esa-snap-s1tbx-gptMAINTAIN
Maintainer guide

Build, test,
publish, document.

Packaging details that should not crowd the README, collected for future release work.

Key files

  • versions.yaml - supported SNAP versions and retained platform cells
  • recipe/meta.yaml - conda-build metadata
  • recipe/build.sh - Linux/macOS installer execution and pruning
  • recipe/bld.bat - Windows installer execution and pruning
  • recipe/activate.* - expose SNAP_HOME, gpt, snap, and Java settings
  • scripts/resolve.py - turns versions.yaml into build variables
  • scripts/compute-hashes.py - verifies installer URLs and fills sha256 values
  • .github/workflows/build.yml - package build and publish workflow

Retained matrix

The public channel intentionally retains:

SNAPlinux-64win-64osx-arm64
9.0.0yesnono
10.0.0yesnono
11.0.0yesnono
12.0.0yesnono
13.0.0yesyesyes

Other cells are represented as null in versions.yaml; the workflow resolves them as intentional skips.

Build numbers

The package version is the SNAP version. Use the conda build number for packaging-only fixes:

13.0.0-1  Java 21 packaging fix
13.0.0-2  Linux/headless startup defaults
11.0.0-1  Linux/headless startup defaults
12.0.0-1  Linux/headless startup defaults

Local build

mamba install -n base -c conda-forge conda-build pyyaml ruamel.yaml
python scripts/compute-hashes.py --only 13.0.0/linux-64
eval "$(python scripts/resolve.py 13.0.0 linux-64)"
conda-build recipe/ --output-folder out

Or use the helper:

scripts/build-local.sh 13.0.0 linux-64

Publishing

Manual workflow inputs:

versions = 11.0.0 12.0.0 13.0.0
subdirs  = linux-64
publish  = true

Required repository secrets:

ANACONDA_TOKEN
ANACONDA_OWNER

The workflow resolves versions.yaml, skips null cells, builds on native runners, runs smoke tests, runs pyroSAR discovery on Unix, and uploads to Anaconda.org when publish=true.

GitHub Pages

GitHub Pages is configured to publish the docs/ directory from main.

Settings -> Pages -> Source -> Deploy from a branch
Branch: main
Folder: /docs

Expected URL:

https://pmuguda.github.io/snap-gpt-conda/