From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from neet.inri.net ([107.191.102.52]) by ur; Fri Sep 9 23:38:26 EDT 2016 Received: (qmail 37096 invoked from network); 10 Sep 2016 03:37:42 -0000 Received: from ds.inri.net (HELO dl.inri.net) (107.191.103.21) by neet.inri.net with SMTP; 10 Sep 2016 03:37:42 -0000 Date: Fri, 9 Sep 2016 23:38:19 -0400 From: sl@stanleylieber.com To: 9front@9front.org Subject: Re: [9front] fwd: [werc] ROADMAP FOR UPCOMING RELEASES Message-ID: <79b8182ff6589336977f142a083ca725@dl.attlocal.net> In-Reply-To: <20160910032228.GA17158@wopr> References: <7b832205137140e5144c6a5454e69a8c@dl.attlocal.net> <20160910032228.GA17158@wopr> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit List-ID: <9front.9front.org> List-Help: X-Glyph: ➈ X-Bullshit: optimized markup wrapper AJAX over SSL app standard-oriented database >> werc - 1.5 release >> apps/mdir # by khm, test it! > > It is slow as shit. This might be okay with sufficient warnings added. Disclosure: This list was copied directly from my personal todo file. I agree; this stuff is not yet ready for primetime. Particularly because of the system load generated by its use. But we can work on it. >> apps/paste # by khm # DONE > > I consider apps/paste a proof-of-concept. It is crippled by rc's > env-var size limit and should be replaced with something that uses > maht's upload code or such. Sounds good. >> apps/upload # generic file uploads # disabled by default >> /n/lists/werc/1240113746.00 # maht's upload code > > This code works and I have not been able to break it. I also don't have > a patch prepared to integrate it into werc's mainline. I have not even looked at this yet. >> apps/wman # FIX >> mysterious (erroneous) escapes are printed here: >> http://man.cat-v.org/9front/3/ip > > A cursory glance didn't show obvious problems. I can assume this duty > if someone can give me details about the bug. Weird shit gets printed in the web browser. Scroll down to input packets delivered to higherfrlaegvmeelntpsroctroecaotlesd I started looking into this once but lost my notes. No memory of what I found. >> make it possible to never load the default css # bind? > > I think this should be a config flag, since filesystem games are not > necessarily portable -- you need root to bind on some OSes, and while > that sucks, werc should keep it in mind. Agree. Some refactoring will be needed to make this happen but it's not hard to make it a knob. >> werc - 1.6 release "UNSETTLING DREAMS" >> apps/gregor >> bugtracker > > We need a separate todo for this. The 9front bugtracker is a beautiful > start; I propose it needs only further integration into the email side > to be perfect. Specifically, replying to a bug on the list should show > up on the bug report in werc. We can cannibalize apps/mdir for pieces > of this. Perhaps too ambitious for inclusion in the following release, but that's why it's not listed for 1.5. Agree we could break it out into a solo project and then add it in once it works. >> apps/paginate > > What is this? Arbitrarily break up the contents of a given page into multiple pages, inserting navigation buttons on each paginated page. I implemented this in barf but my solution is not general enough for werc (this should work with ANY content). >> apps/tags >> tagfs? > > IMO this should support the ability to use external tools like kenji's > database thing. Yes. A simple implementation could just be a _werc/tags index file, formatted like this: index.md hj,dicks,murray_hill,hi_rob,fuck_you Or something along those lines. I implemented tagging in barf in a slightly more complicated and less general way. >> mkfile - use werc to generate static sites > > I feel like this has been at least started on by someone, somewhere > in the mists of time. Am I hallucinating? I implemented this at least twice and threw it away both times. >> http://plan9.stanleylieber.com/src/markdown.go >> revise whitelist >> test with cat-v.org > > MORE LIKE MARKDOWN.NO > > If we're going to include markdown.go, we should make a contrib/go > directory and bundle cgo in there too. Well, the existing contrib/ includes stuff in more than one language. This note (and all the non-portable) stuff was mainly for testing on 9front; some of the stuff can be made general enough to run anywhere. sl