Hiya, A quick note that the icedtea-web-1.6.1-r1 ebuild wasn't tested before it was pushed. The respect-ldflags patch fails to apply, and checking the error output it shows that it was attempting to patch Makefile.am.orig (and checking the 1.5 version of the same patch, it looks like it was fixed up manually to point to Makefile.am despite being based off of Makefile.am.orig to being with). Carrying out patching this way seems like it would lead to this kind of mistake often. I'd strongly recommend changing the patching method from "applying a patch, fixing it up and then doing a manual diff on the .orig file" to one where the entire code repo is committed to git, and then a git diff done at the end. It would ensure that mistakes like this can't creep into the code base again in the future. As it stands, the patch file needs fixing in the tree. Let me know if you need any additional information.
You managed to catch this in the 12 minutes before I set it right. :( As a general rule, I do test my changes thoroughly before I push but my daughter was distracting me at the time. Updating that patch wasn't the main issue I was seeking to address in that commit, I just thought it would be good to replace it with the fuller patch I ended up submitting upstream. I noticed soon after that something was wrong and figured I must have based the upstream patch on newer code so I effectively reverted it. I'm not sure how the Makefile.orig thing crept in. I often do create my patches from git but upstream uses Mercurial in this case, which I am less comfortable with.
Hehehe, no problem, sorry I was a bit quick off the mark. Even if upstream's using mercurial I'd still do my patching by downloading their official source tarball and then doing a "git init; git add .; git commit", but each to their own. Thanks for getting it all fixed up! 5:)
