summaryrefslogtreecommitdiff
path: root/src/sp_harden_rand.c (follow)
AgeCommit message (Collapse)Author
2021-09-23config is stack allocated now + some code improvements (see details)Ben Fuhrmannek
* for easier memory manegement, the entire sp_config struct was merged into snuffleupagus_globals and allocated on stack where possible * SNUFFLEUPAGUS_G() can be written as SPG(), which is faster to type and easier to read * execution_depth is re-initialized to 0 for each request * function calls with inline string and length parameters consistently use ZEND_STRL instead of sizeof()-1 * execution is actually hooked if recursion protection is enabled * some line breaks were removed to make the code more readable
2019-06-19fix snufflepagus_globals linking issues and one mac compatibility issueBen Fuhrmannek
2018-09-03Fix #241 segfault on 32-bit related to bad protype (#242)Remi Collet
This should close #241
2018-08-30Minor code cleanupxXx-caillou-xXx
2018-07-19Fix the issue with rand hardeningxXx-caillou-xXx
2018-02-05Massive simplification of functions hookingjvoisin
2018-01-04Eval blacklistjvoisin
Add support for eval filtering, only blacklist for now
2018-01-03Remove some dead codejvoisin
2017-11-29Code formatting pass on harden-randjvoisin
2017-11-27Initialize some possibly uninitialized variablesjvoisin
Thanks to coverity
2017-11-24Fix harden_rand (#72)jvoisin
This one was tricky. It was a great half-hour of joy, full of macros, ctags, gdb, radare2, tears, hardcoded `int3`, … to finally find that php calls `return` when it fails to parse some parameters for various reasons, even if everything goes fine. This must be a better way to do this, but this is good enough™ for now. This closes #66
2017-10-18.drop() is not a `nop` anymoreThibault "bui" Koechlin
`.drop()` is now baillout out, instead of nop'ing the call. This closes #13
2017-10-11Cheat a bit with the coveragejvoisin
2017-09-20Initial importSebastien Blot