Along the way I learned that when you tweak the base xsl you have to run 'build.sh bootstrap' to recreate the language-specific files. In 20 years I have never had to do that. Apparently we were still running the original stuff that Joshua Slive lovingly hand crafted in 2003 or thereabouts when we switched from writing raw html.
So cheers to Joshua, wherever you are! Rich On Sun, May 31, 2026, 11:25 Lucien Gentis <[email protected]> wrote: > Le 31/05/2026 à 13:48, Rich Bowen a écrit : > > > On Sat, May 30, 2026, 18:33 Rich Bowen <[email protected]> wrote: > >> I’ve gone ahead and patched trunk, but note that this means I’ve also >> patched the build/ repo, so I need to go ahead and patch 2.4 pretty soon, >> or things will be kinda broken in the meanwhile. >> > > > https://validator.w3.org/nu/?doc=https%3A%2F%2Fhttpd.apache.org%2Fdocs%2Ftrunk%2Fbind.html > > > We're down to one warning and one error, and neither one of them > particularly concern me. I'll push this over to 2.4 today or tomorrow. > > Note that there's related problems with the httpd-site repo but since that > site is generated completely differently I haven't looked at that yet. I > expect it's related but that the fix will be different. > > I tested trunk : > > === TESTING === > > 1. Apply patch, rebuild one language: ./build.sh en > ./build.sh validate-xml : OK > ./build.sh fr : OK > ./build.sh validate-xhtml : OK > > 2. Inspect generated .html — verify DOCTYPE is HTML5 > First two lines of bind.html.fr.utf8 : > <!DOCTYPE html SYSTEM "about:legacy-compat"> > <html lang="fr"><head><META http-equiv="Content-Type" content="text/html; > charset=UTF-8"> > > 3. Run: java -jar vnu.jar --skip-non-html ../bind.html.en.utf8 > Ran directly ./build.sh validate-xhtml : OK > > 4. Confirm void elements are not self-closed (<br> not <br/>) > No <br/> found in manual directory > > 5. Spot-check pages render identically in browser > I committed bind.html.fr.utf8, didn’t see any rendering problem. > > > All things seem to be OK > > Great job Rich ! >
