Since our last report in March, 2018 we have continued work on several parallel tracks:
- We have released our final GMT 4 version in July 2018, the GMT 4.5.18 release.
- We have maintained the GMT 5.4 branch which is the current stable release branch, releasing 5.4.4 (the latest official release) on July 1, 2018.
- We have ported all of our subversion repository and issue tracking to GitHub.
- We have secured some funding from EarthScope for the new modern mode.
- We are working on stabilizing GMT 6.0.0 beta and modern mode implementation.
- We have secured funding for the upcoming Portugal and California developer summits in 2019.
We have issued our final GMT 4 release this July 1. This is version 4.5.18 and this terminates our involvement with bug fixes and point releases for this series. GMT 5 was released 5 years ago and any issues discovered in 4.5.18 need to be reproduced in GMT 5 where we will make corrections.
Because all our development is now taking place in GMT 6, the 5.4 branch is only updated when a bug is found. We anticipate to make a 5.4.5 release shortly, which has been held up by the GitHub transition. The 5.4.5 release will simplify installation and building of the MB-system with GMT5 and 6.
Wessel, Luis and Uieda attended the annual GMTSAR workshop at Scripps in August. Due to a late conflict, Wessel missed most of the workshop but arrived on the last day. GMT developers then stayed the weekend at Scripps and collaborated on GMT transition activities (below).
During the GMTSAR 2018 workshop at Scripps, three of the GMT Developers (Wessel, Luis, Uieda) spent a weekend where we ported the entire GMT 5/6 subversion repository to GitHub. Using git will be the fourth version control system that GMT has used over the decades (SCCS, CVS, subversion, now git). The new repository (https://github.com/GenericMappingTools/gmt) is online at GitHub and is independent of the University of Hawaii. We expect that this transition will make it easier for community members to make contributions to the GMT code and documentation. For now, only the source code and the issue tracking has been migrated, but we are working on rebuilding the entire documentation process as well. We have started planning the first of two GMT developer summits to prepare our plans for how a GMT transition can be accomplished. All six GMT developers are confirmed to attend the summit which will be held in Faro, Portugal from January 20-27, 2019. Being off-season and with 3 developers in Europe, this was the cheapest and soonest meeting we could have.
With the move to GitHub we have laid the foundation for future development. We have added Continuous Integration (CI) which means each time a pull request is made the whole suite is compiled and tests are run to make sure the changes in the pull request do not break anything in the repository. GitHub also provides an Issues list similar to our old Issues on the Wiki at SOEST. We still need to port older bug reports to GitHub as well as establish a forum similar to the one at SOEST. Finally, we need to automate the building of the documentation so that we can move off the GMT server at SOEST. By not relying on SOEST computers and hosting we are making it easier for a community of developers and users to assist with the maintenance of GMT. GMT 6 is slowly getting more stable, and with the separate funding from EarthScope we are implementing and testing the modern mode plus the new modules for movie making (movie.c) and Google Earth quadtree imaging (grd2kml.c). We hope GMT 6 beta will be ready early in 2019.
We have secured NSF funding for our two proposed summits. At this point we have made reservations for the GMT Developer Summit in Faro, Portugal for Jan 20-27, taking advantage of very cheap rates for accommodation and travel. During this summit we will work up plans for the transition and these will be circulated do the GMT Steering Committee as well as NSF. We will incorporate their feedback and prepare a White Paper on the GMT transition for further discussion at the California Summit in the summer of 2019.
Development of the Python wrapper has continued with added support for netCDF grids through the xarray library, a new sphinx extension for inserting GMT/Python plots into documentation, new module wrappers developed by new contributors to the project, functions to download sample datasets, and better reporting of C API errors.
The current implementation is running into a few limitations around building new module wrappers and tutorial documentation. A re-implementation of some of the underlying code is under way, which should make it easier to keep the documentation in sync with examples and the GMT documentation.
Some of the GMT developers (Wessel, Luis, Uieda) met with a group of Scripps graduate students and postdocs during the GMTSAR workshop to gather feedback and talk about GMT development. In particular, they discussed opportunities for getting involved in the project and how this process can be made more accessible. The overall reception was positive and resulted in the some feedback filed as issues in the new Github repository. We hope to keep this communication channel open to recruit new contributors in the near future.
Uieda gave an update on the GMT/Python status at the 2018 Scipy Conference in July (a recording of the talk is available on Youtube: https://youtu.be/6wMtfZXfTRM).