summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorBonfaceKilz2021-12-03 15:42:19 +0300
committerBonfaceKilz2021-12-03 15:42:19 +0300
commitcbe94eb40a84a6a071f671d685d83284030d4cd8 (patch)
tree7186fba35d7e97f83053ee981e4b559d4b6c6f32
parent96b9a25c0911b6e8411ee12cc3eac048729e217b (diff)
downloadgn-gemtext-cbe94eb40a84a6a071f671d685d83284030d4cd8.tar.gz
deploy.gmi: New issue to track configuration settings
-rw-r--r--issues/deploy.gmi26
1 files changed, 26 insertions, 0 deletions
diff --git a/issues/deploy.gmi b/issues/deploy.gmi
new file mode 100644
index 0000000..f701805
--- /dev/null
+++ b/issues/deploy.gmi
@@ -0,0 +1,26 @@
+# Handle Configuration Files
+
+How we handle our configuration is a mess; and this makes deployment a
+hassle. In genenetwork2, our configs live in different places, and it
+takes some fiddling tinged with a bit of domain (gn2) knowledge to
+work out what setting live where. With the introduction of blueprints
+in our code-base, this duplication became a necessary evil because
+most of the settings QC checks happen in a "utility/settings.py" which
+unfortunately imports the "app" object. This import leads to circular
+imports whenever we use blueprints. Unfortunately, a huge chunk of
+the code base uses "utility/settings.py" to get configuration
+settings; and removing this would make gn2 as is disfunctional.
+
+# Tags
+
+* assigned: bonfacem
+
+# Notes
+
+Here's an example of how to use json as our settings:
+
+=> https://stackoverflow.com/questions/15122312/how-to-import-from-config-file-in-flask
+
+Here's Flask's official documentation on handling configuration:
+
+=> https://flask.palletsprojects.com/en/1.1.x/config/