On Thu, 27 Sep 2001, Stas Bekman wrote: > why the following pseudo-code screws everything. I think it messes up > the SP, but I cannot tell since I cannot run the debugger :( I guess I > could test it outside of mod_perl env.
perldoc perlapi =item dSP Declares a local copy of perl's stack pointer for the XSUB, ... so SP is not the same in foo() as it is in the calling function. > My point is that it's not easy to call the code that messes up with > items and SP from C functions which are wrapped in XS wrapper. you shouldn't be calling function that mess with items and SP. there are two forms of xsubs in modperl-2.0 that do not have a fix prototype. those with '...' in the .map file are passed items, MARK and SP. and those prefixed MPXS_ which are hooked directly into newXS(). any messing with items and SP should happen at the level of those functions. > So is it OK that I protory functions that use TABLE_GET_SET not to work > with SP and items but instead be prototyped with object, key=NULL, > val=NULL? ? --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
