pike.git / CHANGES

version» Context lines:

pike.git/CHANGES:3:      o Bugfix in Parser.C. "!=" is now considered one token instead of two.   o depend_p() bugfix in las.   o Bugfixed loop over-optimization.   o Postgres 7.3+ correctly detected.      o "Backported funcion_object fixes from Pike 7.5"/"function_object()    now behaves as in Pike 7.2. The other cases have been moved to    function_program()."   o "OOB callback fixes for select(2) mode." - o "Fixed stack bug in push_int64 for large negative integers." +    o "Fix for running in UNLOCKED mode."   o "Backported identifier lookup fixes from Pike 7.5."   o "Fixed identifier sort bug in the old program decode method." - o "Retain the dumpable flag over the implicit sete?id calls for new -  threads on Linux." +    o "Backported object index lvalue fixes from Pike 7.5."   o "Bugfix in F_{INC,DEC}{,_NEQ}_LOOP." - o "Before invoking the trampoline magic, check properly that it's the -  `() lfun in the trampoline object that is to be called." +    o "MKREADSET: If elements e and e+1 have their places swapped, it    doesn't mean that there are overlapping ranges, just that the sort    function is unstable. It doesn't matter. Live with it."   o "Backported native argument handling fixes from 7.5." (java)   o "Backported fpsetmask() paranoia from Pike 7.5." (_math)   o "Fixed LARGE_FILE bug. Thanks to Dan Nelson." (files) - o "Backported System.dumpable() from Pike 7.5." (system) +     -  + o Added System.dumpable() which can be used on Linux to enable +  coredumping even for seteuid/setegid processes.   o Image.FreeType now selects the hopefully best character map when    loading a font, instead of just picking a random one.   o Fixed a bug that made searching for 8 bit strings in wider strings    fail.   o Fixed fencepost bug on overflow of the sprintf format info stack.   o Fixed bugs in float parsing code, preventing very large positive or    negative floats from being turned into 1.0 or -1.0.   o More robustness has been added to dlinit() and __alldiv(), which    fixes some start up crashes in Microsoft Windows when invoking Pike    with the wrong path.
pike.git/CHANGES:72:    last reading (i.e. the last gc run) to be weighted very heavily.    Thus a single "abnormal" gc run could cause the gc to schedule a    fairly short or long gc interval the next time which might make it    erratic.   o More solutions to new and fresh autoconf compatibility problems,    willfully introduced by the autoconf team.   o Threads are disabled by default on OpenBSD/alpha <= 3.x.   o Pike will now look for the master path in SOFTWARE\Pike\ in the    registry on Microsoft Windows.   o Fix icon transparency on Microsoft Windows. + o Fixed a bug that could cause wrong line numbers in backtraces when +  ia32 machine code is used. + o Fixed bug that could cause the stack to get out of synch when large +  64 bit negative integers were converted to bignum objects. + o Fixed bug in trampoline objects so that they can be indexed and +  printed as normal objects.      Changes since Pike 7.2.30:   ----------------------------------------------------------------------      New/improved functionality:   ---------------------------      o New Hilfe    The interactive pike environment you get when running pike without    giving a program name, has been thoroughly rewritten. Most notably