bulk88

UCRTbase.dll toupper() is 133x slower wall time than perl/msvcrt.dll

22

Module: **Description** A certain profiling call stack caught my eye and the final report from my profiler said 8% of all cpu time of perl is spent inside. ```isupper()```/```toupper()``` from...

**Description** `win32_realloc()` was passed 0x30, `HeapSize()` on the ptr says the block size is already 0x30. inside `S_change_engine_size` ((RExC_state_t *)pRExC_state)->precomp is "(?x) \b strict \.pmc? \z" ((RExC_state_t *)pRExC_state)->size is 1...

- macros Perl_error_log/Perl_debug_log internally call exported function PerlIO_stderr(my_perl) which may or may not call Newxz() or calloc(). A naive person would think these macros are PL_something my_perl->Isomething vars or they...

5.41.13 thru EU::MM 7.72 breaks all i386 Win32 XS .dll during linking if they add any extra "-lsome_c_lib"es

16

**Module:** EU::MM in blead **Description** This bug blocks Perl 5.42.0. I don't have time to rewrite all this ATM. ``` Revision: 179356156963a4b43629d91b6c8abc3030e76739 Author: Chris 'BinGOs' Williams Date: 3/14/2025 7:39:15 AM...

- high usage in perl5xx.dll and because these are macros, high usage in both P5P/.git and CPAN XS .dll of a certain category. - pthreads variants of COND_*(), have error...

feature.pm add __END__, split POD, less IO/CPU parse time

1

-feature.pm is 29 KB big, don't IO read or parse 22 KB of POD feature.pm's POD in the middle is from its initial commit 69bcf1d3d727aba8df8f279ed93ee56eea0cd15f 12/22/2011 1:46:41 AM Create regen/feature.pl...

Win32: htonl/htons/ntohl/ntohs change slow winsock exports -> 1 CPU op/ins

5

This patch started when I saw this call stack while single stepping Storeable.pm/.xs. And to my horror, why on earth is Storeable.pm **trying to go on the WWW**. It has...

-type PVGV and GV* API will eventually SEGV through GvNAME and PP tie so bounds limit it to the 2 scalar string types, that are capable of MG and not...

extra thoughts https://github.com/Perl/perl5/blob/a73cdaecc7e625c976c1806ebdcb715321d892e3/sv.c#L5247 +(PTRSIZE-1) &0x7 trick https://github.com/Perl/perl5/blob/a73cdaecc7e625c976c1806ebdcb715321d892e3/sv.c#L5265 msize? IDK, but does Perl_safesysmalloc_size()/LIBC backend, in-place stretch the input string alloc block, to in-place realloc to "MAX", whatever MAX is for that...

Most OSes/libcs have an optimization that calloc() sometimes, or most of the time, do not call memset() in userland wasting CPU to zeroize brand new memory blocks/pages obtained fresh from...