You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, if we let Scm_Eval/Scm_Load to catch errors (that is, no "propagate error" flag), original stack trace is lost. (Cf. #521 ). We don't want to attach stack trace to the thrown condition unconditionally, for taking stack trace is expensive.
However, we may use a dynamilally scoped flag that tells VM to save the stack trace (or, save the continuation chain and keep it) in the condition. Especially, Scm_Load is typically load programs but does not execute long-running code, so it is less critical for error performance.
The text was updated successfully, but these errors were encountered:
#1081
This is a temporary solution. We need to repeat cleanup code
for `SCM_UNWIND_PROTECT` and error-trapping `Scm_Load` call, which
is awkward.
Better solution is with the line of #1079,
but that depends on continuation management overhaul.
Currently, if we let
Scm_Eval/Scm_Loadto catch errors (that is, no "propagate error" flag), original stack trace is lost. (Cf. #521 ). We don't want to attach stack trace to the thrown condition unconditionally, for taking stack trace is expensive.However, we may use a dynamilally scoped flag that tells VM to save the stack trace (or, save the continuation chain and keep it) in the condition. Especially,
Scm_Loadis typically load programs but does not execute long-running code, so it is less critical for error performance.The text was updated successfully, but these errors were encountered: