[conda] conda install from a *.bz2 file appears not to check dependencies

[PeterC]

I am new to conda packaging and still trying to fully understand how to use
it.

I have a conda built package with a dependency on logstash_formatter. I am
working on creating a conda package version of logstash_formatter so that
the dependency can be fulfilled by conda (as an aside, this is currently
failing and I've filed an issue with conda-build (#851)).

However, along the way I noticed that when I install directly from the
*.bz2 file using

conda install -n porttoy_0_0_2 .\porttoy-0.0.2-py35_0.tar.bz2

The porttoy package is installed without any warning or issues. However,
when I import the package and attempt to run it, it will fail to run
because the logstash_formatter dependency was not installed (until I get
the conda package built for it you have to pip install it).

When I install the same package via a channel search into my conda-build
local packages (--use-local) as in:

conda install porttoy --use-local --override-channels

I get an appropriate warning that the dependency on logstash is failing:

Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata: ..
Solving package specifications: .
Error: Dependency missing in current win-64 channels:
- porttoy -> logstash_formatter

Is this an issue conda install or my meta.yaml?

Here is the meta.yaml

package:
name: porttoy
version: "0.0.2"

source:
path: .

build:
script:
- pip install logstash_formatter
- python setup.py install


requirements:
build:
- python
- setuptools
- cython
- numpy
- logstash_formatter

requirements:
run:
- python
- numpy
- logstash_formatter

[Ilan Schnell]

Hello,

installing conda packages directly, for example using "conda install -n
porttoy_0_0_2 .\porttoy-0.0.2-py35_0.tar.bz2", does not check any
dependencies. This way of installing conda packages is only meant for
developing and debugging purposes.
The normal way you install a conda packages is by adding it a (either local
or remote) conda package repository. The "--use-local" options adds the
local package repository to the list of conda channel for packages to look

conda install --use-local porttoy

- Ilan

I am new to conda packaging and still trying to fully understand how to
use it.
I have a conda built package with a dependency on logstash_formatter. I am
working on creating a conda package version of logstash_formatter so that
the dependency can be fulfilled by conda (as an aside, this is currently
failing and I've filed an issue with conda-build (#851)).
However, along the way I noticed that when I install directly from the
*.bz2 file using
conda install -n porttoy_0_0_2 .\porttoy-0.0.2-py35_0.tar.bz2
The porttoy package is installed without any warning or issues. However,
when I import the package and attempt to run it, it will fail to run
because the logstash_formatter dependency was not installed (until I get
the conda package built for it you have to pip install it).
When I install the same package via a channel search into my conda-build
conda install porttoy --use-local --override-channels
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata: ..
Solving package specifications: .
- porttoy -> logstash_formatter
Is this an issue conda install or my meta.yaml?
Here is the meta.yaml
name: porttoy
version: "0.0.2"
path: .
- pip install logstash_formatter
- python setup.py install
- python
- setuptools
- cython
- numpy
- logstash_formatter
- python
- numpy
- logstash_formatter
--
You received this message because you are subscribed to the Google Groups
"conda - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an
Visit this group at https://groups.google.com/a/continuum.io/group/conda/.

[PeterC]

Ilan

Got it, thank you. Peter

[Amit Sudharshan]

While I understand that there are probably good reasons not to support
loading dependencies from bz2, I think there should be a warning to
indicate that dependencies are not being honored

Amit

Post by PeterC
Ilan
Got it, thank you. Peter
--
You received this message because you are subscribed to the Google Groups
"conda - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an
Visit this group at https://groups.google.com/a/continuum.io/group/conda/.