26

Re: BSD sockets in API browsers

Hello, fddima, you wrote: EP>> I told that it faster similar C# the code, instead of a C ++. F> Fejspalm to me. , late - blunted. . EP from a subject in a subject compares call cost  a method in pluses,  in web asm sorry, in asm.js. . (I.e. to JS it strictly speaking does not concern) and the delegate . To do from this global outputs quite seriously - trolling from a series "we measure operation with numbers in a python without numPy". I already time three hinted about "when performance is important, write a little in other style", but something does not help. It is pleasant to distort - a flag in hands, the main thing that remaining were not conducted.

27

Re: BSD sockets in API browsers

Hello, Sinix, you wrote: S> Uv. EP from a subject in a subject compares call cost  a method in pluses,  in web asm sorry, in asm.js. . (I.e. to JS it strictly speaking does not concern) and the delegate . To do from this global outputs quite seriously - trolling from a series "we measure operation with numbers in a python without numPy". S> I already time three hinted about "when performance is important, write a little in other style", but something does not help. It is pleasant to distort - a flag in hands, the main thing that remaining were not conducted. Though I also am skeptical about it - nevertheless true grain in its words is, in a context of it . In that big   a reluctance, I it raised to read last time. I completely agree that modern JS VM - rather suitable pieces. Another matter that.NET as a whole - on two heads above, is simple because bare VM very little to whom is necessary. And.NET NET both generates the code and fulfills it without any knee-bends, and has a heap of libraries, and GC it is better, and itself VM where is more powerful (at least at level of primitive types). Me here does not suffice in JS int64 - it is necessary manages in a place, than is (lines - if id). At me  in one of hot path shows that time is stupid to leave on ... Well there from std:: string it is transformed in any WTF:: String. And such pretty conversions on  only in  - happen 3 times, sometimes to copying, sometimes in addition and with conversion (utf8/utf16), naturally it pulls behind itself and -> as a result it is already bad. And irregular  lines in 0.5-4 mbytes, with such conversions - and performance turns in . Well anything, happens, and percents and  at all of us equally rubber. We will suffer while all  do not pass to single line, and it is obvious std:: string goes wood. But here a call  in a cycle. . Well, it is cool to be able , adaptive/speculative JIT, etc. On the other hand here here GCC HashTable shows the best results (the second, but the first benefits not for the account ) though it is a pure C and the comparator cannot be  in any way. So in my opinion - significance as a whole I think it is overestimated. Naturally somewhere it is very necessary, and in synthetics without it generally in any way it is impossible. Update: And in  for a long time-for a long time I was engaged both manual , and manual  algorithms with adaptation under different variations of the input data (type a line, , the pointer). Though who is better somehow to be able to tell to the compiler here.

28

Re: BSD sockets in API browsers

Hello, fddima, you wrote: F> I completely agree that modern JS VM - rather suitable pieces.  the Problem that productivity of this VM is proved through "we here threw a format which without problems is translated in the binary code, and about a miracle - it works as quickly, as the binary code!" , whatever one may do. Well, i.e. a problem not in most "Xxx faster", here all  just. And here with the argumentation oh. F> Update: And in  for a long time-for a long time I was engaged both manual , and manual  algorithms with adaptation under different variations of the input data (type a line, , the pointer). Though who is better somehow to be able to tell to the compiler here. Well. A heavy heritage of the code ground under an enterprise. As soon as it was a question of ordinarier things then in  MS research and realtime gc appeared, and translation in  normal and a heap of other buns. An ambush what "a prototype? A prototype" (in the main branch to drag there is on the order more difficult, than a prototype . For compatibility, abnormal clients and an other hell.

29

Re: BSD sockets in API browsers

Hello, Evgeny. Panasyuk, you wrote: EP> Without a difference, it can be and from the browser. Even I can build in this message img url which will to pull your local router from within networks. EP> restrictions are, unconditionally, nevertheless similar attacks are applied for a long time already, and are not a fantasy.  even do bruteforce passwords to routers. Well, for the sake of justice, it is vulnerability , and it does not prove that BSD sockets in the browser - good idea

30

Re: BSD sockets in API browsers

Hello, vsb, you wrote: F>> the Same  the client on  - already demands access to . vsb> Access to  at  is many years, File API is called. Well here try to write a file manager using the pure browser. You learn a lot of the new.

31

Re: BSD sockets in API browsers

Hello, Ikemefula, you wrote: F>>> the Same  the client on  - already demands access to . vsb>> Access to  at  is many years, File API is called. I> well here try to write a file manager using the pure browser. You learn a lot of the new. At what here a file manager and the torrent-client? For the torrent-client there suffices .

32

Re: BSD sockets in API browsers

Hello, vsb, you wrote: vsb>>> Access to  at  is many years, File API is called. I>> well here try to write a file manager using the pure browser. You learn a lot of the new. vsb> at what here a file manager and the torrent-client? For the torrent-client there suffices . ,  - and who to you, on one moment gives a file on  to a way? What browsers it are able? What to do with remaining?