Closed Bug 126883 Opened 22 years ago Closed 22 years ago

bugzilla.dtd isn't quite correct

Categories

(Bugzilla :: Bugzilla-General, defect, P2)

2.15
x86
Linux

Tracking

()

RESOLVED FIXED
Bugzilla 2.16

People

(Reporter: bbaetz, Assigned: bbaetz)

References

()

Details

Attachments

(1 file)

As pointed out on npm.webtools, the dtd has a couple of optional elements
(exporter and urlbase) which shouldn't be there.

The stuff exported from landfill fails validation however, because it is missing
op_sys. This is because the enum value is invalid. Not sure how that happened,
and it occurs on bmo as well. However, the last bug # affected by this on bmo is
bug 56209, so whatever the corruption this was has presumably been fixed.

On the newsgroup, I said that we should still output empty values, because of
this. I think I've changed my mind - bugzilla is not guarenteed to give correct
information if sanity checks fail (Of course, they should never be allowed to
fail, but theres not much option with mysql). I've therefore added a sanity
check for this so that we do fail the checks.

Anyway, an op_sys of '' isn't likely to be valid on the other end.
Attached patch patchSplinter Review
want this for 2.16
Severity: normal → major
Status: NEW → ASSIGNED
Keywords: patch, review
Priority: -- → P2
Target Milestone: --- → Bugzilla 2.16
however, even with GIGO...  shouldn't we validate on the way out, and if it's
garbage, report the error instead of generating a faulty file?

If the bug is corrupted, generate an error.  That way the admin has to fix the
corruption before anyone can export the data (which in my opinion would be a
good thing).
But we don't do that anywhere else. I think an assumption that the database
passes all sanity checks is a reasonable one. Should I check that bugs in the
depends/blocks field are valid, using an extra sql query, for example? What
about checking that all the voters are correct, and that the vote cache is up to
date? (Actually, bug.pm doesn't use the vote cache, so thats not an issue...)
Comment on attachment 70656 [details] [diff] [review]
patch

r=gerv

Gerv
Attachment #70656 - Flags: review+
Comment on attachment 70656 [details] [diff] [review]
patch

r= justdave
Attachment #70656 - Flags: review+
Checked in.
Status: ASSIGNED → RESOLVED
Closed: 22 years ago
Resolution: --- → FIXED
QA Contact: matty_is_a_geek → default-qa
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: