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20) 3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible. 4. Use [code] tags to format code blocks. | ||
|
Birdie
New Zealand4438 Posts
should assign 4 bytes and place h e y \0 in them. The output would be: | ||
|
beg
991 Posts
cout << size(ca); It returns 3, so I guess the while loop really is undefined eventually. Thanks for inspiring me, Birdie^^ | ||
|
Birdie
New Zealand4438 Posts
On October 06 2016 13:04 beg wrote: I just had genius idea *cough cough* cout << size(ca); It returns 3, so I guess the while loop really is undefined eventually. Thanks for inspiring me, Birdie^^ The size() function returns the length of the string excluding the null terminator. sizeof(ca) would be much more relevant to what you want to know. | ||
|
beg
991 Posts
char c1[] = { '\0' }; char c2[] = { '\0', '\0' }; and it did return the number of null terminators. Anyway. I checked sizeof(ca) and it returned the same numbers. So thanks again. (The bazillion details of C++ ... ) | ||
|
Djagulingu
Germany3605 Posts
| ||
|
Birdie
New Zealand4438 Posts
| ||
|
icystorage
Jollibee19350 Posts
 | ||
|
Djagulingu
Germany3605 Posts
On October 06 2016 15:34 Birdie wrote: Yup learned that pretty early on at uni, there's different ways around it like calling getch() right after if I recall correctly, been a while though. It doesn't clear the newline from the stdio buffer IIRC. Calling getchar() right in between worked for me, but I'm pretty sure it wouldn't work if my eol was \r\n. Still need to learn how to clear input buffer. | ||
|
Deleted User 101379
4849 Posts
On October 06 2016 16:12 Djagulingu wrote: Calling getchar() right in between worked for me, but I'm pretty sure it wouldn't work if my eol was \r\n. Still need to learn how to clear input buffer. Essentially you'll want to not use scanf in the first place. Use fgets and parse the whole line instead. scanf is a fairly dangerous function, because it doesn't check for overflow and causes problems when users enter nonsense - as users tend to do. There are methods to clear the input buffer, but they are all not very portable or reliable. | ||
|
BluzMan
Russian Federation4235 Posts
On October 06 2016 12:32 beg wrote: Hope no one minds me asking a few C++ questions every now and then :S So, I'm wondering about an excersise from the book C++ Primer. "What does this code do?" const char ca[] = { 'h', 'e', 'y' };Sooo, cp is a pointer to &ca[0] and we keep incrementing it until we encounter *cp == 0. But... ca[] is not null-terminated. I was thinking the while loop should eventually show undefined behaviour, but everytime I run this code it seems to work fine. Is it undefined? If not, what am I missing? You're perfectly right, this is undefined behavior which might include "working alrite" because by some twist of luck ca ended up in zero-filled memory. Try playing with optimisation flags, you might get different results with say, -O3. | ||
tofucake
Hyrule19177 Posts
On October 06 2016 18:25 BluzMan wrote: You're perfectly right, this is undefined behavior which might include "working alrite" because by some twist of luck ca ended up in zero-filled memory. Try playing with optimisation flags, you might get different results with say, -O3. It's not undefined behavior. The compiler automatically adds a null on the end of the array, and sizeof doesn't count it. It doesn't matter what optimization is done. If you want to create undefined behavior you would need to declare the array with malloc, then instantiate with your characters. The compiler won't automatically add a null in that case. | ||
|
beg
991 Posts
Anyway, I'm somewhat sure I figured it out now. char ca[] = "hey"; //null-terminator added! sizeof(ca) == 4; //true size(ca) == 4; //true But: char ca[] = { 'h', 'e', 'y' }; //no null-terminator! sizeof(ca) == 3; //true size(ca) == 3; //true This method of testing didn't come to my mind. I should have slept over this another night... | ||
|
Acrofales
Spain18143 Posts
On October 06 2016 22:36 beg wrote: I'm sorry for kicking off this discussion with so much disagreement ~ Anyway, I'm somewhat sure I figured it out now. char ca[] = "hey"; //null-terminator added! sizeof(ca) == 4; //true size(ca) == 4; //true But: char ca[] = { 'h', 'e', 'y' }; //no null-terminator! sizeof(ca) == 3; //true size(ca) == 3; //true This method of testing didn't come to my mind. I should have slept over this another night... Sure, that makes sense. But it doesn't explain why your initial code worked properly: why was that null terminator added? Was it by chance? Was it some kind of magic that means it will always work properly for type char? Or always work properly in whatever OS you're using? | ||
|
iaretehnoob
Sweden741 Posts
2) The most likely reason the original code worked properly is zero-initialization of memory in a debug build. Just in case it isn't obvious: The while loop doesn't check for '\0', it checks for 0. But it could also be some form of buffer overflow prevention or divine intervention. | ||
|
Hhanh00
34 Posts
On October 06 2016 22:36 beg wrote: I'm sorry for kicking off this discussion with so much disagreement ~ Anyway, I'm somewhat sure I figured it out now. char ca[] = "hey"; //null-terminator added! sizeof(ca) == 4; //true size(ca) == 4; //true But: char ca[] = { 'h', 'e', 'y' }; //no null-terminator! sizeof(ca) == 3; //true size(ca) == 3; //true This method of testing didn't come to my mind. I should have slept over this another night... Just to clarify a few things. "hey" is null terminated because it is a string constant. http://aboutc.weebly.com/string-constants.html 2) Every string constant ends up with a NULL character which is automatically assigned (before the closing double quotation mark) by the compiler. In the code below,
p is a pointer (not an array of char). so sizeof(p) is 8 on a 64 bit machine q is an array initialized to 'h', 'e', 'y', '\0' and sizeof(q) is 4. The size is automatically determined by the initializer list. r works too because it has a length equal to the sizeof("hey")-1. So r is non null terminated and sizeof(r) = 3 Now s is allowed too but will produce a warning (on gcc) because it is too short. You can also use a larger size and the array will be padded with 0. The argument of f is not an array but a pointer. So sizeof(r) is 8 when called as f(p) and f(q). q is converted to the pass the address of its first element. sizeof is a compile time expression and can't tell you the length of runtime string. It behaves very differently from strlen.
sizeof(t) = 8, wcslen(t) = 3 Finally, p is pointer to static storage. You can't change it. p[0] = 'a' is undefined (it will probably segfault because static data is in a protected segment). But q, r, s are char arrays and you can modify them. q[0] = 'a' is ok. Hope it helps | ||
|
toadtoad
9 Posts
I was wondering if anyone had done the opposite, so moving from a pure programming position to a IT OP/sysadmin position, if you did, how do you feel? do you feel like creating scripts and managing stuff less tiring? Do you regret your choice Thanks, | ||
|
mantequilla
Turkey781 Posts
I am trying to find out how I am supposed to realize this scenario: Clients visit my website, purchase my web app by filling out a form. Cloud does not play any role yet. Then, I need to deploy my app to azure and send customer an email containing the url of app, admin username/pass etc. required info. Or similarly, a client requests a trial of my web app, then I deploy a trial instance to azure. Question is how am I supposed to deploy an app to cloud programmatically (automatically) ? Azure has a web portal with a good interface, it is easy to deploy and configure things like tomcat, mysql etc manually. But I need to do that programmatically, since I can't deploy and configure every instance. I need to create tomcat, mysql and upload some configuration files and configure env variables: - powershell, there are various examples that use powershell scripts to manage azure services but they are 1) too cryptic, I don't know what a wall of commands do. 2) I need a windows server to run these and integrate them with my website so after purchase these commands can be run. And what happens something doesn't work? Powershell scripts can't debug, log etc.. - azure management rest api: the api is huge and finding out what parameters needs to be sent in what format, plus writing a custom rest client for this, is not an easy job - azure sdk, there are some examples using azure sdk to create resource groups etc. but I couldn't find how the heck can I create a tomcat or mysql containers and upload files to them... And I feel like java sdk does not cover all azure functionality, I may need to write a .net app for this (don't know if .net sdk is more complete) What is the usual way to perform such a task with cloud providers? Do I need a 3rd party tool, or is it a trivial thing that I don't know how to do? | ||
|
BluzMan
Russian Federation4235 Posts
On October 06 2016 21:47 tofucake wrote: It's not undefined behavior. The compiler automatically adds a null on the end of the array, and sizeof doesn't count it. It doesn't matter what optimization is done. If you want to create undefined behavior you would need to declare the array with malloc, then instantiate with your characters. The compiler won't automatically add a null in that case. You're wrong, it doesn't. It only automatically adds terminators to string literals, not arbitrary arrays of char. Undefined behaviour means literally "anything can happen" and since your code working as intended falls under "anything", it can happen too. But might change at any moment. Furthermore, compilers often use detectable cases of UB as a source of information (compilers assume that code is well-formed and might for example flag code that triggers UB as unreachable) which may result in completely magical and non-obvious code transformations. | ||
|
beg
991 Posts
On October 06 2016 22:54 Acrofales wrote: Sure, that makes sense. But it doesn't explain why your initial code worked properly: why was that null terminator added? Was it by chance? Was it some kind of magic that means it will always work properly for type char? Or always work properly in whatever OS you're using? Tested it on three different compilers, Win10 and Linux. The program always worked as if the char array was null terminated. So weird. @Hhanh00: I enjoyed going through your code as a mental excersise. Thanks for that. On a sidenote, I'm using C++, so the following examples were actually illegal (because "hey" is const char [4]) char r[3] = "hey"; @everyone else: Thanks for all the other contributions to my question. Not gonna quote everyone now, hehe. Gonna lay this case to rest and blame some hidden compiler magic (or divine intervention.... or simply undefined behaviour). | ||
|
Requizen
United States33802 Posts
I missed the GameMaker bundle on Humble Bundle last month, and while Clickteam Fusion seems like a much lesser tool, $15 seems like a good deal. I was just thinking of making some simple phone games as a hobby, would it be a good enough tool? Or is there something free out there that's worth checking out? | ||
| ||
