I am used to profiling with xdebug, and then use wincachegrind (on wine) to look at the results (I prefer wincachegrind to kcachegrind).
The output from xdebug allows to really inspect the complete call tree. This can be quite hairy big data for long processes, but I find it super useful.
I don't really care about visual representation, I am happy to click through the list to explore the call tree.
Now everyone is telling me to use xhprof instead.
It seems that xhprof does not really distinguish between functions and calls, and does not record a full call tree. It only remembers how often function A calls function B, but not anything about the call trace, or which call instance of A calls which call instance of B.
Is this a correct observation?
Would it be possible to implement the full call tree capture in xhprof?
Thanks!
I am used to profiling with xdebug, and then use wincachegrind (on wine) to look at the results (I prefer wincachegrind to kcachegrind). The output from xdebug allows to really inspect the complete call tree. This can be quite hairy big data for long processes, but I find it super useful. I don't really care about visual representation, I am happy to click through the list to explore the call tree.
Now everyone is telling me to use xhprof instead.
It seems that xhprof does not really distinguish between functions and calls, and does not record a full call tree. It only remembers how often function A calls function B, but not anything about the call trace, or which call instance of A calls which call instance of B.
Is this a correct observation? Would it be possible to implement the full call tree capture in xhprof? Thanks!