regarding FS#2864 my approach would be to see if HTTP_REMOTE_USER is available and copy it to REMOTE_USER if that one's empty. and I would do that centrally in auth.php so all auth backends can profit. any better ideas?
I guess we're not using the roadmap in the bug tracker for this next release ... do you have a list of bugs to fix somewhere? ... it's not all of devel:reviewablebugs, is it?
Michitux, yes, using the bug tracker's roadmap is usually preferred, but this time it's too late for that and no-one seems to be motivated to do that anyway
E_STRICT is supposed to be triggered when the php is compiled (ie. the file is read), so we should have most, if not all. i doubt there are many files untouched by tests
because while the button and link code still supports it the actionOk()-function doesn't which means the subscribe link is always displayed even when subscriptions are disabled
the arctic template has a lot more rather old code, for example it tries to remove the ids from the headers in the sidebar but still checks for the old html code
selfthinker: I think it isn't worth trying to fix all the templates, there are more serious problems that are relevant for users like not using tpl_favicon
yeah ... on the other hand (in a slightly sadistical way) I think it is *good* that templates break (visibly) ... otherwise users will never be aware of the fact that they are not updated ... but a template can break in a bad way (e.g. when we had tpl_sidebar() in the core) or in a "good" way (e.g. the broken TOC)
I think if we disable the subscription action compatibility people will complain that they can't find the subscription functionality when they enable it and they won't realize that the problem is the template
hm, I have a rewritten info plugin. it doesn't contain a single piece of html code anymore and is almost independent of the output format. the only remaining problem is how to determine if the last section before the plugin output needs to be closed/if the last one needs to be left opened.