51

Re: Than to explain popularity ?

Hello, DenisCh, you wrote: DC> Hello, Serginio1, you wrote: S>> Here When 1 will be on "odoo on a python" then it is possible and to try. DC> and this an ode is able more than one  ? She is able from a box of a web client and the mobile client? And that on  it is not clear.... The System works on postgres only. Naturally, it is possible to be connected to exterior bases by means a python. The Web client from a box is, and it is the core. The Desktop-client is not present, the mobile client is for andriod and ios.

52

Re: Than to explain popularity ?

Hello, alex_public, you wrote: _> There is quite a uniquely identifying strict (it strong) typifications. The last some years it is not so unambiguous. Probably, once  with transfer, now there are questions then, there.

53

Re: Than to explain popularity ?

Hello, neFormal, you wrote: F> why the author himself does not launch the tests if he writes them? Launches, certainly! Only than it helps? I think, you and can easily describe a variant when all breaks to . How any software is written? Wrote, debugged, handed over. In two months something is necessary  - the developer from other command adds new parameter to a method. The tests of sobriquets - all works. The nearest code viewed - like everywhere changed, where it is necessary.  - , fell off twenty seven tests, written by other command three weeks ago. Because there that method too is caused, but VSCode/PyCharm it does not show. When over different parts of the project tens people from different commands without checks at a compilation stage it becomes difficult work. Well and classics who breaks  generally in any language - "race condition in GIT". Two developers changed two different pieces of the code, depending from each other. Everyone on-separateness checked up , transited  the code and received all signatures on  the code. Both send the code in GIT and  collapses, therefore as on-separateness code pieces correct, but together give a bug.

54

Re: Than to explain popularity ?

Hello, alex_public, you wrote: _> There is quite a uniquely identifying strict (it strong) typifications. Can result this determination? And that Wikipedia tells "terms are not unambiguously treated".

55

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: F>> why the author himself does not launch the tests if he writes them? AK> launches, certainly! Only than it helps? I think, you and can easily describe a variant when all breaks to . Because at me it did not turn out to invent, I and ask. I do not understand a situation. If it is integration tests why falls at start from other tests? They should not be intersected.

56

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: it seems To me shared problems of development of the big projects are described by the big command. Really at static typification do not break , do not spoil the friend the friend pull request and NullPointerException then does not appear in ? Yes, the compiler a part of problems to catch at a compilation stage. Depends, of course, on the project and the compiler, but project recompilation what is the time can occupy that all to check up? How much it is comparable with starts of tests? Though yes, in dynamic language of change of the interface is a cheerful piece, it is necessary to know how to cook

57

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: AK> we wait for 5 hours It in one process or with multisequencing of tests?

58

Re: Than to explain popularity ?

S> As it is known, in the bad languages to write much more difficult. But why that the bad languages win love of audience. S> in what the reason? Here it in any way I can not understand. Language it always a member of equation look at remaining libraries, a runtime environment, support of this environment by indirect services, a development environment, the price of developers, availability of developers etc., etc. look with reference to your task, instead of to the Universe as a whole

59

Re: Than to explain popularity ?

Hello, Shmj, you wrote: S> In what the reason? Here it in any way I can not understand. Formation low level . S> to Take the same JS, PHP or Python even. Well JS that is clear, though and here stars of mankind which install Web-standards, could arrive on mind. An input threshold. In JS and PHP it is easy to come and make something working, , on a knee and to sell, and receiving the first 1000$ to shout at every turn that it the great businessman and JS/PHP it is the best. All more or less serious they to implement cannot any more, and will not be, and  about scalability do not prove to be true  experience. S> to me the logic of this world and all here here is not clear. To themselves cactuses invent. Let invent. The more fans  there where it is not necessary, the more expensively service of the pro in alteration of this everything, principal that usage  did not type . Mass.

60

Re: Than to explain popularity ?

Hello, Kernan, you wrote: K>  about scalability do not prove to be true  experience. Tell about the real experience.

61

Re: Than to explain popularity ?

Hello, Wolverrum, you wrote: W> using System; W> using System. Linq; W> using System. IO; W> using System. Collections. Generic; Using' it is not necessary to quote. Them generally rarely read. To them pay attention only at adding/removal not to drag the unnecessary. W> class A {W> static void Main () {it too. It is simply wrapping construction which the eye which have got used to the code at all does not note. W> at other equal I will prefer . Because of more short and . I do not know that such , but "short" I in the example resulted by you do not see. The code on a pearl on volume about same.

62

Re: Than to explain popularity ?

Hello, msorc, you wrote: M> Yes, the compiler a part of problems to catch at a compilation stage. Depends, of course, on the project and the compiler, but project recompilation what is the time can occupy that all to check up? Much less than tests chase. Even if  all OS.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

63

Re: Than to explain popularity ?

Hello, CreatorCray, you wrote: CC> Hello, msorc, you wrote: M>> Yes, the compiler a part of problems to catch at a compilation stage. Depends, of course, on the project and the compiler, but project recompilation what is the time can occupy that all to check up? CC> much less than tests chase. CC> even if  all OS. Well it is admissible. And for statically  language it is not necessary to launch tests absolutely? And to recompile too?

64

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: _>> There is quite a uniquely identifying strict (it strong) typifications. AK> can result this determination? And that Wikipedia tells "terms are not unambiguously treated". It you where looked? I here see here https://en.wikipedia.org/wiki/Strong_and_weak_typing quite unambiguous statement about the Python: "Smalltalk, Perl, Ruby, Python, and Self are all" strongly typed "in the sense that typing errors are prevented at runtime and they do little implicit type conversion".

65

Re: Than to explain popularity ?

Hello, neFormal, you wrote: F> because at me it did not turn out to invent, I and ask. F> I do not understand a situation. If it is integration tests why falls at start from other tests? They should not be intersected. The lion's part of test code normally goes in the form of libraries, that not  a bicycle on five times in a year. Here those general libraries also are intersected. The test at us generally goes not in the form of the code on the Python, and in the form of small files of scenarios in format Yaml where actions and executable checks are enumerated. Probably, to testers so it is more convenient. Those Yaml-files  the topology of a network specified in a configuration forms by the Python, the surrounding demanded for the test and further already is adjusted, one for another, instructions of the test scenario are carried out. At us the typical test in free retelling on the Russian would sound approximately so - "Two virtual networks to unite in one logical network, to check up that  VM1.1 from network VNet1 can incorporate to REST-service on  VM2.1 in network VNet2. After that to block connection  VM2.1 with , to be convinced that  it is sent in quarantine and it is inaccessible to others ".

66

Re: Than to explain popularity ?

Hello, msorc, you wrote: M> it seems To me shared problems of development of the big projects are described by the big command. Not, here those errors of types and signatures of a call of methods are characteristic only for  languages. I do not claim for a role of the expert on the Python, I look, how with it others and while something I did not see the scenarios of operation protecting or at least reducing to minimum here these jambs work. But I, will repeat, while watched usage of the Python only on auxiliary tasks like integration tests. I yet did not work over projects where the prodakshn-code would be on the Python. Though I watch for a long time JS in  and I see there all same problems. For comparing I can remember the operation in General Electric Healthcare where test  has been constructed on the basis of SpecFlow, with the code on C#. There such problems with the code of test components was not. M> Really at static typification do not break , do not spoil the friend the friend pull request and NullPointerException then does not appear in ? All it is. NullPointerException in  takes off, but it already problems of the following step, a problem of a correctness of algorithm. They and in the Python are. And here problems with a wrong amount or type of arguments are not present - they are caught by the compiler.  break, but there is it on the order less often, than test code breakages on the Python. As though at all on two orders is more rare. Pull-rekvesty do not spoil, but quite often it is necessary to do manual  if someone already changed the same file. I do not say that development in languages with static typification is problemless process. Jambs and bugs are everywhere and always. But the compiler of language with static typification gives check of types at an assembly stage, than already removes set of potential problems. Theoretically, each error which to me deduces the compiler, would demand start of tests, waiting and  dens in a case with dynamic typification. Even if to assume that in this case I would spend more time more attentively to check up the changes and such errors would meet three times less often, it all the same would be, on the average, one bug in day. I do not pose as the expert on dynamic languages. I look, how others work with the Python and JS and I see, as professional programmers in these languages continually attack this rake. Integration tests at us break on two times in the course of the day, and  in the web-interface we watch in flow of all time of development of a web muzzle. Thus that there at us Angular2, i.e. TypeScript, instead of bare JavaScript. M> Yes, the compiler a part of problems to catch at a compilation stage. Depends, of course, on the project and the compiler, but project recompilation what is the time can occupy that all to check up? It is how much comparable with starts of tests? In a case to our integration tests - it is absolutely non-comparable. I cannot tell, compilation our test  what is the time would occupy, it be written in compiled language. But the assembly of two main services over which I work, occupies 1 minute and 3 minutes. Units-tests demand even 2 minutes for the first service and 7 minutes for the second. Preparation for start of our integration tests (i.e. yet test code, and development of Docker-containers, surrounding adjustment, registration of components) occupies minutes 30. It is a separate song - I since autumn butt our testers, I ask to alter all it to divide into two actions - separately , separately start of tests. They now gradually work over it, but meanwhile is, that is - simply to launch one test for the Python occupies 30 minutes on , plus 2. 5 minutes on the test. The complete set of tests for our subsystem is fulfilled hours for 5 - once 30 minutes  a surrounding, then are already launched all test set. As well as follows from meanness laws, breaks normally somewhere in the end. M> though yes, in dynamic language of change of the interface is a cheerful piece, it is necessary to know how to cook Well here yes. On my observations, it is the main source of problems. At least, there, where I saw.

67

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: AK> The test at us generally goes not in the form of the code on the Python, and in the form of small files of scenarios in format Yaml where actions and executable checks are enumerated. Probably, to testers so it is more convenient. Those Yaml-files  the topology of a network specified in a configuration forms by the Python, the surrounding demanded for the test and further already is adjusted, one for another, instructions of the test scenario are carried out. AK> at us the typical test in free retelling on the Russian would sound approximately so - "Two virtual networks to unite in one logical network, to check up that  VM1.1 from network VNet1 can incorporate to REST-service on  VM2.1 in network VNet2. After that to block connection  VM2.1 with , to be convinced that  it is sent in quarantine and it is inaccessible to others ". And, clearly. It is possible to impose it with tests

68

Re: Than to explain popularity ?

Hello, msorc, you wrote: AK>> we wait for 5 hours M> It in one process or with multisequencing of tests? I can not precisely tell. I suspect that in one. Integration tests cannot be launched simultaneously - they check system operation as a whole, instead of separate components. If two tests change parallely system adjustments at us there porridge turns out. There is, however, one thing which can be . Tests are launched on different configurations. Tests are fixed for one or more "topology".  unrolls supported topology one for another, launching complete set of tests which are supported for each topology. At check  at us two-three basic topology is used. It would be possible to tear all topology simultaneously and to launch on them tests parallely. Can be and becomes - it would be necessary to ask testers.

69

Re: Than to explain popularity ?

Hello, Artem Korneev, you wrote: Me normally suffices at interface change to walk by simple search, to look, where the changed method is caused. Tried  the jambs, the compiler I do not think that would rescue. From colleagues for last couple of years was time or two as you speak, changed in library the interface and there were problems in other place. But I think, small experience there is faster. They came to language recently. In last time sometimes began to catch itself on thoughts that here static typification would help. Because it is necessary to write sometimes superfluous checks and to cover with place tests where typeful all would be easier. Still it would be desirable macroes

70

Re: Than to explain popularity ?

Hello, msorc, you wrote: M> In last time sometimes began to catch itself on thoughts that here static typification would help. Some even invent any game https://medium.com/byteconf/stripe-is-b … cd7cee6abf

71

Re: Than to explain popularity ?

Hello, kov_serg, you wrote: _> Therefore as they are capable to surprise unpleasantly: With ++ can surprise not less.

72

Re: Than to explain popularity ?

Hello, Shmj, you wrote: S> As it is known, in the bad languages to write much more difficult. But why that the bad languages win love of audience. Why they bad? Probably, the got accustomed widespread decisions it is simple : balance, standards and , a way of the least harm. I think, advantages php, python and js for beginners - a low threshold of entrance. And for skilled coders that enumerated that others - about same when cones are already typed, the code saw enough, I am turned out the approach etc. when after villages Si for js, some days used foul language only the first, yet did not understand, as on this shit  it is possible. Watched reverse passage js - With ++ at one guy, there all was much worse. For it With ++ it was awful. More shortly, stereotypes hinder. I will assume that at the majority of the people knowing of some languages, problems does not arise, "the bad language" they can give an estimation only in a specific context.