by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Nov 11 2018 09:31

    vinipsmaker on socket

    Restore behaviour changed in la… (compare)

  • Nov 10 2018 22:39

    vinipsmaker on socket

    Refactor socket impl to rely on… (compare)

  • Oct 28 2018 05:23

    vinipsmaker on socket

    Update `chunk_ext` token to inc… Merge branch 'master' into sock… Move test utils to separate fol… and 3 more (compare)

  • Oct 28 2018 05:23

    vinipsmaker on master

    Update `chunk_ext` token to inc… (compare)

  • Oct 06 2018 00:00

    vinipsmaker on socket

    Update Tufão examples Add `chunk_ext` token Now user… Fully qualify references to ASI… and 9 more (compare)

  • Oct 06 2018 00:00

    vinipsmaker on master

    Fully qualify references to ASI… Add noisy code to clear GCC ann… (compare)

  • Oct 03 2018 13:18

    vinipsmaker on socket

    Replace test's helper use_yield… (compare)

  • Oct 01 2018 12:49

    vinipsmaker on socket

    Disable `write_state::finished`… Refactor async_write_response t… Refactor async_write to use asi… and 4 more (compare)

  • Sep 02 2018 03:36
  • Aug 05 2018 10:03

    vinipsmaker on master

    Add `chunk_ext` token Now user… (compare)

  • Jul 29 2018 07:27

    vinipsmaker on master

    Update Tufão examples (compare)

  • Apr 23 2018 11:33

    vinipsmaker on socket

    Remove unused code Simplify some lambda declaratio… Refactor async_write_response_m… (compare)

  • Mar 26 2018 11:30

    vinipsmaker on traviszzz

    (compare)

  • Mar 26 2018 11:29

    vinipsmaker on socket

    Update CI (Travis and AppVeyor)… (compare)

  • Mar 26 2018 11:00

    vinipsmaker on traviszzz

    Update CI (Travis and AppVeyor) (compare)

  • Mar 26 2018 10:59

    vinipsmaker on socket

    Update CI badges Merge branch 'parser' into sock… (compare)

  • Mar 26 2018 10:59

    vinipsmaker on master

    Update CI badges (compare)

  • Mar 26 2018 10:41

    vinipsmaker on traviszzz

    Update CI (Travis and AppVeyor) (compare)

  • Mar 26 2018 10:21

    vinipsmaker on traviszzz

    Update CI (Travis and AppVeyor) (compare)

  • Mar 26 2018 10:17

    vinipsmaker on traviszzz

    Update CI (Travis and AppVeyor) (compare)

Vinícius dos Santos Oliveira
@vinipsmaker
no code, but a lot of questions
lamtf
@lamtf
may i send to you my current (incompleted) thesis? (i will update it because i'm gonna make a single algorithm [using a patricia tree i guess] and i'll just compare with crow.. i guess.. using a brenchmark tool)
Vinícius dos Santos Oliveira
@vinipsmaker
feel free to send me your thesis: vini.ipsmaker@gmail.com
Vinícius dos Santos Oliveira
@vinipsmaker
but crow performs better in this benchmark than nodejs alternatives
requests/s: the more the better
lamtf
@lamtf
ok.. i have some warnings about my thesis.. it is very very incomplete (even if the text seems long).. and i will have to change the introduction.. because i have to update the concept of middleware.. and i won't make all that algorithms which i described.. i will make a single one.
Vinícius dos Santos Oliveira
@vinipsmaker
no worries
lamtf
@lamtf
i have to include a lot about chow
lamtf
@lamtf
hello guys!
Vinícius dos Santos Oliveira
@vinipsmaker
I'm currently writing/designing/battle-testing request router ideas for Boost.Http
I always like to see Node.JS frameworks do because they all care about async and there are plenty
One thing that happens in Node.JS routers is the next argument: https://github.com/senchalabs/connect#use-middleware
and the handler assumes the form of void handler(request_t request, response_t response, function<...> next)
Vinícius dos Santos Oliveira
@vinipsmaker
in Boost.Http, there is no need for separate request_t and response_t, both are abstracted away into socket, but it is still required to pass the read message_type imessage
and, of course, it is still useful to pass the message_type omessage along, so chaining handlers can allow them partially handle the response
therefore, if we try to mirror the Node.JS design, the Boost.Http handler signature could be void handler(socket_t &socket, message_t &imessage, message_t &omessage, function<...> next)
However, I don't like the function<...> next argument and I thought a lot about removing it. My current idea is to embed an opaque type into the message_t &imessage depending on the router type (it could be an iterator) as it'd be a lot cheaper
Vinícius dos Santos Oliveira
@vinipsmaker
I also thought about routers done using TMP were each append_route(rule, handler) call would return a different router object. Embeding the next argument into the imessage object would still work because I can embed a type erased type that doesn't incur heap allocation (which would occur if we used function<...>) and we would only pay for the type erasure when we "call the next func".
there still a lot of questions to solve regarding the difference between tree-based routers and middleware routers and things that I'm not 100% certain about
I should really start writing the prototype routers to clear some questions out by this time
I know this channel isn't very "heated", but I though I could share a "progress report" in case anyone here read stuff
lamtf
@lamtf
@vinipsmaker, my professor told me that everything is "middleware" and there is no separation between the concepts of "middle based routing" and "three based routing", because according to his words middleware is some "software layer" between your program and some resource which you want to access on your program.
so he told me that the "resource" is exactly a "function call"
Vinícius dos Santos Oliveira
@vinipsmaker
this means I need better names to communicate the concepts I see better
thanks =)
lamtf
@lamtf
haha.. you also send me before a figure about it.. idk if you remember.. https://docs.djangoproject.com/en/1.9/topics/http/middleware/
Vinícius dos Santos Oliveira
@vinipsmaker
"And they're always arguing with each other. And what they don't realize is [...] neither one can take his position without the other person because you wouldn't know what you're advocating [..]. And that's the answer to philosophy. You see? I'm a philosopher and I'm not gonna argue very much because if you don't argue with me I don't know what I think."
-- https://www.youtube.com/watch?v=LhkGtd_OGCQ
lamtf
@lamtf
there is a figure on this page.. nothing related to "routes" specifically.. but.. we can imagine on the same way a "routing layer"
lamtf
@lamtf
Hello guys.. i'm trying to make a router implementation for boost.http.. and i came to this.. https://github.com/uisleandro/Router_for_Boost.Http/blob/master/imp4.mar2016/mainApp.cxx .. at first i was trying to understand how the Crow routing works (https://github.com/ipkn/crow) so i take it as reference, specially to make type checking at compile time. It still didn't solve the "routing problem" itself. I'm a little bit concerned about legal issues, specially because i'm taking their "numeric system" idea due to the fact that i could not imagine a better way to do it. So in this version of the code i'm trying to make reference about where i found each part of code which didn't came from my own mind. I hope you guys look on it and give your opinion.
Vinícius dos Santos Oliveira
@vinipsmaker
This weekend I had trouble trying to reenable a complex test: BoostGSoC14/boost.http@cd01cad
Vinícius dos Santos Oliveira
@vinipsmaker
I managed to reenable the test: BoostGSoC14/boost.http@43a18ae
lamtf
@lamtf
Vinicius
Vinícius dos Santos Oliveira
@vinipsmaker
hi
lamtf
@lamtf
Hello Vinicius!
I need to speak
Vinícius dos Santos Oliveira
@vinipsmaker
hi
Vinnie Falco
@vinniefalco
Greetings folks. Congrats on Boost.Http GSoC 2016
Vinícius dos Santos Oliveira
@vinipsmaker
@vinniefalco: thanks =)
Vinícius dos Santos Oliveira
@vinipsmaker
hi guys
I'm gonna close this chat at some point in the next few weeks as the discussion will be conducted in this mailing list: https://groups.google.com/forum/#!forum/boosthttp-dev
I welcome all of you to join on the mailing list
wangxiusong
@kspine
C++ 11 & Boost
Vinícius dos Santos Oliveira
@vinipsmaker
The parser can be used C++98
I'm thiking about using C++11 or higher for a type-safe router
lamtf
@lamtf
Hello everybody, good morning!
JasonWinston
@JasonWinston
Hello everyone ,How to compile CGAL
lamtf
@lamtf
Steve Jobs!! hello
Hodor123456
@hOdOrNight
Hello everyone, I am newcomer in Boost. Where can I start?
Vinícius dos Santos Oliveira
@vinipsmaker
boost.org