> On Oct. 29, 2014, 12:18 a.m., Cody Maloney wrote:
> > src/master/master.cpp, line 3912
> > <https://reviews.apache.org/r/26859/diff/6/?file=735286#file735286line3912>
> >
> >     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.

Yes, I feared this would come up. I think that we won't touch all areas of this 
implementation for a loooong time, hence I simply took the freedom to get it 
done now.


> On Oct. 29, 2014, 12:18 a.m., Cody Maloney wrote:
> > src/module/authenticator.hpp, line 1
> > <https://reviews.apache.org/r/26859/diff/6/?file=735287#file735287line1>
> >
> >     This needs to be added to a Makefile.am so that distcheck will pass.

Ow, great point - thanks a bunch! As this file should have been added in 26857, 
I will fix it there.


> On Oct. 29, 2014, 12:18 a.m., Cody Maloney wrote:
> > src/master/master.cpp, line 3873
> > <https://reviews.apache.org/r/26859/diff/6/?file=735286#file735286line3873>
> >
> >     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.

Yes indeed.


> On Oct. 29, 2014, 12:18 a.m., Cody Maloney wrote:
> > src/master/master.cpp, line 3876
> > <https://reviews.apache.org/r/26859/diff/6/?file=735286#file735286line3876>
> >
> >     This assuems crammd5 can only appear first in the list of authenticator 
> > modules.

MESOS-1939 will fix this, as noted.


> On Oct. 29, 2014, 12:18 a.m., Cody Maloney wrote:
> > src/master/master.cpp, line 3875
> > <https://reviews.apache.org/r/26859/diff/6/?file=735286#file735286line3875>
> >
> >     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.

Using a default makes much more sense, even though it complicates the flags 
description a little. Changed that accordingly - thank you!.


- Till


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


On Oct. 29, 2014, 5:19 p.m., Till Toenshoff wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26859/
> -----------------------------------------------------------
> 
> (Updated Oct. 29, 2014, 5:19 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/examples/test_authenticator_module.cpp PRE-CREATION 
>   src/master/flags.hpp c931fd9 
>   src/master/master.hpp b1a2cd0 
>   src/master/master.cpp 762d2ff 
>   src/tests/cram_md5_authentication_tests.cpp PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/26859/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> NOTE all four CRAM-MD5 authenticator module related RRs need to get applied 
> before running make check.
> 
> 
> Thanks,
> 
> Till Toenshoff
> 
>

Reply via email to