Hi, I will be unable to work on Dillo or reply for the next weeks, probably around mid-October I can come back. On Fri, Sep 25, 2026 at 05:09:47PM +0100, 256@256-32.com wrote:
I want to improve Dillo in a bunch of ways, most urgently by adding support for the "expires" HTTP header field, and improving the HTML/CSS rendering. I'm worried, though, that if I submit a huge patch to do one of these things it'd get rejected for not being minimalist enough.
The current blocker from implementing this is the infrastructure to test that new features work fine. I wrote dilloc with this kind of tests in mind so we can do more complicated HTTP tests and see if it works properly. Until we don't have the tests I don't think I want to implement new features we cannot test (other than manually). I'll suggest implementing first the tests that fail due to unsupported cache support. Take a look at https://github.com/dillo-browser/testsuite (pending to be moved to our server) and see if there is something in WPT that we can reuse to test dillo (without JS). We can merge the failing tests first, then focus on the implementation in dillo for expires and other headers and check that the test pass. In the testsuite we can use python or other high-level tools that make writing tests easier than C (that is why it is separated from the dillo code).
My question is, would substantially fleshing out the layout and style engines be seen as going against the spirit of Dillo?
The layout engine has nothing to do with expires. I would recommend not touching the layout engine for now (mostly for the same reason, lack of tests).
Would it push too far against the 1.44MB limit?
It is unlikely that you will pass that limit with handwritten code, as gzip will compress your code with the current codebase. Don't worry too much about that, there are stuff pending to be removed that will make some room and I have some provisions to make more room if that limit is reached. Thanks!, Rodrigo.