-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/5186/#review8217
-----------------------------------------------------------



src/linux/fs.hpp
<https://reviews.apache.org/r/5186/#comment17768>

    Separate these two with a newline please.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17769>

    Newline unless common prefix directories.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17771>

    s/Lock */Lock* /



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17772>

    s/Lock */Lock* /



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17773>

    s/char */char*/ (here and below please).



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17793>

    How about just 'return ::hasmntopt(...) != NULL;'



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17774>

    Just call this 'table' instead of rv. In general, we prefer names which are 
more descriptive than 'rv', 'ret', 'r', etc.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17777>

    s/FILE */FILE* /



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17775>

    Is an errno set which can be used to determine whether or not you just need 
to try again with a larger buffer?



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17776>

    This line should not compile.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17778>

    Don't think this comment is necessary (or the one in the previous 
function). ;)



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17785>

    Again, 'table' is preferred (and it reads better other places, i.e., 
'table.entries.push_back(...)').



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17779>

    s/fstab */fstab* /



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17782>

    We prefer to do explicit null checks, i.e., 'if (fstab == NULL) {'. Also, 
if the semantics here are that NULL means end of entries, please say so in a 
comment.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17789>

    Formatting seems wrong, +4 for this please.



src/linux/fs.cpp
<https://reviews.apache.org/r/5186/#comment17790>

    +4 spaces please.



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17794>

    s/rv/table/



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17795>

    Great idea! But you should use Option instead of Result here.



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17799>

    This should be an ASSERT_TRUE since you're doing a '.get()' on the next 
line.



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17796>

    s/rv/table/



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17797>

    Again, Option instead of Result.



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17798>

    This should be an ASSERT_TRUE.



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17800>

    s/rv/table/



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17801>

    s/Result/Option/



src/tests/fs_tests.cpp
<https://reviews.apache.org/r/5186/#comment17802>

    ASSERT_TRUE.


- Benjamin


On 2012-05-30 15:25:22, Jie Yu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/5186/
> -----------------------------------------------------------
> 
> (Updated 2012-05-30 15:25:22)
> 
> 
> Review request for mesos and Benjamin Hindman.
> 
> 
> Summary
> -------
> 
> Add two files (src/linux/fs.h|cpp) to handle file system operations on Linux.
> 
> 
> Diffs
> -----
> 
>   src/Makefile.am 96cb4c6 
>   src/linux/fs.hpp PRE-CREATION 
>   src/linux/fs.cpp PRE-CREATION 
>   src/tests/fs_tests.cpp PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/5186/diff
> 
> 
> Testing
> -------
> 
> On linux machines, make check.
> 
> Mount/unmount is not tested since root permission is required.
> 
> 
> Thanks,
> 
> Jie
> 
>

Reply via email to