This is my week Week 7 report for the project "Full support of Python
3 in GRASS GIS"
1) What did I complete this week?
Updated ctypes to deal with bytes[0]
Worked on 'temporal' library [1], lot of errors were encountered
Looked into issues for 'pygrass' testsuites which were failing too
Created a spreadsheet[2] detailing errors and information
regarding the temporal testsuites
Also, updated the pygrass testsuite details on spreadsheet[3]
2) What am I going to achieve for the next week?
Start exploring GUI library, starting with mapdisp and imgr
Continue working on fixing ctypes module
Continue working on 'temporal' library
3) Is there any blocking issue?
Ctypes are not fully compatible with the Python3 even if they are
compiling without any issues. Discussed witb mentors about the next
steps and how to deal with ctypes.
Will discuss more about the upcoming milestones and how ctypes
handling is to be done so as not to interfare too much with the
pending work.
On Mon, Jul 2, 2018 at 6:39 AM Helmut Kudrnovsky <hellik@web.de> wrote:
Is there any blocking issue?
Ctypes are not fully compatible with the Python3 even if they are
compiling without any issues. Discussed witb mentors about the >next
steps and how to deal with ctypes.
ctypes seems not to be actively maintained anymore. Are there any
alternatives?
ctypes is standard Python package, ctypesgen is what is not much maintained, although there are forks with some Python3 changes Sanjeet was using. There are alternatives to ctypes such as cffi, Pietro did some initial work. However given the fact that neither Sanjeet nor we mentors are familiar with this, we decided to attempt to fix the ctypesgen library, any other option would not fit the timeline of this GSoC. Sanjeet got pretty far so hopefully we can eventually got it working fully.
ctypes is standard Python package, ctypesgen is what is not much maintained
yes I meant ctypesgen.
although there are forks with some Python3 changes Sanjeet was using.
There are alternatives to ctypes >such as cffi, Pietro did some initial
work. However given the fact that neither Sanjeet nor we mentors are
>familiar with this, we decided to attempt to fix the ctypesgen library,
any other option would not fit the >timeline of this GSoC. Sanjeet got
pretty far so hopefully we can eventually got it working fully.