Bug #2994
Should not attempt to manage content without checksum
| Status: | Closed | Start: | 12/29/2009 | |
|---|---|---|---|---|
| Priority: | Normal | Due date: | ||
| Assigned to: | % Done: | 0% |
||
| Category: | plumbing | |||
| Target version: | 0.25.2 | |||
| Affected version: | 0.25.2rc3 | Branch: | http://github.com/MarkusQ/puppet/tree/ticket/0.25.x/2994 | |
| Keywords: | ||||
| Votes: | 0 |
Description
Attempting to call a constructed method name with a nil component (the checksum type in this case) causes an unhelpful error message in a case where either 1) the results were not needed anyway or 2) a more useful error message would otherwise be generated. It also opens (at least hypothetically) then possibility that some unrelated method (in this case “_file” would be called if it existed.
History
Updated by Markus Roberts 7 months ago
- Status changed from Accepted to Ready for Testing
- Branch set to http://github.com/MarkusQ/puppet/tree/ticket/0.25.x/2994
Updated by Markus Roberts 7 months ago
- Affected version changed from 0.25.1 to 0.25.2rc2
Updated by James Turnbull 7 months ago
- Status changed from Ready for Testing to Closed
Pushed in commit:“d1ff4b3fca31ee99276c26bc1df6391cfc79254a” in branch 0.25.x
Updated by Markus Roberts 7 months ago
- Status changed from Closed to Code Insufficient
The first change simply exposed an underlying problem; metadata for links does not contain the checksum, which is needed when links is set to follow.
Updated by Markus Roberts 7 months ago
- Status changed from Code Insufficient to Ready for Testing
Branch updated with additional fix.
Updated by Markus Roberts 7 months ago
- Status changed from Ready for Testing to Code Insufficient
The fix works as reported by the user, but I’m not satisfied with the error handling. Revised (and hopefully final) branch up shortly.
Updated by Markus Roberts 7 months ago
- Status changed from Code Insufficient to Ready for Checkin
The branch has been updated with a version that has better tests and error handling (and still works at the user’s site). It’s good to go.
Updated by James Turnbull 7 months ago
- Status changed from Ready for Checkin to Closed
Pushed in commit:“7e64393dd10023d528d2fc21383ead30c9ee94dd” and commit:“d1ff4b3fca31ee99276c26bc1df6391cfc79254a” in branch 0.25.x
Updated by Mark Plaksin 7 months ago
- Status changed from Closed to Re-opened
- Affected version changed from 0.25.2rc2 to 0.25.2rc3
I swear the fix tested just fine for me but the “unknown checksum” and filebucket-every-time problem is back for me on all of our OSes with 0.25.2rc3.
Updated by Markus Roberts 7 months ago
- Status changed from Re-opened to Investigating
Updated by Markus Roberts 7 months ago
Mark —
I tested it on your box too, and just retested with the same results you saw; it worked before, but is not working in 0.25.x. I’m digging in to it now.
— Markus
Updated by Markus Roberts 7 months ago
Branch updated with a version of the patch which always checksums links (going back to the earlier strategy) with tests adjusted to correspond.
This was needed to eliminate the partial reintroduction of the original bug in the case where the checksum of a followed link is being requested via rest from the client (which doesn’t know that it’s not just a regular file, and thus can’t know to pass {links => follow} in the request).
Updated by Markus Roberts 7 months ago
- Status changed from Investigating to Ready for Checkin
Updated by James Turnbull 7 months ago
- Status changed from Ready for Checkin to Closed
Third time is the charm – pushed in commit:“cbc2ef003d93e5162979cebc927b5d9a69397439” in branch 0.25.x