quick syntax question
-
Hi all, Can someone let me know what is happening with the following line of code? I don't understand the syntax used in after j == (arr[i].Length - 1) Specifically what's happening with ? "" : I do know what " " is doing at the very end.
System.Console.Write("{0}{1}", arr[i][j], j == (arr[i].Length - 1) ? "" : " ");
-
What happens with the sequence points though? For the ternary operator there is a sequence point between the condition and the second/third operand (whichever it chooses), but AFAIK not after the second/third operand (unless you close it with a semicolon, of course) Or did MS sneakily add some sequence points there, like they did with method arguments?
Disclaimer: I don't know the finer details as I don't really care; with 20+ years of C under the belt (using all kinds of compilers) I have learned to stay well away from the borders. Defensive programming is what I call it. Besides, doing so improves readability and robustness of the code. But then I decided to browse the C# 3.0 spec anyway, and here are some findings: 1. the notion of "sequence point" does not appear in the C# standard. 2. "The order of evaluation of operators in an expression is determined by the precedence and associativity of the operators (§7.2.1)." 3. "If the array creation expression contains an array initializer, then each expression in the array initializer is evaluated and assigned to its corresponding array element. The evaluations and assignments are performed in the order the expressions are written in the array initializer" 4. "A method decorated with the Conditional attribute is a conditional method. The Conditional attribute indicates a condition by testing a conditional compilation symbol. Calls to a conditional method are either included or omitted depending on whether this symbol is defined at the point of the call. If the symbol is defined, the call is included; otherwise, the call (including evaluation of the parameters of the call) is omitted." I did not find: - a single thing about side-effects - evaluation order of method call arguments which does not mean it isn't covered, there are over 500 pages in the spec. Note: as 4.0 supports named arguments (and hence accepts them in any order), the problem potentially gets larger. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
modified on Monday, April 26, 2010 3:32 PM
-
Disclaimer: I don't know the finer details as I don't really care; with 20+ years of C under the belt (using all kinds of compilers) I have learned to stay well away from the borders. Defensive programming is what I call it. Besides, doing so improves readability and robustness of the code. But then I decided to browse the C# 3.0 spec anyway, and here are some findings: 1. the notion of "sequence point" does not appear in the C# standard. 2. "The order of evaluation of operators in an expression is determined by the precedence and associativity of the operators (§7.2.1)." 3. "If the array creation expression contains an array initializer, then each expression in the array initializer is evaluated and assigned to its corresponding array element. The evaluations and assignments are performed in the order the expressions are written in the array initializer" 4. "A method decorated with the Conditional attribute is a conditional method. The Conditional attribute indicates a condition by testing a conditional compilation symbol. Calls to a conditional method are either included or omitted depending on whether this symbol is defined at the point of the call. If the symbol is defined, the call is included; otherwise, the call (including evaluation of the parameters of the call) is omitted." I did not find: - a single thing about side-effects - evaluation order of method call arguments which does not mean it isn't covered, there are over 500 pages in the spec. Note: as 4.0 supports named arguments (and hence accepts them in any order), the problem potentially gets larger. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
modified on Monday, April 26, 2010 3:32 PM
I don't think there can be any side effects in C#, because objects (and everything is an object), are essentially static in nature. For example in C, an operation like x++?: could introduce some ambiguity, while in C#, x and x++ are distinct objects. Am I out of my depth here? Absolutely :-)
-
I don't think there can be any side effects in C#, because objects (and everything is an object), are essentially static in nature. For example in C, an operation like x++?: could introduce some ambiguity, while in C#, x and x++ are distinct objects. Am I out of my depth here? Absolutely :-)
There always is some potential for side-effects. First example:
int v = ...;
int y = (v++==100) ? 0 : v;will the auto-increment happen before or after the evaluation of the second/third operand? Second example:
int y = someFlag ? method1() : method2();
We all hope this causes exactly one method to get executed; in general it is not equivalent to:
int v1=method1();
int v2=method2();
int y = someFlag ? v1 : v2;as those methods could have all kinds of side-effects. However I did not see that stated explicitly in the spec. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
I don't think there can be any side effects in C#, because objects (and everything is an object), are essentially static in nature. For example in C, an operation like x++?: could introduce some ambiguity, while in C#, x and x++ are distinct objects. Am I out of my depth here? Absolutely :-)
Someone disliked your message enough to down-vote it. The topic is interesting enough IMO, albeit a little OT but that's OK. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
There always is some potential for side-effects. First example:
int v = ...;
int y = (v++==100) ? 0 : v;will the auto-increment happen before or after the evaluation of the second/third operand? Second example:
int y = someFlag ? method1() : method2();
We all hope this causes exactly one method to get executed; in general it is not equivalent to:
int v1=method1();
int v2=method2();
int y = someFlag ? v1 : v2;as those methods could have all kinds of side-effects. However I did not see that stated explicitly in the spec. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
Luc Pattyn wrote:
will the auto-increment happen before or after the evaluation of the second/third operand?
Neither, because the code won't compile. "int v = ...;" ;P
I know the language. I've read a book. - _Madmatt
-
Disclaimer: I don't know the finer details as I don't really care; with 20+ years of C under the belt (using all kinds of compilers) I have learned to stay well away from the borders. Defensive programming is what I call it. Besides, doing so improves readability and robustness of the code. But then I decided to browse the C# 3.0 spec anyway, and here are some findings: 1. the notion of "sequence point" does not appear in the C# standard. 2. "The order of evaluation of operators in an expression is determined by the precedence and associativity of the operators (§7.2.1)." 3. "If the array creation expression contains an array initializer, then each expression in the array initializer is evaluated and assigned to its corresponding array element. The evaluations and assignments are performed in the order the expressions are written in the array initializer" 4. "A method decorated with the Conditional attribute is a conditional method. The Conditional attribute indicates a condition by testing a conditional compilation symbol. Calls to a conditional method are either included or omitted depending on whether this symbol is defined at the point of the call. If the symbol is defined, the call is included; otherwise, the call (including evaluation of the parameters of the call) is omitted." I did not find: - a single thing about side-effects - evaluation order of method call arguments which does not mean it isn't covered, there are over 500 pages in the spec. Note: as 4.0 supports named arguments (and hence accepts them in any order), the problem potentially gets larger. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
modified on Monday, April 26, 2010 3:32 PM
Luc Pattyn wrote:
I have learned to stay well away from the borders. Defensive programming is what I call it.
Couldn't agree more, and liberal use of parentheses :-)
-
Luc Pattyn wrote:
will the auto-increment happen before or after the evaluation of the second/third operand?
Neither, because the code won't compile. "int v = ...;" ;P
I know the language. I've read a book. - _Madmatt
you're right, you'd better insert some expression that the compiler can't evaluate right away. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
Disclaimer: I don't know the finer details as I don't really care; with 20+ years of C under the belt (using all kinds of compilers) I have learned to stay well away from the borders. Defensive programming is what I call it. Besides, doing so improves readability and robustness of the code. But then I decided to browse the C# 3.0 spec anyway, and here are some findings: 1. the notion of "sequence point" does not appear in the C# standard. 2. "The order of evaluation of operators in an expression is determined by the precedence and associativity of the operators (§7.2.1)." 3. "If the array creation expression contains an array initializer, then each expression in the array initializer is evaluated and assigned to its corresponding array element. The evaluations and assignments are performed in the order the expressions are written in the array initializer" 4. "A method decorated with the Conditional attribute is a conditional method. The Conditional attribute indicates a condition by testing a conditional compilation symbol. Calls to a conditional method are either included or omitted depending on whether this symbol is defined at the point of the call. If the symbol is defined, the call is included; otherwise, the call (including evaluation of the parameters of the call) is omitted." I did not find: - a single thing about side-effects - evaluation order of method call arguments which does not mean it isn't covered, there are over 500 pages in the spec. Note: as 4.0 supports named arguments (and hence accepts them in any order), the problem potentially gets larger. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
modified on Monday, April 26, 2010 3:32 PM
I didn't find much, but I did find "The expressions of an argument list are always evaluated in the order they are written. " on page 182/553 of the C# 2 spec (4th edition / June 2006) I'm not sure whether that still applies to C# 4 - I'll download the new spec (I don't really know why I still had that old one) And in 14.2 (page 172) of the same document I found "The order in which operands in an expression are evaluated, is left to right. [Example: In F(i) + G(i++) * H(i), method F is called using the old value of i, then method G is called with the old value of i, and, finally, method H is called with the new value of i. This is separate from and unrelated to operator precedence. end example] " edit: IMO that [the first quote] supports my claim that MS had inserted sequence points in method argument evaluation, they're just not calling it that..
-
I didn't find much, but I did find "The expressions of an argument list are always evaluated in the order they are written. " on page 182/553 of the C# 2 spec (4th edition / June 2006) I'm not sure whether that still applies to C# 4 - I'll download the new spec (I don't really know why I still had that old one) And in 14.2 (page 172) of the same document I found "The order in which operands in an expression are evaluated, is left to right. [Example: In F(i) + G(i++) * H(i), method F is called using the old value of i, then method G is called with the old value of i, and, finally, method H is called with the new value of i. This is separate from and unrelated to operator precedence. end example] " edit: IMO that [the first quote] supports my claim that MS had inserted sequence points in method argument evaluation, they're just not calling it that..
Thanks Harold for "The expressions of an argument list are always evaluated in the order they are written"; that is what I knew had to be there, and I now found it in C# 3.0 too. I'm afraid there is some semantic ambiguity involved; to evaluate can either be "determine the value" or "process the expression yielding a value and execute all side-effects", and I'm pretty sure the latter is what was meant for argument lists.
harold aptroot wrote:
And in 14.2 (page 172)
the numbers clearly have changed since. I now found that sentence too, it is under 7.2; and it supports the broader meaning of "to evaluate". :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
Thanks Harold for "The expressions of an argument list are always evaluated in the order they are written"; that is what I knew had to be there, and I now found it in C# 3.0 too. I'm afraid there is some semantic ambiguity involved; to evaluate can either be "determine the value" or "process the expression yielding a value and execute all side-effects", and I'm pretty sure the latter is what was meant for argument lists.
harold aptroot wrote:
And in 14.2 (page 172)
the numbers clearly have changed since. I now found that sentence too, it is under 7.2; and it supports the broader meaning of "to evaluate". :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
The second meaning is also what's used by the compiler, so it's probably that :) But I can't really find any good "C# 4 language specification" documents, do you know one? Google didn't seem in the mood to give me useful results.. I found some drafts, but well, they're drafts.. or does that mean that the C# 4 specs aren't "official" yet?
modified on Monday, April 26, 2010 5:31 PM
-
The second meaning is also what's used by the compiler, so it's probably that :) But I can't really find any good "C# 4 language specification" documents, do you know one? Google didn't seem in the mood to give me useful results.. I found some drafts, but well, they're drafts.. or does that mean that the C# 4 specs aren't "official" yet?
modified on Monday, April 26, 2010 5:31 PM
This[^] is what Google is willing to offer; I see two candidates, nothing final though. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
The second meaning is also what's used by the compiler, so it's probably that :) But I can't really find any good "C# 4 language specification" documents, do you know one? Google didn't seem in the mood to give me useful results.. I found some drafts, but well, they're drafts.. or does that mean that the C# 4 specs aren't "official" yet?
modified on Monday, April 26, 2010 5:31 PM
harold aptroot wrote:
the C# 4 specs aren't "official" yet?
yes, I expect that is part of the game: propose a standard, get it accepted, then propose extensions and implement them faster than anyone else can, giving you de facto control. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
This[^] is what Google is willing to offer; I see two candidates, nothing final though. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
Google is NOT your friend :-) http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=dfbf523c-f98c-4804-afbd-459e846b268e[^] But I haven't got the impression that it's different from 3.0 regarding the ternary operator.
-
Google is NOT your friend :-) http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=dfbf523c-f98c-4804-afbd-459e846b268e[^] But I haven't got the impression that it's different from 3.0 regarding the ternary operator.
Thanks Michel. I have been expecting Bing to become better at finding MS stuff. They wouldn't dare putting in some meta tags to fool other search engines, would they?
Michel Godfroid wrote:
Google is NOT your friend
Yes he is, he is my best mate; he knows almost everything I don't know (and that is a lot). Can't do without him. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.
-
The second meaning is also what's used by the compiler, so it's probably that :) But I can't really find any good "C# 4 language specification" documents, do you know one? Google didn't seem in the mood to give me useful results.. I found some drafts, but well, they're drafts.. or does that mean that the C# 4 specs aren't "official" yet?
modified on Monday, April 26, 2010 5:31 PM
The sentence "The expressions of an argument list are always evaluated in the order they are written" is still present in the 4.0 document Michel provided a link for. And it now has an example with named parameters. Great. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
Prolific encyclopedia fixture proof-reader browser patron addict?
We all depend on the beast below.