| 1 |
Greetings, following is the weekly update for g-pypi project. |
| 2 |
It's main purpose is to generate ebuilds from PyPi. |
| 3 |
|
| 4 |
Important: My FTTH link is broken, so fubar.si domain is part-time down. |
| 5 |
I'm very sad about this, should be fixed in few days. |
| 6 |
|
| 7 |
Quick links to project info: |
| 8 |
* http://bitbucket.org/iElectric/g-pypi2- Repository, issue tracker |
| 9 |
* http://hudson.fubar.si/job/g-pypi2/ - Hudson CI instance |
| 10 |
* http://docs.fubar.si/gpypi2/ - Sphinx documentation generated by |
| 11 |
hudson on each commit |
| 12 |
* http://neurogeek.ath.cx/g-pypi - wiki/user stories for scrum |
| 13 |
development |
| 14 |
|
| 15 |
|
| 16 |
Previous week (08-14 June) |
| 17 |
========================= |
| 18 |
|
| 19 |
--------------------------------------------------------------------- |
| 20 |
Task: get working version of ebuild.py |
| 21 |
--------------------------------------------------------------------- |
| 22 |
|
| 23 |
-> This will be main goal of the week. Refactor ebuild.py and make it work |
| 24 |
with enamer.py and portage_utils.py |
| 25 |
|
| 26 |
Done. Only ebuild writing is left to refactor, which can be done at the very |
| 27 |
last stage of prepairing for first prototype. |
| 28 |
|
| 29 |
List of taks assigned: |
| 30 |
|
| 31 |
* get rid of last snippets for generating template with python code - done |
| 32 |
* port some functions (like format_depend) to jinja environment filters - not needed, jinja default filters were used |
| 33 |
* get rid of all non-sense if/else statements (I will probably make |
| 34 |
ebuild inherit from dict) - most of it was removed, other portion with handle configurator |
| 35 |
* add support to detect Sphinx documentation - TODO |
| 36 |
* do not regex parse setup.py files but rather mock the setup object |
| 37 |
and collect metadata (still not sure if that will cover most of cases) - done |
| 38 |
* repair flow of the code to be easier to understand and document it - done (docs still needed) |
| 39 |
|
| 40 |
Most of the work was done to not regex parse setup.py, now we are monkeypatching it. |
| 41 |
All other parts of Ebuild were needed to support this change. |
| 42 |
|
| 43 |
Now we extract from setup.py: |
| 44 |
|
| 45 |
* tests_require |
| 46 |
* setup_requires |
| 47 |
* install_requires |
| 48 |
* extras_require |
| 49 |
|
| 50 |
Additional, extra dependencies are converted to USE flags. |
| 51 |
|
| 52 |
Category guessing function was also added to Enamer, I will probably make first implementation based |
| 53 |
on CLASSIFIERS info from PyPi. |
| 54 |
|
| 55 |
|
| 56 |
Upcoming week (15-21 June) |
| 57 |
========================== |
| 58 |
|
| 59 |
Main goal of upcoming week will be to get closer to first prototype. I would love to get some initial feedback |
| 60 |
from Gentoo users. So the goal is: get gpypi2 echo ebuilds. |
| 61 |
|
| 62 |
---------------------------------------------- |
| 63 |
Task: First snippet of command line interface |
| 64 |
---------------------------------------------- |
| 65 |
|
| 66 |
Put enough together to assemble ebuilds from command line. I think |
| 67 |
many issues may pop up, and lots of bugs shall be fixed. I'm currently |
| 68 |
mostly doing stuff from my head, with not much practise whatsoever. |
| 69 |
|
| 70 |
Mostly, we should get a result of better test coverage (currrently about |
| 71 |
50%). |
| 72 |
|
| 73 |
The only component left untouched is the configuration. I would like to |
| 74 |
touch this topic after all my exams are over, leaving me time and |
| 75 |
productivity. |
| 76 |
|
| 77 |
|
| 78 |
---------------------------------------------------- |
| 79 |
Task: get_dependencies - another refactor to enamer |
| 80 |
---------------------------------------------------- |
| 81 |
|
| 82 |
Title says it all. Also handling of multiple version operators should be |
| 83 |
added. |
| 84 |
|
| 85 |
Example: foobar>=0.5 && foobar<0.6 |
| 86 |
|
| 87 |
should resolve in <0.6 blocker and >=0.5 requirement. |
Archives