2007/11/11, Anselm R. Garbe <[EMAIL PROTECTED]>: > On Sat, Nov 10, 2007 at 10:47:23PM +0100, y i y u s wrote: > > 2007/11/10, Anselm R. Garbe <[EMAIL PROTECTED]>: > > > Hi there, > > > > > > I plan to update the suckless.org website completely, before I'm > > > going to continue the st development. So I ask if people know > > > some sane and simple wiki/blog implementations, which I might > > > consider? > > > > > > Actually there were reasons why I wrote diri a while ago, but it > > > just feels too lowlevel and I more and more dislike the look and > > > feel of the suckless.org diri pages. > > > > > > If there are no wikis which might gain my interest, I will > > > rewrite diri from scratch using mainly C filter programs. The > > > first thing is a simplified Markdown -> html generator. > > > > > > I believe the basic principles and ideas in diri are right - > > > using a tree structure for glueing related pages together, > > > having a Markdown'ish wiki syntax which looks like > > > well-formatted ASCII text files (but lacks tables ;() etc. > > > > > > But maybe in the meantime some rocky wiki appeared which I > > > didn't noticed so far. Let me know! > > > > > > Regards, > > > -- > > > Anselm R. Garbe >< http://www.suckless.org/ >< GPG key: 0D73F361 > > > > I sent you my markdown implementation in awk a while ago. I know it is > > not perfect but it could be improved. I had the idea of applying your > > suggestions (also what webrc people had to say), but you didn't look > > interested. Tell me if this has changed, I can work on it now. > > cheers, > > I know your implementation and I had a look on it. From a > Markdown perspective it is quite well implemented. But what I > dislike about Markdown (which is not related to your awk > version) is that it has no higher-level features for tables and > that it recognizes different syntactical elements for the same > thing, e.g. you can define headlines differently with the same > result, where only one way should be allowed to make the format > consistent. > > I.e. what I'd like to have is Markdown without embedded HTML > (html will be escaped as in your version, at least as far as > discussed with Kris on IRC) but with the following table > syntax: > > > col | col | col > --------------- > val | val | val > --------------- > val | val | val > > and with a couple of other simplifications. I believe such > construct is quite context-sensitive and very hard to parse with > awk - hence I'm looking forward ;) > > Another idea I'd prefer is having a simple reference element, > e.g. > > ---snip--- > For further information about those problem refer (J.R. Hoare)[1]. > > ... > > References > ---------- > [1] http://www.jrhoare.com/papers/1963/anatomy.ps > ---snip--- > > What I have in mind is, that the markdown replacement should > add a link for the "J.R. Hoare" portion pointing to the site > referenced with the label "1". I believe that is heavily > context-sensitive and can only be done in two passes - so I see > heavy difficulties with awk here. > > That might give you some ideas about the silence from my side > about your implementation and markdown lately. > > Regards, > -- > Anselm R. Garbe >< http://www.suckless.org/ >< GPG key: 0D73F361 > >
Probably you are true and awk is not the best tool for such a parser, but I'll probably give it a try as an exercise (I'm still learning awk and I find it really interesting and... funny!). Anyway, I just wanted to tell that the reference method you are proposing could be a pain making modifications, i.e. you would have to change all the further references if you add a new one. That's why I propose a name system (you could use number as names if you want, but the parser would take care of putting them in order later). I will let you know if I give with something useful after some awk experiments... -- - yiyus || JGL .
