> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 190
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line190>
> >
> >     Related to the previous comment, I don't like something called create 
> > that also truncates.
> >     
> >     If the use case is to have a new truncated file, one can simply do:
> >     
> >     utils::os::rm(path);
> >     utils::os::touch(path);
> >     
> >     That is, I'd prefer this was the 'touch' abstraction we know and love, 
> > without the truncate. I would consider the ability to pass a 'bool 
> > truncate' flag to 'touch', but I'm not sure it's necessary.

ok..removed O_TRUNC option and renamed create to touch.


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 196
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line196>
> >
> >     We really shouldn't be logging here (and everywhere else in utils, 
> > since they all return Try or Result).

killed. here and everywhere else


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 217
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line217>
> >
> >     Shouldn't need this line, just use 'strerror(errno)' on the next line.

done


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 253
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line253>
> >
> >     Okay, I reneg my previous comment. I think it makes sense to use Result 
> > for read because there are three possibilities: (1) the data; (2) an error; 
> > (3) EOF. We can represent EOF as Result::none.

fixed for all versions of read(). do we want Try or Result for write(). 
Technically, write shouldnt return a none() but I like the symmetry.


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 265
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line265>
> >
> >     Did you want an 'else if' here? And you should kill this check anyway 
> > and return a Result::none below where you check if length == 0.

killed


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 303
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line303>
> >
> >     Kill.

killed


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 304
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line304>
> >
> >     s/open file./open " + path

done. here and everywhere else.


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 543
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line543>
> >
> >     Kill.

killed


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 555
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line555>
> >
> >     What if it's an error? If your intention is that it will never be an 
> > error, make this a CHECK first before you do a .get().

added the check.


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/common/utils.hpp, line 748
> > <https://reviews.apache.org/r/4899/diff/5/?file=106624#file106624line748>
> >
> >     Right, here, using Result makes sense (like the read above). Just all 
> > the other places didn't make sense.

fixed


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/slave/slave.cpp, line 737
> > <https://reviews.apache.org/r/4899/diff/5/?file=106625#file106625line737>
> >
> >     s/rea/Try<

:)


> On 2012-05-08 23:16:57, Benjamin Hindman wrote:
> > src/tests/utils_tests.cpp, line 89
> > <https://reviews.apache.org/r/4899/diff/5/?file=106627#file106627line89>
> >
> >     UtilsTest should really be a fixture which makes sure the file gets 
> > deleted even if the test fails (see https://reviews.apache.org/r/4981 for 
> > an example).

done


- Vinod


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/4899/#review7702
-----------------------------------------------------------


On 2012-05-04 00:28:13, Vinod Kone wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/4899/
> -----------------------------------------------------------
> 
> (Updated 2012-05-04 00:28:13)
> 
> 
> Review request for mesos and John Sirois.
> 
> 
> Summary
> -------
> 
> We can now read/write strings to a file.
> 
> Some common utils for working with protobuf objects.
> 
> 
> Diffs
> -----
> 
>   src/Makefile.am cd503a8 
>   src/common/protobuf_utils.hpp PRE-CREATION 
>   src/common/utils.hpp 1d81e21 
>   src/slave/slave.cpp 09a8396 
>   src/tests/protobuf_io_tests.cpp 22f37ac 
>   src/tests/utils_tests.cpp 49ec107 
> 
> Diff: https://reviews.apache.org/r/4899/diff
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Vinod
> 
>

Reply via email to