Ivan Smirnov comments

Results 324 comments of


                                            Ivan Smirnov

Support combination of features `mpio` and `static`

@magnusuMET It looks like `mpi-sys` needs mpi compiler anyway to do its bindgen-style magic to build itself, so it's not like we're requiring anything extra I think (https://github.com/rsmpi/rsmpi).

.required() on Parsers doesn't make them fail the parse

Also just bumped into this, and #28 seems to have been hanging there for a while now.

Exception on mandatory options?

Found myself checking them all the time too manually, with the new version -- would sure be a nice thing to have built-in.

Do not build namedtuple on each function call

@deadpixi Wonder if this could be merged?..

llvm-related build errors on OSX

An update: killed all previous llvm installations, reinstalled from brew with `--HEAD`, now libc++/libstdc++ seems to be resolve but there's this: ``` sh $ make -f Makefile.osx MACOSX_DEPLOYMENT_TARGET=10.7 clang++ -std=c++11...

llvm-related build errors on OSX

To this point, I have no idea how to build it on OS X. Has anyone succeeded with the homebrew's llvm? Also, `MACOSX_DEPLOYMENT_TARGET=10.7` hardcoded in the OSX makefile doesn't seem...

[WIP] FIX CDClassifier with n_jobs != 1 in py36 (#114)

@fabianp @mblondel Could this get possibly merged?

Make LDAP bind failure cause an ldap_error, rather than an invalid_credentials

changing callback_path to callback_url to account for relative root url

Has anyone figured how to resolve this? Including the sub URI in the `:callback_path` seems wrong, because then `on_callback_path?` would fail (as it matches against it). So the correct solution...

changing callback_path to callback_url to account for relative root url

TL;DR: this PR does look like a correct solution, are there are any reasons not to merge it in?