Bugmail filtering

2026-03-06T00:00:00+00:00

Keeping up with Bugmail can be a challenging task. After using Bugzilla for a while, you become CC'ed on quite a few bugs and maybe also subscribe to one or more components. With time, the amount of mail you receive can become overwhelming and you might wonder how to keep on top of it. In this post, I'm going over the setup that I've built in the 4 years I've been at Mozilla.</p>

Setup</h2>
My setup has changed over time. In the beginning I used Thunderbird filters</a> to filter on bugmail and added rules over time. This already allows to sort mail into the useful folders and tag important mails. However, you can't modify mails with Thunderbird filters.</p>
The new setup I'm going to present here allows rewriting mails and do complex filtering with the mail filter language sieve</a>. The setup has been in place for half a year.</p>
I'm hosting a mail server on a Raspberry Pi to fetch new mails and run sieve scripts on them, before I retrieve them via imap from my Thunderbird client. The mails are fetched via fetchmail</a> on the Raspberry Pi, and then delivered to dovecot</a> with postfix</a>. Dovecot runs my sieve scripts on all incoming mail.</p>
</p>

Bugzilla Mail Headers</h2>

Bugzilla adds meta-information to all bugmail it sends out. The headers look like this:</p>

X-Bugzilla-Reason: Reporter AssignedTo</span></span>
X-Bugzilla-Type: changed</span></span>
X-Bugzilla-Watch-Reason: None</span></span>
X-Bugzilla-Classification: Components</span></span>
X-Bugzilla-ID: 1992712</span></span>
X-Bugzilla-Bug-Type: enhancement</span></span>
X-Bugzilla-Product: Core</span></span>
X-Bugzilla-Component: Privacy: Anti-Tracking</span></span>
X-Bugzilla-Version: unspecified</span></span>
X-Bugzilla-Keywords: </span></span>
X-Bugzilla-Severity: N/A</span></span>
X-Bugzilla-Who: csabou@mozilla.com</span></span>
X-Bugzilla-Status: RESOLVED</span></span>
X-Bugzilla-Resolution: FIXED</span></span>
X-Bugzilla-Priority: P2</span></span>
X-Bugzilla-Assigned-To: manuel@mozilla.com</span></span>
X-Bugzilla-Target-Milestone: 145 Branch</span></span>
X-Bugzilla-Flags: </span></span>
X-Bugzilla-OS: Unspecified</span></span>
X-Bugzilla-Changed-Fields: Last Resolved Resolution Status status-firefox145</span></span>
 Target Milestone Comment Created</span></span>
X-Bugzilla-Changed-Field-Names: cf_last_resolved resolution bug_status</span></span>
 cf_status_firefox145 target_milestone comment</span></span></code></pre>Enhancing Subject/From</h2>
In my sieve script I take these headers and make quite a few modifications:</p>


I take the X-Bugzilla-Who</code> and set the From</code> to this value making it show up as correspondent in Bugzilla.</p>
</li>

I take X-Bugzilla-Priority</code>, X-Bugzilla-Severity</code>, X-Bugzilla-Status</code>, X-Bugzilla-Resolution</code> and extend the brackets in the mail header that usually only contains the bug number with these fields when they are at non-default value. Upper casing the values if they have changed with this mail.</p>
</li>

The X-Bugzilla-Changed-Field-Names</code> is used to add emojis after the []</code>, to make it clear on a glance what the change is about.</p>
Field</th> Emoji</th></tr></thead>

short_desc</td> 🏷️</td></tr>
comment</td> 💬</td></tr>
attachment</td> 📎</td></tr>
assigned_to</td> 👤/👥</td></tr>
blocked</td> 🧱</td></tr>
depends_on</td> 🔨</td></tr>
see_also</td> 👀</td></tr>
cc</td> 🅭</td></tr>
bug_file_loc (url)</td> 🌐</td></tr>
status_whiteboard</td> 🪧</td></tr>
keywords</td> 🔑</td></tr>
cf_user_story</td> 📇</td></tr>
regresses</td> 📉</td></tr>
regressed_by</td> 📈</td></tr>
bug_group (security group)</td> 🙈</td></tr>
dep_changed (not this bug changed, but depending)</td> ⛓️</td></tr>
</tbody></table>
</li>

The X-Bugzilla-Bug-Type</code> turns into an emoji at the beginning of the subject line</p>
Type</th> modified</th> unchanged</th></tr></thead>

defect</td> 🐞</td> 🪲</td></tr>
enhancement</td> 🧰</td> 🔧</td></tr>
task</td> 📌</td> 📝</td></tr>
</tbody></table>
</li>
</ol>
Here is a visualization of how the mail looks like after the info was added and where the info is coming from:</p>
</p>
Involvement</h2>
Then I'm also using the bugmail headers to sort the mails by how involved I am within the bug.
I'm sorting the bugs into these categories:</p>

Directly involved: Sort into Assignee</code> / Reporter</code> / needinfo</code> / CC</code></li>
Team components: All components my team owns, some of which I help triage</li>
Interesting components: I loosely read bugmail there. Mostly components where I have past involvement (Necko) or larger team involvement, where it is helpful for me to sometimes chime in. I sometimes mark all mail of these components as read when I haven't taken a look for a longer time there.</li>
</ol>
</p>
Highlighting bugs</h2>
I also tag some more bugs to highlight them in Thunderbird:</p>

Actions from outside contributors: I usually try to read all actions by outside contributors, because they usually don't set the needinfo tag and their comments have the tendency to get lost.</li>
New bugs: To be aware of what is on file.</li>
I also directly mark some known-to-me uninteresting bugs as read</li>
</ul>
Further Resources</h2>
This blog post is based on a lightning talk I gave on 2025-11-13. You can watch the recording or take a look at the slides.
I gave a brief view into the source code.</p>

Lightning Talk Recording</a></li>
Lightning Talk Slides</a></li>
</ul>


Integrating a Rust library into Gecko
2025-07-10T00:00:00+00:00
Integrating a Rust library to Gecko / Firefox is unexpectedly easy, but Rust adoption varies significantly between teams.
I started in the Firefox networking Team Necko</a>, which has a few Rust components like the HTTP/3 stack neqo</a> or the URL implementation.
After switching to Privacy / Anti-Tracking team</a> I'm now in a team that has no Rust code at all.
The interest for Rust components is there, but usually projects are started in C++ due to all code being in C++.</p>
Rust in Gecko</h2>
</a></p>
Looking at current language stats, we can see that Rust is currently at around 12% of Firefox code with almost 5 million lines of code.
This is not neglectable, but falls short behind C / C++ and also JavaScript.</p>

Note:</strong> Using tokei with --exclude=third_party</code> suggests Rust has about 8% of the LOC of C++ for code written in mozilla-firefox directly. Not 50% as this diagram might suggest. (feedback after lightning talk)</p>
</blockquote>
Error categories rules out by Rust</h2>

Use after free bugs</strong>: No more kungFuDeathGrip</code> security fixes to extend the lifetime of an object.</li>
Data races</strong>: No more forgetting to question "Do I need to lock a Mutex to access this variable?"</li>
Nullptr dereferences</strong>: Call by pointer is very common in C++, especially in idl interfaces. "Can this pointer I’ve got passed be null?" is now answered by the Rust type system.</li>
Iterator invalidation</strong>: Another big class of errors. In Gecko this is also memory safe for nsTArray</a>, but wrong code.</li>
Other undefined behavior</strong></li>
</ul>
Other Rusty reasons</h2>
In my opinion, Rust is just a very well-designed language.
It has great documentation infrastructure and awesome compiler errors aiding you in writing safe code.</p>
Case Study: RIIR "Copy Clean Link"</h2>
The patch for the case study can be found in D225390</a>.</p>
Almost always you probably want to split up the library into two parts:</p>
One crate for functionality and bindings.
The functionality crate can be developed out-of-tree. In this example I'm going to have the crate be developed in-tree.</p>
Add crate directories to moz.build</h3>
Add directory where the crates are located need to be included in the parent moz.build:</p>
DIRS</span> +=</span> [</span></span>
    "urlstrip"</span>,</span></span>
    "urlstrip_glue"</span>,</span></span>
]</span></span></code></pre>Functionality crate</h3>
The crate implementing the logic with no dependency on xpcom types simply needs to have a single-line (plus license text) moz.build</code> next to the Cargo.toml</code>:</p>
FINAL_LIBRARY</span> =</span> "xul"</span></span></code></pre>

Note:</strong> To combat build failue you need to regenerate the Cargo.toml</code> by running ./mach vendor rust --allow-dirty</code>.</p>
</blockquote>
cbindgen / XPCOM bindings crate</h3>
You need a few more lines for the bindings crate to autogenerate a C++ header with cbindgen.</p>
FINAL_LIBRARY</span> =</span> "xul"</span></span>
EXPORTS</span>.mozilla</span> +=</span> [</span></span>
    "UrlStrip.h"</span>,</span></span>
]</span></span>
if</span> CONFIG</span>[</span>"COMPILE_ENVIRONMENT"</span>]:</span></span>
    CbindgenHeader(</span></span>
        "UrlStrip_ffi.h"</span>,</span></span>
        inputs</span>=</span>[</span>"/toolkit/components/antitracking/urlstrip_glue"</span>]</span></span>
    )</span></span>
    EXPORTS</span>.mozilla</span> +=</span> [</span></span>
        "!UrlStrip_ffi.h"</span>,</span></span>
    ]</span></span></code></pre>
You can depend on nsstring</code> and other xpcom</code> crates. However, you won't</a> be able to use Rust's native #[test]</code> directive. This is one of the reasons you likely really want to have the separation between the two crate types.</p>
To configure cbindgen, get inspired by taking a look at other cbindgen.toml files</a>.</p>
You also want to have somewhat nicer C++ headers around your autogenerated cbindgen header for C++ UrlStrip.h</code>:</p>
struct</span> UrlStrip</span> {</span></span>
public:</span></span>
 static void</span> Create</span>(</span>UrlStrip</span>**</span> pThis</span>) {</span> urlstrip_new</span>(pThis); }</span></span>
 void</span> Delete</span>() {</span> urlstrip_drop</span>(</span>this</span>); }</span></span>
</span>
 void</span> Init</span>() {</span> urlstrip_init</span>(</span>this</span>); };</span></span>
 void</span> Uninit</span>() {</span> urlstrip_uninit</span>(</span>this</span>); };</span></span>
</span>
 nsresult Strip</span>(</span>const</span> nsACString</span>&</span> aURI</span>,</span> nsCString</span>*</span> aOutURI</span>) {</span></span>
   return</span> urlstrip_strip</span>(</span>this</span>,</span> &</span>aURI, aOutURI);</span></span>
 }</span></span>
 bool</span> CanStrip</span>(</span>const</span> nsACString</span>&</span> aURI</span>) {</span></span>
   return</span> urlstrip_canstrip</span>(</span>this</span>,</span> &</span>aURI);</span></span>
 }</span></span>
</span>
private:</span></span>
 UrlStrip</span>()</span> = delete</span>;</span></span>
 ~UrlStrip</span>()</span> = delete</span>;</span></span>
 UrlStrip</span>(</span>const</span> UrlStrip</span>&</span>)</span> = delete</span>;</span></span>
 UrlStrip</span>& operator</span>=</span>(</span>const</span> UrlStrip</span>&</span>)</span> = delete</span>;</span></span>
};</span></span></code></pre>
Maybe cxx</a> could ease with this in the future: Bug 1921139 - Allow creating C++ bindings for Rust code with cxx</a>.</p>
Other files to touch</h3>
To use the Rust library from C++/JS, it must be linked. And for that you need to both add it as a dependency to toolkit/library/rust/shared/Cargo.toml</code>:</p>
urlstrip</span> =</span> { path</span> =</span> "../../../components/antitracking/urlstrip"</span> }</span></span>
urlstrip_glue</span> =</span> { path</span> =</span> "../../../components/antitracking/urlstrip_glue"</span> }</span></span></code></pre>
And also important (due to still using 2015 rust edition) extern crate</code> definitions in toolkit/library/rust/shared/lib.rs</code>:</p>
extern crate</span> urlstrip;</span></span>
extern crate</span> urlstrip_glue;</span></span></code></pre>Done</h2>
This is all.
Feel free to reach out to me if you have any questions!</p>
Further Resources</h2>

https://firefox-source-docs.mozilla.org/build/buildsystem/rust.html</a></li>
https://firefox-source-docs.mozilla.org/writing-rust-code/</a></li>
Blog as Lightning Talk</a></li>
Lightning Talk slides</a></li>
</ul>
Updates</h2>

2025-07-10: Added note on Rust usage within gecko</li>
2025-11-02: note on how to combat build failure after adding rust crate</li>
</ul>


Introducing Torpatch.es
2025-04-23T00:00:00+00:00
torpatch.es</a> is a dashboard to aid uplifting patches from Tor-Browser into Firefox for the Tor Fusion</a> project.
It shows currently worked items and links uplift bugzilla bugs to tor browser bugs.
The dashboard has been in progress since end of 2024, but has reached a state of completeness with only minor changes outstanding.
torpatch.es</a> was inspired by torpat.ch</a>.</p>
The project collects information from several sources and mashes them together in a single html file per tor-browser branch:</p>

Git repositories:

gecko-dev</a> for Firefox source.</li>
tor-browser</a>, commit on esr branches, that follow the scheme described in the repository overview</a>.</li>
mullvad-browser</a></li>
</ul>
</li>
Gitlab issues (downloading all issues):

tor-browser</a>: most issues reside here</li>
trac</a>: some issues are still being referenced, but weren't migrated to tor-browser issue tracker</li>
mullvad-browser</a></li>
tor fenix</a>: nowadays part of gecko-dev / tor-browser, but issues are still referenced from commit messages</li>
tor android-components</a>: also archived, but still referenced</li>
</ul>
</li>
Bugzilla: all bugs that have [tor]</code> or [tor #]</code> in the whiteboard. The [tor #]</code>-whiteboard tag is used to match uplift bugs in bugzilla to tor-browser bugs.</li>
</ul>
The source code is available under MPLv2 on gitlab.torproject.org</a>.</p>

- firefox

Bugmail filtering

Integrating a Rust library into Gecko

Other Rusty reasons</h2> In my opinion, Rust is just a very well-designed language. It has great documentation infrastructure and awesome compiler errors aiding you in writing safe code.</p>

Functionality crate</h3> The crate implementing the logic with no dependency on xpcom types simply needs to have a single-line (plus license text) moz.build</code> next to the Cargo.toml</code>:</p>

Done</h2> This is all. Feel free to reach out to me if you have any questions!</p>

Updates</h2> 2025-07-10: Added note on Rust usage within gecko</li> 2025-11-02: note on how to combat build failure after adding rust crate</li> </ul>

Introducing Torpatch.es

Other Rusty reasons</h2>
In my opinion, Rust is just a very well-designed language. It has great documentation infrastructure and awesome compiler errors aiding you in writing safe code.</p>

Done</h2>
This is all. Feel free to reach out to me if you have any questions!</p>

Updates</h2>

2025-07-10: Added note on Rust usage within gecko</li>
2025-11-02: note on how to combat build failure after adding rust crate</li> </ul>