----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/66002/#review199204 -----------------------------------------------------------
src/files/files.cpp Lines 314-328 (original), 314-329 (patched) <https://reviews.apache.org/r/66002/#comment279505> const src/files/files.cpp Line 350 (original), 352 (patched) <https://reviews.apache.org/r/66002/#comment279507> const src/files/files.cpp Line 384 (original), 387 (patched) <https://reviews.apache.org/r/66002/#comment279508> Ha, this ought to be const too even though you didn't add the variable. I sure wish C++ default to const like Rust :( src/files/files.cpp Line 453 (original), 458 (patched) <https://reviews.apache.org/r/66002/#comment279509> const src/files/files.cpp Line 628 (original), 634 (patched) <https://reviews.apache.org/r/66002/#comment279510> const src/files/files.cpp Line 769 (original), 776 (patched) <https://reviews.apache.org/r/66002/#comment279511> const src/tests/files_tests.cpp Lines 256-257 (original), 256-257 (patched) <https://reviews.apache.org/r/66002/#comment279513> With the above fixes, `files.attach` now accepts URI like paths, right? Or did we skip `attach`? Should we? I don't know. src/tests/files_tests.cpp Lines 336-345 (original), 336-349 (patched) <https://reviews.apache.org/r/66002/#comment279515> Ew. Can we test these without `stat`? I don't want to add to [MESOS-8275](https://issues.apache.org/jira/browse/MESOS-8275) - Andrew Schwartzmeyer On March 9, 2018, 12:17 p.m., John Kordich wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/66002/ > ----------------------------------------------------------- > > (Updated March 9, 2018, 12:17 p.m.) > > > Review request for mesos, Akash Gupta, Andrew Schwartzmeyer, Jeff Coffler, > and Joseph Wu. > > > Repository: mesos > > > Description > ------- > > Many mesos frameworks assume that path separators are forward slashes, > like dcos-ui and marathon-ui. dcos-cli also assumes this. Previous to > this patch, if a forward slash was given in the path variable to an HTTP > API function call to the mesos agent on a Windows system, the Windows > API would fail at recognizing the path, because the Windows API accepts > only backslashes as path separators. To remedy this issue, we now > convert all forward slashes passed as a path to the HTTP API to an agent > to back slashes on Windows agents by using the path::from_uri function. > > > Diffs > ----- > > src/files/files.cpp 5f92d2af722c3201200fa1d6a75dc0b87fdc6078 > src/tests/files_tests.cpp c703cae03345112715aeab83cb0a74abe3e12469 > > > Diff: https://reviews.apache.org/r/66002/diff/2/ > > > Testing > ------- > > Tested on both Windows and Linux, all tests pass, including the two newly > enabled tests on Windows. > > > Thanks, > > John Kordich > >
