Document how production installations should be updated
The README currently documents multiple ways to use dvwebloader, including pointing Dataverse directly at the GDCC GitHub Pages deployment:
curl -X PUT \
-d 'https://gdcc.github.io/dvwebloader/src/dvwebloader.html' \
http://localhost:8080/api/admin/settings/:WebloaderUrl
It also recommends using a fork or locally hosted copy for production, but it does not explain how administrators should update those installations when a new dvwebloader version is available.
Could the documentation include an explicit installation and upgrade strategy for each supported deployment method?
Suggested topics to document
GDCC-hosted GitHub Pages deployment
Please clarify that:
- No local update is required.
- Changes published to the GDCC GitHub Pages site can become available to Dataverse users automatically.
- The application code is loaded and executed in the user's browser.
- This method is primarily intended for testing, demonstration, or environments that accept upstream-controlled updates.
- Production administrators should understand that the deployment is not pinned to a release, tag, or commit.
Locally hosted deployment
I'm making an assumption that this is actually the process for updating locally.
Please document a recommended upgrade process, such as:
- Select a release, tag, or reviewed commit.
- Download or clone that revision.
- Copy the
src files to institution-controlled web hosting.
- Run
localinstall.sh if remote JavaScript and CSS dependencies should also be hosted locally.
- Test the new version against a staging Dataverse installation.
- Update the deployed files or change
:WebloaderUrl.
- Retain the previous version for rollback.
It would also help to clarify that localinstall.sh vendors remote JavaScript and CSS dependencies, but does not update dvwebloader itself.
Versioning clarification
The documentation should also identify the authoritative source for the current version.
For example:
- Is the GitHub Pages deployment associated with a specific release?
- Should production installations use the latest GitHub release or the latest commit on
main?
- How can an administrator determine which version is currently deployed?
- Are release artifacts expected to be created for every user-facing version?
A small table could make the behavior clearer:
| Deployment method |
Who controls updates? |
How updates are applied |
| GDCC GitHub Pages |
Upstream maintainers |
Automatically when the hosted site changes |
| Forked GitHub Pages |
Fork owner |
Merge or pull upstream changes |
| Local hosting |
Dataverse administrator |
Manually deploy a selected release or commit |
The current installation options are useful, but documenting the update and rollback behavior would make the production implications much clearer.
Document how production installations should be updated
The README currently documents multiple ways to use dvwebloader, including pointing Dataverse directly at the GDCC GitHub Pages deployment:
curl -X PUT \ -d 'https://gdcc.github.io/dvwebloader/src/dvwebloader.html' \ http://localhost:8080/api/admin/settings/:WebloaderUrlIt also recommends using a fork or locally hosted copy for production, but it does not explain how administrators should update those installations when a new dvwebloader version is available.
Could the documentation include an explicit installation and upgrade strategy for each supported deployment method?
Suggested topics to document
GDCC-hosted GitHub Pages deployment
Please clarify that:
Locally hosted deployment
I'm making an assumption that this is actually the process for updating locally.Please document a recommended upgrade process, such as:
srcfiles to institution-controlled web hosting.localinstall.shif remote JavaScript and CSS dependencies should also be hosted locally.:WebloaderUrl.It would also help to clarify that
localinstall.shvendors remote JavaScript and CSS dependencies, but does not update dvwebloader itself.Versioning clarification
The documentation should also identify the authoritative source for the current version.
For example:
main?A small table could make the behavior clearer:
The current installation options are useful, but documenting the update and rollback behavior would make the production implications much clearer.