[Bug 1150105] Review Request: python-pynn - Simulator-independent
specification of neuronal network models
by bugzilla@redhat.com
https://bugzilla.redhat.com/show_bug.cgi?id=1150105
--- Comment #12 from Ankur Sinha (FranciscoD) <sanjay.ankur(a)gmail.com> ---
Thanks for the review Jerry. Apologies for the delay. Here are the updated
Spec/SRPM files:
Spec: https://ankursinha.fedorapeople.org/python-pynn/python-pynn.spec
SRPM:
https://ankursinha.fedorapeople.org/python-pynn/python-pynn-0.9.6-1.fc33....
Changelog:
* Tue Feb 09 2021 Ankur Sinha <ankursinha AT fedoraproject DOT org> - 0.9.6-1
- Remove duplicate neuron-devel BR
- Add patch to correct python operator usage
- Use arch dependent and version specific requirement on base package
- Add comment about NEURON files
- Filter out private libnrnmech from provides
The rpath bit came from the NEURON package. I've fixed that and pushed updates
too. So if you test this one for F32/F33, you'll have to ensure that the
updates-testing repo is enabled for the build.
I've submitted a PR for the python issue, which was already accepted:
https://github.com/NeuralEnsemble/PyNN/pull/706
I've also filed a ticket for the compilation warnings:
https://github.com/NeuralEnsemble/PyNN/issues/707
The x86_64 folder is NEURON specific. Mod files that are to be accessible in
NEURON are compiled together to create a libnrnmech.so library, which is loaded
into NEURON by a script called "special" (that calls `nrniv -dll
$ARCH/.libs/libnrnmech.so`). And then, `.libs/libnrnmech.so` is a sym link to
`../../$ARCH/libnrnmech.so`. I guess they could just use the libnrnmech.so file
instead of the symlink, but this is how it is---probably because of some
historical reasons. It also creates the empty x86_64 directory as you noticed:
really not sure what that is about but I see it everywhere we compile mod
files. I'll double check with colleagues that are running NEURON on other
platforms to see if this one is necessary and also confirm with upstream.
Cheers,
Ankur
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
3 years, 2 months
[Bug 1924924] New: Review Request: R-MatrixGenerics - S4 Generic
Summary Statistic Functions that Operate on Matrix-Like Objects
by bugzilla@redhat.com
https://bugzilla.redhat.com/show_bug.cgi?id=1924924
Bug ID: 1924924
Summary: Review Request: R-MatrixGenerics - S4 Generic Summary
Statistic Functions that Operate on Matrix-Like
Objects
Product: Fedora
Version: rawhide
Hardware: All
OS: Linux
Status: NEW
Component: Package Review
Severity: medium
Priority: medium
Assignee: nobody(a)fedoraproject.org
Reporter: spotrh(a)gmail.com
QA Contact: extras-qa(a)fedoraproject.org
CC: package-review(a)lists.fedoraproject.org
Target Milestone: ---
Classification: Fedora
Spec URL: https://spot.fedorapeople.org/R-MatrixGenerics.spec
SRPM URL: https://spot.fedorapeople.org/R-MatrixGenerics-1.2.1-1.fc33.src.rpm
Description:
S4 generic functions modeled after the 'matrixStats' API for alternative matrix
implementations. Packages with alternative matrix implementation can depend on
this package and implement the generic functions that are defined here for a
useful set of row and column summary statistics. Other package developers can
import this package and handle a different matrix implementations without
worrying about incompatibilities.
Fedora Account System Username: spot
NOTE: This package depends on F34+ versions of R components, accordingly, it
will only have builds in F34+.
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
3 years, 2 months