-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26859/#review58921
-----------------------------------------------------------



src/master/master.hpp
<https://reviews.apache.org/r/26859/#comment100101>

    There shouldn't be a newline added here



src/master/master.cpp
<https://reviews.apache.org/r/26859/#comment100105>

    Since all this code only can load one authenticator module it would be good 
to check the authetnicatorModules argument only produces a list of length 1 max 
and warn / exit early. Currently extra arguments are silently ignored.



src/master/master.cpp
<https://reviews.apache.org/r/26859/#comment100103>

    The strings::split() API makes this a lot more complicated than it should 
be. split of an empty string returning an empty vector I think would make more 
sense... But alas...
    
    I think it would make sense though to default the argument to a list of one 
element ['crammd5'] by default. Then you should just error if not 
--authenticator= is passed without listing any authenticators, rather than try 
to fall back. Ideally the error would list available authentication modules.



src/master/master.cpp
<https://reviews.apache.org/r/26859/#comment100104>

    This assuems crammd5 can only appear first in the list of authenticator 
modules.



src/master/master.cpp
<https://reviews.apache.org/r/26859/#comment100113>

    If authenticator should really be Owned here it would be a lot better to 
make it owned much earlier so that it is never unowned / a resource which could 
potentially be lost if / when the unexpected happens.



src/master/master.cpp
<https://reviews.apache.org/r/26859/#comment100114>

    All the right angle bracket fixes are nice, but the ones in chunks of code 
which aren't otherwise touched make the diff harder to read.



src/module/authenticator.hpp
<https://reviews.apache.org/r/26859/#comment100115>

    This needs to be added to a Makefile.am so that distcheck will pass.


- Cody Maloney


On Oct. 28, 2014, 12:44 p.m., Till Toenshoff wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26859/
> -----------------------------------------------------------
> 
> (Updated Oct. 28, 2014, 12:44 p.m.)
> 
> 
> Review request for mesos, Adam B, Benjamin Hindman, and Vinod Kone.
> 
> 
> Bugs: MESOS-1889
>     https://issues.apache.org/jira/browse/MESOS-1889
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> Enables selecting a module based authenticator via the new --authenticators 
> flag for mesos master.
> 
> Additionally, all "> >" have been fixed towards ">>" in master.hpp and 
> master.cpp.
> 
> 
> Diffs
> -----
> 
>   src/master/flags.hpp c931fd9 
>   src/master/master.hpp b1a2cd0 
>   src/master/master.cpp 762d2ff 
>   src/module/authenticator.hpp PRE-CREATION 
>   src/tests/crammd5_authentication_tests.cpp PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/26859/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> NOTE all three CRAM-MD5 authenticator module related RRs need to get applied 
> before running make check.
> 
> 
> Thanks,
> 
> Till Toenshoff
> 
>

Reply via email to