rfc:error-optimizations
Differences
This shows you the differences between two versions of the page.
Both sides previous revisionPrevious revisionNext revision | Previous revisionNext revisionBoth sides next revision | ||
rfc:error-optimizations [2010/08/23 22:14] – kalle | rfc:error-optimizations [2010/08/24 03:37] – kalle | ||
---|---|---|---|
Line 60: | Line 60: | ||
} else { | } else { | ||
/* no logging, increase the stack size */ | /* no logging, increase the stack size */ | ||
+ | |||
++EG(error_stack_size); | ++EG(error_stack_size); | ||
} | } | ||
Line 82: | Line 83: | ||
} | } | ||
</ | </ | ||
+ | |||
+ | ==== $php_errormsg ==== | ||
+ | |||
+ | This one is a tricky one, as we do not have any hooks for altering variables at reading, nor do we have JIT assignments. I think the best solution here is to simply remove $php_errormsg and require userland to use error_get_last() if they *REALLY* want the last errors, without depending on the track_errors ini options to be on. | ||
+ | |||
+ | If its not removed, then we cannot gain any optimization with track_errors = On at all, so if thats the case, it has to be taken into account when initializing the error_stack executor globals. | ||
==== Memory usage ==== | ==== Memory usage ==== | ||
Line 87: | Line 94: | ||
Obvious the memory usage here can grow quite rapidly, but people who would use this feature already takes great care of their code to not assume the opposite. | Obvious the memory usage here can grow quite rapidly, but people who would use this feature already takes great care of their code to not assume the opposite. | ||
- | ===== Notes ===== | + | ===== Additional notes ===== |
- | * This is a very early WIP, the idea may change, or render obsolete in the future | + | * While altering |
rfc/error-optimizations.txt · Last modified: 2017/09/22 13:28 by 127.0.0.1