1

Topic: call to arms

2

Re: call to arms

Hello, Arsen. Shnurkov, you wrote: AS> Well here,  I ask to write article on normal regular expressions, instead of on defective, as now. AS> the English-speaking world made it in 1964, it is time to reduce lag a little. As the person which is able unite written engine  N  by means of any N' a boolean function I can to tell that: 1) the Intersection of automatic machines is not necessary. I do not know any vital example. 2) and negation is frankly harmful. At people a roof takes down. I at first added it, and then removed. For people were confused also delirium wrote. 3) it is more useful in practice there was a subtraction of automatic machines. In expression NotEscs from all set not blank lines are subtracted all the line long which contain inverted commas, the character of the beginning of special sequence and transfers of lines. [SpanClass (String)] token StringLiteral1 = Quote StringPart* Quote {regex Quote = ' \"'; regex Esc = ' \\' (Quote | EscChar); regex Escs = Esc +; regex NotEscs = Any + - (Any* (Quote | ' \\' | NewLine) Any *); token StringPart {| Escs; | NotEscs;}}... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

3

Re: call to arms

WH> written engine  which is able to unite N  by means of any N' a boolean function Many people made surprising achievements in the past, however articles educational they did not write also anything from them does not remain. Here and your achievements help nothing to me. And article would help. WH> it is useful in practice there was a subtraction of automatic machines see, it appeared it is useful in practice, but is accessible only to elite instead of being taught everywhere.

4

Re: call to arms

Hello, Arsen. Shnurkov, you wrote: AS> And still there are the regular expressions to intersection operations ^ and negations! And still happen to review forward-back and with conditions))

5

Re: call to arms

AS>> And still there are the regular expressions to intersection operations ^ and negations! S> and still happen to review forward-back and with conditions)) Specially for you I will tell that else happen context-free grammar. You after all for certain do not know it also to you it will be interesting! And I know! I know! Also I burn with the desire to educate you. Here what I clever!

6

Re: call to arms

Hello, Sheridan, you wrote: AS>> And still there are the regular expressions to intersection operations ^ and negations! S> and still happen to review forward-back and to conditions)) do not happen. Simply on determination of the regular expressions.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

7

Re: call to arms

Hello, WolfHound, you wrote: AS>>> And still there are the regular expressions to intersection operations ^ and negations! S>> and still happen to review forward-back and to conditions)) WH> do not happen. Simply on determination of the regular expressions. Aha

8

Re: call to arms

Hello, Sheridan, you wrote: AS>>>> And still there are the regular expressions to intersection operations ^ and negations! S>>> and still happen to review forward-back and to conditions)) WH>> do not happen. Simply on determination of the regular expressions. S> aha that you show, concerns various types so-called _extended_ regular expressions which included a heap of features which in "normal" regular expressions are not found. For example, there there is a possibility to refer on earlier  through \n, where n - number. Introduction of these features forced down concept of the regular expressions to grammar, which in any way regular on classics like Homsky. I do not know, to whom here to prefer in this dispute, especially after colleague Wolfhound was passed by the supertank on all theorists of a subject  - if "" was softest of terms, also hierarchy Homsky for it in any way the sample. But anyway dispute not as a matter of fact.

9

Re: call to arms

Hello, netch80, you wrote: N> Introduction of these features forced down concept of the regular expressions to grammar, which in any way regular on classics like Homsky. And what in this point in question there are other classics? N> I do not know, to whom here to prefer in this dispute, especially after colleague Wolfhound was passed by the supertank on all theorists of a subject  - if "" was softest of terms, also hierarchy Homsky for it in any way the sample. But anyway dispute not as a matter of fact. You generally understood nothing from this that I spoke. My claim to the book of a dragon that there a heap of poor and at the same time difficult algorithms. Also is not present simple and qualitative. For on them it is impossible to write articles to scientific logs.  there all is simple. For example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. And it thus that is the best algorithm for analysis of programming languages. Actually,  also it is constructed around TDOP poured by sugar. ... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

10

Re: call to arms

Hello, WolfHound, you wrote: WH> My claim to the book of a dragon that there a heap of poor and at the same time difficult algorithms. There is there no heap of algorithms, you did not read any of 3 "dragons", certainly. This book about principles of creation of compilers, i.e. about their architecture, instead of about specific algorithms. One of the most important heads in each issuing of "dragon" is about syntactically controlled translation. It simply principle and all book it is written for the sake of this chapter, . And in kach-ve algorithms of a lexer and a parcer the most general algorithms in minimum them kol-ve are taken. WH> Also is not present simple and qualitative. At Aho it is full of publications on specific algorithms on which all now lives modern IT. In general, flies separately, cutlets separately. WH> for example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. In this book it is a lot of still that is not present, because a dragon not about it it is perfect.

11

Re: call to arms

Hello, vdimas, you wrote: WH>> For example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. V> in this book it is a lot of still that is not present, because a dragon not about it it is perfect. There is no the most practical algorithm of analysis of programming languages. Hence, it is the book not about that how to write compilers, and about what the hell. Actually, because of it it is impossible to take a dragon and to write on it the normal compiler.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

12

Re: call to arms

Hello, WolfHound, you wrote: WH>>> For example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. V>> in this book it is a lot of still that is not present, because a dragon not about it it is perfect. WH> there is no the most practical algorithm of analysis of programming languages. Hence, it is the book not about that how to write compilers, and about what the hell. Actually, because of it it is impossible to take a dragon and to write on it the normal compiler.  as you look in real examples (something actively used) - so where spit, everywhere  a descending parcer. In the book of a dragon such parcer is, chapter 2. Means, to write on it the normal compiler it is possible Q.E.D. And concerning TDOP is, of course, very tasty variant, but as a matter of fact it only the event-driven is tabular-organized representation of a descending recursive parcer "it is necessary to generalize the analysis code From idea of a type, reducing a choice  in the table" it is deduced simply on one . I do not belittle Pratt, for 1973 idea ingenious, but for now anything special in it is not present. From the third party, after - normally to describe it would occupy page three - at a thickness of all book it is tiny expenses so I there  would enter it. Not for the general idea of the book, but for those who wants to test there and then in the beginning all theory practice and only then to reflect.

13

Re: call to arms

Hello, WolfHound, you wrote: WH>>> For example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. V>> in this book it is a lot of still that is not present, because a dragon not about it it is perfect. WH> there is no the most practical algorithm of analysis of programming languages. Hence, it is the book not about that how to write compilers, and about what the hell. Actually, because of it it is impossible to take a dragon and to write on it the normal compiler. I  that you rest. The dragon does not forbid to take any algorithm lexical or parse and to write the compiler. Because this book not about algorithms of parsing , and about application of these algorithms in . It is badly believed that this basic difference you do not see in an emphasis.

14

Re: call to arms

Hello, netch80, you wrote: N> Chtoj-to as you look in real examples (something actively used) - so where spit, everywhere  a descending parcer. And on that that about TDOP anybody at all did not hear it. Simply compare the size of article in Wikipedia. https://en.wikipedia.org/wiki/Pratt_parser https://en.wikipedia.org/wiki/LALR_parser and it thus that LALR is useless. And all on that that about LALR is written in the book of a dragon. And about the book of a dragon the myth is strenuously spread that there is written about all that it is necessary to know about creation of compilers. N> in the book of a dragon such parcer is, chapter 2. N> Means, to write on it the normal compiler it is possible N> Q.E.D. On  too it is possible to write any program.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

15

Re: call to arms

Hello, WolfHound, you wrote: N>> Chtoj-to as you look in real examples (something actively used) - so where spit, everywhere  a descending parcer. WH> and on that that about TDOP anybody at all did not hear it. WH> simply compare the size of article in Wikipedia. WH> https://en.wikipedia.org/wiki/Pratt_parser WH> https://en.wikipedia.org/wiki/LALR_parser WH> and it thus that LALR is useless. $ find/usr/src/| grep '\.y $ ' | fgrep-v/test/| wc-l 84 It freebsd base sources. Thus that there not bison, and byacc, and at it pure LALR (1) without variants. So who to whom is more useless? With LALR we have the general-purpose mechanism, suitable for implementation and attending, with a minimum of mental nestings in a grammar subject (the author of the code writes as a matter of fact only the ), and also built-in check of a correctness of grammar (that crucial for those who) Only it is not necessary about  to recall - outside of your sandbox about it is not audible. WH> and all on that that about LALR is written in the book of a dragon. And about the book of a dragon the myth is strenuously spread that there is written about all that it is necessary to know about creation of compilers. So struggle with this myth. That to certain level write on LALR (1)-bound means, and then suddenly sharply pass on self-peep. Perhaps to that there are serious reasons? Or at least lay out (in two languages) beautiful description TDOP without all these null denotation which dement at the first reading. (I know that such articles already are. But it is not enough of them.) In your conditions you could break easily it  articles on . Cheaply and effectively. And that now even for classical TDOP is only simple dry transfer of Krokforda. N>> In the book of a dragon such parcer is, chapter 2. N>> Means, to write on it the normal compiler it is possible N>> Q.E.D. WH> On  too it is possible to write any program. On  do not write industrial compilers. And on manual their self-peep full (I will not recall a C ++ with its cockroaches in the half-sky, but C# and Go are already indicative). To you, in my opinion, all already said that matter is not in the academic snobbery or in something similar, but your development it is necessary to shift on something more suitable to download-and-go.

16

Re: call to arms

Hello, WolfHound, you wrote: WH>>> For example, where in the book of dragon Top down operator precedence? And after all it is operation of 1973. V>> in this book it is a lot of still that is not present, because a dragon not about it it is perfect. WH> there is no the most practical algorithm of analysis of programming languages. Hence, it is the book not about that how to write compilers, and about what the hell. Actually, because of it it is impossible to take a dragon and to write on it the normal compiler. What does "most practical" mean? You in it put what sense?

17

Re: call to arms

Hello, netch80, you wrote: N> With LALR we have the general-purpose mechanism, suitable for implementation and attending, with a minimum of mental nestings in a grammar subject (the author of the code writes as a matter of fact only the ), and also built-in check of a correctness of grammar (that crucial for those who) It is simple not truth. There a step aside and greetings.  grammar to rewrite. And that it to rewrite it is necessary to know, how this most LALR works. N> so struggle with this myth. That to certain level write on LALR (1)-bound means, and then suddenly sharply pass on self-peep. Perhaps to that there are serious reasons? So at all institutes to all students this LALR . Here people also think that anything else is not present. But as soon as people come far enough they understand what LALR a shit. Also are forced to start to write a parcer hands. And the only thing that it is possible to write stupid recursive descent by hands if not to know about TDOP. N> to You, in my opinion, all already said that matter is not in the academic snobbery or in something similar, but your development it is necessary to shift on something more suitable to download-and-go. I here at all about my development speak. I speak about algorithm which is higher me.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

18

Re: call to arms

Hello, WolfHound, you wrote: N>> Chtoj-to as you look in real examples (something actively used) - so where spit, everywhere  a descending parcer. WH> and on that that about TDOP anybody at all did not hear it. WH> simply compare the size of article in Wikipedia. WH> https://en.wikipedia.org/wiki/Pratt_parser WH> https://en.wikipedia.org/wiki/LALR_parser WH> and it thus that LALR is useless. WH> and all on that that about LALR is written in the book of a dragon. And about the book of a dragon the myth is strenuously spread that there is written about all that it is necessary to know about creation of compilers. Same at its finest . At you "this such general-purpose explanation for everything with what you do not agree is written in the book of a dragon". Such plot against you and TDOP from citizen Aho. We tell, I began without any dragon - esteemed Virta. , on yours, too a dragon ?

19

Re: call to arms

Hello, Ikemefula, you wrote: I> Same at its finest . At you "this such general-purpose explanation for everything with what you do not agree is written in the book of a dragon". On the contrary. I say that the book of a dragon bad on that that that there is written with what I do not agree. And that is even worse that its all  as the sample for imitation.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

20

Re: call to arms

Hello, WolfHound, you wrote: N>> With LALR we have the general-purpose mechanism, suitable for implementation and attending, with a minimum of mental nestings in a grammar subject (the author of the code writes as a matter of fact only the ), and also built-in check of a correctness of grammar (that crucial for those who) WH> It is simple not truth. WH> there a step aside and greetings.  grammar to rewrite. And that it to rewrite it is necessary to know, how this most LALR works. That on it to make grammar, it is necessary half-ten simple rules. Certainly, if not to try to put a swan, a cancer and a pike as made in a C ++. From 84 majority of these named me  are simple, as a house corner. And, thus, it is total are declarative that is faster in plus - at least because TDOP rigidly anchors the parsing code, and BNF-like declarations with the code hung up to them was general-purpose lay down under any generator of a parcer. And here that would be useful, so it is approaches how not to run at the grammar extension into the same pose where the C ++ got. But TDOP it too does not solve, and Nitra does not solve. N>> so struggle with this myth. That to certain level write on LALR (1)-bound means, and then suddenly sharply pass on self-peep. Perhaps to that there are serious reasons? WH> so at all institutes to all students this LALR . Here people also think that anything else is not present." At all institutes """simultaneously LL and LR. With a different dial-up of specializations, clearly, but both give. WH> but as soon as people come far enough they understand what LALR a shit. WH> also are forced to start to write a parcer hands. And the only thing that it is possible to write stupid recursive descent by hands if not to know about TDOP. I know about TDOP though never earned writing of compilers. And those who seriously was engaged, are capable to fan at least the list of methods in Wikipedia. Therefore all to whom it is necessary, for a long time about it know. And if do not use, the reason not in LALR. N>> to You, in my opinion, all already said that matter is not in the academic snobbery or in something similar, but your development it is necessary to shift on something more suitable to download-and-go. WH> I here at all about my development speak. I speak about algorithm which is higher me. OK, it is possible to reduce a subject only to TDOP. But in that case it strongly suffers. How much I had time to see in , at you  the descriptive approach, and TDOP demands hands to write out the code of all nud/led/etc.

21

Re: call to arms

Hello, WolfHound, you wrote: I>> Same at its finest . At you "this such general-purpose explanation for everything with what you do not agree is written in the book of a dragon". WH> On the contrary. I say that the book of a dragon bad on that that that there is written with what I do not agree. __ _ _, what with what you do not agree? There it is explicitly told "we to you presented the best world development, do not think anywhere more to look"? Something I there such anywhere did not see. Name page number. WH> and that is even worse that its all  as the sample for imitation. It is other question (if it generally so), feeblly connected to the content of the book. But you can confirm, what exactly "" as the sample? And all"? You did not inflate a fly to an elephant?

22

Re: call to arms

Hello, netch80, you wrote: N> It is other question (if it generally so), feeblly connected to the content of the book. But you can confirm, what exactly "" as the sample? And all"? N> You did not inflate a fly to an elephant? When the person asks how to write the compiler the first that offer it the book of a dragon. And on this message the heap of the positive estimations is always formed. This book even in films mention.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

23

Re: call to arms

Hello, netch80, you wrote: N> That on it to make grammar, it is necessary half-ten simple rules. Certainly, if not to try to put a swan, a cancer and a pike as made in a C ++. In with ++ from the point of view of a parcer anything difficult is not present. There complexities with typification begin. N> from 84 majority of these named me  are simple, as a house corner. And, thus, it is total are declarative that is faster in plus - at least because TDOP rigidly anchors the parsing code, and BNF-like declarations with the code hung up to them was general-purpose lay down under any generator of a parcer. It again not truth. 1) TDOP happens quite declarative. Nitra perfectly shows it. 2) all generators  have different syntax. Also that different algorithms of analysis are even worse. It is possible so to transfer only very simple grammar. If the grammar is hardly more difficult, about transfer without a serious rewriting cannot be and speeches. An exception only  parcers. But all of them as one brakes. N> and here that would be useful, so it is approaches how not to run at the grammar extension into the same pose where the C ++ got. But TDOP it too does not solve, and Nitra does not solve. In what was specific a problem? N> "at all institutes" "" simultaneously LL and LR. With a different dial-up of specializations, clearly, but both give. A shit cow and a shit pork. Smell differently. But an essence one. N> I know about TDOP though never earned writing of compilers. And those who seriously was engaged, are capable to fan at least the list of methods in Wikipedia. Therefore all to whom it is necessary, for a long time about it know. And if do not use, the reason not in LALR. And whence you about it learned? N> OK, it is possible to reduce a subject only to TDOP. But in that case it strongly suffers. How much I had time to see in , at you  the descriptive approach, and TDOP demands hands to write out the code of all nud/led/etc. He does not demand anything. Here unpretentious hybrid PEG' and TDOP'. http://rsdn.org/forum/dotnet/6700752.1 the Author: WolfHound Date: 16.02.17 static void Main (string [] args) {var expr = new ChoiceParser (); expr. Add (new IntParser ()); expr. Add ("x"); expr. Add ("(", expr,")"); expr. Add ("+", expr. Get (1000)); expr. Add ("-", expr. Get (1000)); expr. Add (expr. Get (10), "+", expr. Get (10)); expr. Add (expr. Get (10), "-", expr. Get (10)); expr. Add (expr. Get (20), "*", expr. Get (20)); expr. Add (expr. Get (20), "/", expr. Get (20)); foreach (string s in new string [] {"*x", "+x", "x", "x +-x","x - + x", "x+x", "x*123-123/x", "x * (123-123)/x", "x*123-x12", "x*x - (x+x)","x*x + (x +

24

Re: call to arms

Hello, WolfHound, you wrote: N>> It is other question (if it generally so), feeblly connected to the content of the book. But you can confirm, what exactly "" as the sample? And all"? N>> You did not inflate a fly to an elephant? WH> when the person asks how to write the compiler the first that offer it the book of a dragon. Well and? As the high-level overview it quite not bad goes. Though also alternatives full. WH> and on this message the heap of the positive estimations is always formed. WH> this book even in films mention. That is your claim what at it the self-supported reputation because once quitted in a top? Or  what the reading decide, what in it how in the Koran, there is all?

25

Re: call to arms

Hello, WolfHound, you wrote: N>> That on it to make grammar, it is necessary half-ten simple rules. Certainly, if not to try to put a swan, a cancer and a pike as made in a C ++. WH> In with ++ from the point of view of a parcer anything difficult is not present. WH> There complexities with typification begin. When it is impossible to understand, something is function declaration or a variable, yet will not disassemble up to the end and will not read a context (the part from which is set already after this disputable line) is not "complexities with typification", it is parsing. N>> from 84 majority of these named me  are simple, as a house corner. And, thus, it is total are declarative that is faster in plus - at least because TDOP rigidly anchors the parsing code, and BNF-like declarations with the code hung up to them was general-purpose lay down under any generator of a parcer. WH> it again not truth. WH> 1) TDOP happens quite declarative. Nitra perfectly shows it. OK, I will agree. I know implementations TDOP under the text of grammar or on tamping of rules of the general-purpose type (like such), but did not anchor them to Pratt's original method because many consider that TDOP == Pratt and variants cannot be. We converge that Pratt's this extension. WH> 2) all generators  have different syntax. Almost all of them from the most popular classes are different codings of abstract BNF and the controlled rules of language of implementation added to them. Transfer from one on another becomes type changeover ":" On "=" and so forth. If you about possibility, for example, to collect in one rule some variants with priorities in between they are eliminated from similar possibilities - a certain general basic dial-up of possibilities undertakes. WH> also that is even worse different algorithms of analysis. It is possible so to transfer only very simple grammar. If the grammar is hardly more difficult, about transfer without a serious rewriting cannot be and speeches. WH> an exception only  parcers. But all of them as one brakes. Who at you these general? To steam of names at least. N>> and here that would be useful, so it is approaches how not to run at the grammar extension into the same pose where the C ++ got. But TDOP it too does not solve, and Nitra does not solve. WH> in what was specific a problem? In how to predict convenient extensibility of grammar without interpretation conflicts. At the time of the Fortran could not think of by what it will be enveloped them "integer zz". N>> I know about TDOP though never earned writing of compilers. And those who seriously was engaged, are capable to fan at least the list of methods in Wikipedia. Therefore all to whom it is necessary, for a long time about it know. And if do not use, the reason not in LALR. WH> And whence you about it learned? Simply went to read see also on pages. It is accessible to any student. Only earlier for this purpose it was necessary to quit the house. N>> OK, it is possible to reduce a subject only to TDOP. But in that case it strongly suffers. How much I had time to see in , at you  the descriptive approach, and TDOP demands hands to write out the code of all nud/led/etc. WH> he does not demand Anything. See above. WH> here unpretentious hybrid PEG' and TDOP'. Funny. Coding style is not pleasant to me, but the idea is clear. Anyway to an initial question about the book it is relations has no, well except for a long time already told - to add in the book three pages on TDOP would not worsen, but also did not refine. It would be possible as to add at level "on three pages" on each point, the book could increase easily in few times in thickness. They and so cut on live, it seems to me.