On Sat, Feb 22, 2014 at 1:49 PM, Anselm R Garbe <garb...@gmail.com> wrote: > On 22 February 2014 14:42, Anselm R Garbe <garb...@gmail.com> wrote: >> On 22 February 2014 14:37, Markus Teich <markus.te...@stusta.mhn.de> wrote: >>> Anselm R Garbe wrote: >>>> All I want is to have full logging of all output when building various >>>> stuff, >>>> now stali related stuff. >>> >>> I would just pipe it to less. >> >> I rather want persistent logs. And I want it implicit. Sometimes I >> realize, oh dang, what was printed before and then it's too late >> already to launch less ;) Actually this is the main reason for a >> scrollback buffer IMO, it'll be only used if you forgot to pipe the >> output somewhere... >> >> Back in the older days, I actually intended towards a totally >> different st design as it is today. Originally it was planned to >> define a new client/server model, where st would've been just a client >> with an own formatting language to a st server. The st server was >> intended to provide full persistent logging and filter programs for >> the different terms that it would have translated into the st >> formatting instead. >> >> Nowadays st is just a simpler terminal ;) > > Having said this, when I discussed with uriel about it several years > back in time, we really thought it to allow attaching to the same > terminal from various st clients, so taking terminal multiplexing a > bit further. We also envisioned that it could replace ssh and remote > shell hacks at some point, as an st client would define a saner > interface than escape sequences and termcaps/terminfos. All the ugly > task would have been down to the filter in between which must be aware > of the actual escapes and translating them into the st formatting > language.
This is very interesting. The main complaint I have towards dtach(1) (I have the same complaint about screen(1) and tmux(1), but for those it's not the *main* complaint) is the need to first ssh into an interactive session. I'd like to do `dtach -a <user>@<remote>:<path-to-session>`. > > Perhaps this will be achieved in the future. The st formatting was > also thought to provide a "text mode" like interface, but a > *graphical* one. As most people do use graphical environments these > days... > > I hope this sheds some light into the direction I'm looking when > asking those really basic questions ;) Relevant to this discussion: http://lubutu.com/idea/ivo Cheers, -- ______________ Raphaël Proust
