Am Sat, 27 Feb 2010 12:57:59 +0900 (JST) schrieb Teika Kazura <[email protected]>:
> How about debugger? Today for the first time I tried it, but it works > for the simplest case. Since it says byte-compiled files are not > supported, some care is needed. > > Now the procedure: > 1. copy lisp/sawfish/cfg/main.jl.in (in source) as main.jl. > 2a. In main.jl, replace 'nokogiri-dir' with > /usr/share/sawfish/1.6.2/lisp, and touch all lisp/sawfish/cfg/* > files. (If jl source is newer than jlc, then jl is read.) > 2b. Even better, replace 'nokogiri-dir' with /your-git-repo/lisp/, > since there's no jlc files there. > 3. Add (setq debug-on-error t) somewhere in main.jl > 4. rep main.jl > > When you enter the debugger, type 'backtrace'. (Tab completion > available.) > > On Thu, 25 Feb 2010 00:04:31 +0100, Christopher Roy Bratusek wrote: > > Bad arguement: #<closure 10adc40>, (1 1 40 32), 1 > > [...] > > ... since I didn't change SawfishConfig since 1.6.0, I may be some > > incompatibility between rep-gtk and gtk+ 2.19.5 (which I'm using), > > though I'm not 100% sure here. > > Hmm, but it says 'closure'. rep-gtk funcs are C, so 'subr', no? I doubt > add-matcher thingy, or bogus key binding. > > Regards, > Teika (Teika kazura) > compiler doesn't bring any benefit. Well after disabling custom frame-font colors the bad-arg #closure became a bad-arg #subr. Though frame-font colors are *not* the cause, as it still happens afterwards, I've now disabled almost all of my configs and begin to track down the root of evil. Chris -- Re[3]: Nicht ganz glücklich Gesendet von M. Mrotzek am Sa, 1. Aug um 4:55 Was willst du? Bitte sprich Deutsch mit uns. Dieses Fachgesimpel kann doch niemand nachvollziehen... Pff, immer diese Geeks. Halten sich wohl für etwas Beseres. Das war schon immer so. :(
