...sd.org/bugzilla/show_bug.cgi?id=223264
Details
Diff Detail
- Repository
- rP FreeBSD ports repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
I'm not yet familiar with python & flavours so this may need reworking of the dependencies but at least Horst's changes are finally up for review...
Makefile | ||
---|---|---|
5 ↗ | (On Diff #36893) | DISTVERSION while you are here? |
27 ↗ | (On Diff #36893) | Update blocks/order to conform to Chapter 15. of the Porter's Handbook? |
29 ↗ | (On Diff #36893) | concurrent? |
files/graphiteapi.in | ||
23 ↗ | (On Diff #36893) | Include a description? |
39 ↗ | (On Diff #36893) | This shouldn't be hard coded. Maybe PYTHON_CMD or PYTHON_VERSION? |
41 ↗ | (On Diff #36893) | %%GRAPHITEAPI_PIDFILE%% ? |
Makefile | ||
---|---|---|
5 ↗ | (On Diff #36893) | PORTREVISION=1 |
Makefile | ||
---|---|---|
5 ↗ | (On Diff #36893) | Do not change PORTVERSION to DISTVERSION if the version itself does not change. |
www/py-graphite-api: clean up rc.d script and Makefile
Updating D13581: www/py-graphite-api: clean up rc.d script
- fix ordering in rc.d script
- ensure rc.d script returns on startup
- ensure working directories are re-created on startup if required
submitted by Horst Kapfenberger via BZ223264
Makefile | ||
---|---|---|
48 ↗ | (On Diff #42879) | In pkg-plist, PREFIX is called %D. But as PREFIX is not used in there, I guess you can just remove this. |
www/py-graphite-api: clean up rc.d script
- teach rc.d to behave when installing with multiple python & gunicorn versions
- clean up substitutions in SUB_LIST, pkg-list and *.in script
Updating D13581: www/py-graphite-api: clean up rc.d script
- fix ordering in rc.d script
- ensure rc.d script returns on startup
- ensure working directories are re-created on startup if required
submitted by Horst Kapfenberger via BZ223264
Makefile | ||
---|---|---|
27 ↗ | (On Diff #36893) | AFAICT this entails just moving NO_ARCH? Nothing else seems to need changing. |
Makefile | ||
---|---|---|
29 ↗ | (On Diff #36893) | It doesn't support concurrent? |
Makefile | ||
---|---|---|
29 ↗ | (On Diff #36893) | weird, this got lost between resurrecting from backup while my main computer is getting repairs... |
Makefile | ||
---|---|---|
29 ↗ | (On Diff #36893) | So I'm happy with this as-is *without* concurrent. Concurrent doesn't do POLA here very well and making further changes to allow people to install side-by-side versions of a daemon doesn't make much sense to me as a sysadmin (convince me otherwise). If concurrent is /on/ then %%ETCDIR%% becomes graphiteapi_config="/usr/local/etc/py27-graphite-api/graphiteapi.yaml" (which we can obviously fix) but would clearly break anybody updating from previous port version to current port version. It seems less evil to skip concurrent, and require people to choose their FLAVOR of python to install, vs hacking ETCDIR out to do what was previously the most intuitive thing. When they update to py3 the port would still share the same config file. Thoughts? |