Hi Nick, Nick Dokos wrote: >> The "automagic C-c C-c" should be NOT[1] done after each key press or >> some such. That certainly would be a killer feature, in its real >> acception: performance would be unbearable. >> >> In my mind, automatically (re-)parsing the meta options should be each >> time the user presses `C-c C-v C-e' (eval code blocks); that is, when >> the user expects his options to be taken into account. >> >> Does it make sense? >> >> Best regards, >> Seb >> >> Footnotes: >> >> [1] This word was missing (in the original post)! >> > > Well, it might make sense but you can try it out and let us know: > > - make files with 10, 100, 1000 trivial (or even empty) code blocks, just > enough to make sure that org-babel-execute-maybe is really called on them: > I think that it will be called even on empty code blocks, but I'm not sure > if there is some optimization in there. > > - measure the time it takes to export each one to html (say). > > - add a call to org-mode-restart into org-babel-execute-maybe, and time the > same operation again: how significant is the slowdown? > > If the slowdown is bearable in these cases, then it will be bearable in > realistic situations, where block execution is going to be a much more > significant fraction of the total.
I'll give it a shot, and report the pre/post results "à la Weight Watchers". Thanks for pointing out some detailed calls I have to make for testing that idea! > BTW, what's the biggest file you (all, not just Seb) have in terms of the > number of code blocks it contains? In my case, the largest one had about two > dozen code blocks, so the 100 case would easily cover me, but I suspect > there are much bigger ones out there. The biggest number of code blocks in any document I have is around 20. Best regards, Seb -- Sebastien Vauban
