Today spatstat 1.48-0 ‘Model Prisoner’ has been uploaded to CRAN where Windows and OSX binaries are also available. You can also download it from the github repository. The main changes are (full release notes here):

• Sufficient Dimension Reduction for point processes.
• Alternating Gibbs Sampler for point process simulation.
• Intensity approximation for area-interaction and Geyer models.
• New class of spatially sampled functions.
• ROC and AUC extended to other types of point patterns and models.
• More support for linear networks.
• More support for infinite straight lines.
• Simulation of ‘rhohat’ objects.
• Kernel smoothing accelerated.
• Methods for ‘head’ and ‘tail’ for spatial patterns.
• More low-level functionality.
• Improvements and bug fixes.
• spatstat now has more than 1000 help files.