egeezer wrote:FYI- Installation of Firefox 1.5.0.4 disables UndoCloseTab.
There is no 'Firefox 1.5.0.4', and it was already compatible with Firefox 1.5.0.3 at the time of your post.
zeniko wrote:Again a missing release candidate (for Firefox 1.5.0.3), I suppose.
This time there was a release candidate (in releases/1.5.0.3/update/), and I've updated my update.rdf a few days ago. It doesn't seem like there were any announcements regarding the RC.
zeniko wrote:Do you really test your extensions on all released versions of Firefox (Linux and Mac builds and all locales)?
There is no need to test all platforms because authors are not forced to claim that an extension works with a specific platform (as they are with versions).
zeniko wrote:And since the devs promised to keep the 1.5.0.* API compatible [...]
Firefox 1.0.x releases were supposed to not break any extension but it still happened (IIRC one author said that three releases broke her/his extension). While Firefox 1.0.x releases were supposed to only have security related changes Firefox 1.5.0.x releases are supposed to include other changes, so they are much more likely to break extensions.
zeniko wrote:[...] why would you suppose that your extensions could break at a security update?
Because there were already security updates which have broken extensions (various in 1.0.x, but 1.5.0.2 might have broken some with
xpcom). Also undoclosetab uses runtime patching, and as with any form of patch it will fail if the underlying code has been changed and the reference points were modified.
zeniko wrote:As for claiming 2.* compatibility (or even 1.5.*), that's obviously a DON'T since there might be API changes (e.g. nsIFormHistory has been removed from the 1.8.1 branch)...
I wonder why some of your extensions have a maxVersion higher than 1.5.0.* then...
zeniko wrote:[...] you're preventing an offline installation of your extension;
Offline installation of extensions is something the devs intentionally dropped support for (but it is still possible by manually setting app.extensions.version).
zeniko wrote:In any way, your users have a much bigger chance to get an inconvenience than what you're trying to protect them from...
Blame the devs for not providing a way to ignore compatibility checks for security releases (a warning would be enough). It isn't my responsibility to work around an issue they intentionally introduced.
Also I am not 'trying to protect' someone, I simply don't like being forced to claim something I can't possibly know...