On Sunday I climbed into the hayloft for the first time this year. I spent most of the day setting up for the coming “building season”, sweeping out a winter’s worth of pigeon mess and making lists.
And just because I wouldn’t feel like it truly was spring until I’d heard the buzz of an orbital sander I put in a couple of hours sanding primer. I was in such a good mood to be back at work that I caught myself trying to whistle through the respirator.
It took awhile but I finally got around to submitting version 1.4.6 of RMD to addons.mozilla.org.
RMD 1.4.6 supports Firefox and Thunderbird versions 1.5 through 2.0.12 on Windows, OSX and Linux.
https://addons.mozilla.org/en-US/firefox/addon/2131
(And just a reminder that RMD will not be updated to support Firefox 3).
Last month I put together a 1.4.6 build of the Remember Mismatched Domains extension for Firefox and Thunderbird. This update is pretty much just a “maintenance release” making RMD more compatible with SELinux as well as fixing a couple of my usual localization screw ups. (Thanks to Josh for both pointing out the SELinux issue as well suggesting how to solve it).
The more exciting announcement for this update is that I’m hoping version 1.4.6 will be the final release. Back in October I wrote about the new security exception features in Firefox 3 which essentially provide the functionality that RMD has been providing for the past couple of years. Since then Firefox 3 has been steadily moving through the beta stage and soon enough will be released. (Thunderbird users will be happy to hear that the same security exception features are available in the nightly builds of the still very early Thunderbird 3). What this all means in the long run is that RMD will not be available for Firefox 3. Please note that I have every intention to continue maintaining the extension for Firefox 2 for as long as it is a Mozilla supported product.
Now’s the time to voice your questions, concerns…
I just made a quick update to the FDTracer plugin for FlashDevelop. This fix resolves an issue with FlashDevelop beta 4 and selecting FDTracer’s “Settings…” option from the plugin’s context menu.
Along with the revisions to the FD Settings Manager in beta 4 came a couple of API changes so you should only use this version of the plugin with beta 4 (or greater). If you’re still using beta 3 I’d say it’s a good time to update both.
Download FDTracer 0.3.2:
http://www.andrewlucking.com/code/fdtracer/
Today, nearly 4 years, 123 comments and 42 votes after having been opened, Mozilla bug 228684 – Remember overrides of Certificate Domain Name Mismatch was updated to Resolved.
The cold weather is setting in and I’ve started looking for tasks on the boat that don’t need epoxy or paint to cure. A couple of weekends ago I cut holes for the windows.
I’ve essentially closed up shop for the winter but with the crazy fluctuations in temperature we’re having this fall I’m hoping I’ll get in a bonus day or two.
I stumbled across a bit of a performance problem with FDTracer last week. While working in FlashDevelop I noticed that things were beginning to respond a little slowly. The problem gradually grew worse and my system resources manager clearly showed that some part of FD was the culprit. After a bit of digging around it was obvious that FDTracer was at the root of it all.
I happened to be listening to a Flash video stream in the browser while I was working and the video’s swf player was tracing out debug data every 10 milliseconds or so. After about a 1/2 hour of this my debug Flash Player’s log file had grown to many thousand lines and FDTracer was chugging to parse the contents.
Upon replacing a String concatenation or two with a StringBuilder routine FDTracer was once again humming along. The performance gains were pretty astounding and definitely taught me a lesson I won’t forget anytime soon.
Download FDTracer 0.3.1:
http://www.andrewlucking.com/code/fdtracer/
A patch was recently checked into the Mozilla codebase that marks bug 327181 “Improve error reporting for invalid-certificate errors” as Resolved Fixed. This bug significantly changes the way Mozilla will handle invalid certificates and as a result will hopefully make RMD redundant so earlier today I played around with a nightly build of Firefox 3.0 alpha (Minefield). My impression? I think the changes, while still rough around the edges, do a pretty good job of appeasing the various sides in the bad certificate handling saga.
The first thing you’ll notice when browsing to a website that presents an invalid security certificate is that the familiar mismatched domain error dialog is no more. Instead an in page Secure Connection Failed error is presented. Note that there is no way to temporarily accept the mismatch and visit the website:
Before websites that present invalid certificates can be visited the certificate must first be added to the new security exceptions list. The dialog for this list is intentionally buried in the Preferences / Options dialog:
Edit > Preferences > Advanced > Encryption > View Certificates > Servers -> Add Exception…
For current users of RMD I think it would be helpful to get your impressions of the changes and so I put together a little screencast of the new functionality in action. I encourage you to take a look and provide Mozilla with your feedback. Even better would be to download a recent build of Minefield and try it out for yourself.
Version 0.2 of The FDTracer plugin for FlashDevelop 3 adds a couple of setting options:
Open Panel On Compile is the result of an improvement suggestion. When this option is selected the FDTracer panel will always be displayed after a compile completes.
Colour Warnings and Errors adds some colour to the trace statements. Warnings and errors generated by the Flash Player appear in orange and red fonts respectively.
Download FDTracer 0.2:
Download FDTracer 0.3:
http://www.andrewlucking.com/code/fdtracer/
Update:
Argh. I used version 0.2 all week before posting it last night and then this afternoon I came across a bug that was dropping digits that begin a trace statement (eg, trace(”1,2,3″); would display as “,2,3“). Version 0.3 should fix that silliness.
