Closed Bug 71767 Opened 25 years ago Closed 24 years ago

Provide summary view for secure bugs

Categories

(Bugzilla :: Bugzilla-General, defect)

defect
Not set
normal

Tracking

()

RESOLVED FIXED
Bugzilla 2.14

People

(Reporter: jacob, Assigned: myk)

Details

Attachments

(3 files)

The addition of the bug summary to the page <title> and the tooltips of bug links initially had the unwanted side-effect of leaking this information for secure bugs. The "quick fix" was to not show the summary if the bug is secure, regardless of the user's permissions. Effected files: show_bug.cgi (v1.14), duplicates.cgi (v1.4), globals.pl (v1.85) This should be fixed for 2.14 to allow users with proper permissions to benefit from this information.
Target Milestone: --- → Bugzilla 2.14
The only downside to this patch is that the ValidateBugID function, which this patch uses to determine if the user is authorized to view the bug, does not provide an error message to the user which is as helpful as the one currently being provided. In particular, the current error message provides a link for logging in if the user isn't logged in. The solution to this is probably to make ValidateBugID provide a friendlier error message.
Assignee: tara → myk
Keywords: patch
The other thing this latest patch does is have &ValidateBugID use local variables $usergroupset and $userid instead of potentially setting the values of the global variables $::usergroupset and $::userid. The former approach isn't terrible, but the new way conforms to better coding practices.
Status: NEW → ASSIGNED
Adding "review" keyword to get these on the radars of reviewers (if they aren't already).
Keywords: review
Two comments. (I always seem to have two comments :) By removing all checks from bug_form.pl this has the potential to leak the information from process_bug.cgi. If a user somehow gets a secure bug into thier buglist cookie and then modifies the bug before it in the list (or any bug if the secure bug is the first/only bug in the cookie) then process_bug.cgi will gladly spit the entire contents of the bug out to the user (at least I'm 90% sure that's the case, although I haven't looked at the code yet to be sure). The other is whitespace ;) I just noticed that you seem to be using two space indents, but the rest of bugzilla (and the emacs modeline in the files) use 4 spaces.
----- By removing all checks from bug_form.pl this has the potential to leak the information from process_bug.cgi. ----- The solution to this problem is to patch process_bug.cgi to validate the next ID on the user's bug list. ----- The other is whitespace ;) I just noticed that you seem to be using two space indents, but the rest of bugzilla (and the emacs modeline in the files) use 4 spaces. ----- Darn, I was hoping no one would notice. :-) Ok, the next patch will indent blocks 4 spaces. I'll leave multi-line statements indented two spaces since there seems to be no consensus among the code and because it seems to work well to distinguish those statements from blocks.
Patching process_bug.cgi sounds like the right solution, but I think I'd rather see ValidateBugID($::next_bug); Added at line 1038. To me that seems more obvious when looking at the code that this has been though of and also seems less error prone (such as modifications in the future). Multi-line statements seem to just indent to whatever seems right at the time (most of the time they line up w/the '(' if it's a sub or similar). I'd say the ones you're doing are fine w/2 spaces :)
Status: ASSIGNED → RESOLVED
Closed: 24 years ago
Resolution: --- → FIXED
This looks good. I couldn't reproduce the behaviour Jake was concerned about, though if someone can, feel free to reopen the bug. Checking in Myk's patch and closing out.
Moving to Bugzilla product
Component: Bugzilla → Bugzilla-General
Product: Webtools → Bugzilla
Version: Bugzilla 2.11 → unspecified
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: