diff options
author | Sawyer X <xsawyerx@cpan.org> | 2021-05-15 11:27:58 +0200 |
---|---|---|
committer | Sawyer X <xsawyerx@cpan.org> | 2021-05-15 11:27:58 +0200 |
commit | b77403b54fb61cf4a80612dc1c8a78c1c53bf2d7 (patch) | |
tree | e76ea06f4a999b7c408d73c8975524f7b32abc7e | |
parent | facfab624d9f5f04cb29c9106390941e2b8263de (diff) | |
download | perl-b77403b54fb61cf4a80612dc1c8a78c1c53bf2d7.tar.gz |
Perldelta: Podifying a bit more
-rw-r--r-- | pod/perldelta.pod | 52 |
1 files changed, 26 insertions, 26 deletions
diff --git a/pod/perldelta.pod b/pod/perldelta.pod index 2a9feb1342..f5e4671504 100644 --- a/pod/perldelta.pod +++ b/pod/perldelta.pod @@ -564,8 +564,8 @@ Additionally, the following selected changes have been made: =item * -L<perlapi>, L<perlguts>, L<perlxs>, and L<perlxstut> now prefer SvPVbyte -over SvPV. +L<perlapi>, L<perlguts>, L<perlxs>, and L<perlxstut> now prefer C<SvPVbyte> +over C<SvPV>. =item * @@ -649,7 +649,7 @@ also states that the result of the function is always in English. =item * L<msgsnd()|perlfunc/msgsnd> documented a length field included in the -packed C<MSG> parameter to msgsnd(), but there was no such field. +packed C<MSG> parameter to C<msgsnd()>, but there was no such field. C<MSG> contains only the type and the message content. =item * @@ -684,8 +684,8 @@ document how to get a PR as a local branch =item * -L<perlguts> now explains in greater detail the need to consult SvUTF8 -when calling SvPV (or variants). A new "How do I pass a Perl string to a C +L<perlguts> now explains in greater detail the need to consult C<SvUTF8> +when calling C<SvPV> (or variants). A new "How do I pass a Perl string to a C library?" section in the same document discusses when to use which style of macro to read an SV's string value. @@ -848,7 +848,7 @@ given and expected arguments. L<Lost precision when %s %f by 1|perldiag/"Lost precision when %s %f by 1"> -This warning was only issued for positive too-large values when +This warning was only issued for positive too-large values when incrementing, and only for negative ones when decrementing. It is now issued for both positive or negative too-large values. [L<GH #18333|https://github.com/Perl/perl5/issues/18333>] @@ -912,7 +912,7 @@ Filter inc paths out of libpth. =item * stadtx hash support has been removed stadtx support has been entirely removed. Previously, it could be requested -with PERL_HASH_FUNC_STADTX, and was default in 64-bit builds. It has been +with C<PERL_HASH_FUNC_STADTX>, and was default in 64-bit builds. It has been replaced with SipHash. SipHash has been more rigorously reviewed than stadtx. =item * Configure @@ -928,7 +928,7 @@ configuration option C<-Dusedefaultstrict>. These strict defaults do not apply when C<perl> is run via C<-e> or C<-E>. -This setting provides a diagnostic mechanism intended for development +This setting provides a diagnostic mechanism intended for development purposes only and is thus undefined by default. =item * @@ -968,7 +968,7 @@ made: =item * -Split Config-dependent tests in t/opbasic/arith.t to t/op/arith2.t +Split Config-dependent tests in F<t/opbasic/arith.t> to F<t/op/arith2.t> =item * @@ -1063,7 +1063,7 @@ versions < 3.22 now works again. This was broken in Perl 5.31.4. Building with mingw.org compilers (version 3.4.5 or later) using mingw runtime versions >= 3.21 now works (for compilers up to version 5.3.0). -Makefile.mk, and thus support for dmake, has been removed. It is still possible +F<Makefile.mk>, and thus support for dmake, has been removed. It is still possible to build Perl on Windows using nmake (Makefile) and GNU make (GNUmakefile). [L<GH #18511|https://github.com/Perl/perl5/pull/18511>] @@ -1090,8 +1090,8 @@ links C<nlink> field is now always populated. L<< C<${^WIN32_SLOPPY_STAT}> |perlvar/${^WIN32_SLOPPY_STAT} >> previously controlled whether the C<nlink> field was populated requiring a -separate Windows API call to fetch, since nlink and the other -information required for stat() is now retrieved in a single API call. +separate Windows API call to fetch, since C<nlink> and the other +information required for C<stat()> is now retrieved in a single API call. The C<-r> and C<-w> operators now return true for the C<STDIN>, C<STDOUT> and C<STDERR> handles. Unfortunately it still won't return @@ -1132,7 +1132,7 @@ them. They are included in C<LC_ALL>. Corrected handling of double and long double parameters for perl's implementation of formatted output for C<-Dusequadmath> builds. -This applies to PerlIO_printf(), croak(), warn(), sv_catpvf() and +This applies to C<PerlIO_printf()>, C<croak()>, C<warn()>, C<sv_catpvf()> and their variants. Previously in C<quadmath> builds, code like: @@ -1164,9 +1164,9 @@ And also, shifting INT_MIN bits used to yield the shiftee unchanged =item * -A set of cop_hints_exists_{pv,pvn,pvs,sv} functions was added, to support checking -for the existence of keys in the hints hash of a specific cop without needing to -create a mortal copy of said value. +A set of C<cop_hints_exists_{pv,pvn,pvs,sv}> functions was added, +to support checking for the existence of keys in the hints hash of a +specific cop without needing to create a mortal copy of said value. =item * @@ -1181,7 +1181,7 @@ calls. =item * -Make REFCOUNTED_HE_EXISTS available outside of core +Make C<REFCOUNTED_HE_EXISTS> available outside of core =item * @@ -1229,14 +1229,14 @@ clearly named way to find how many elements are in an array. =item * -Setting %ENV now properly handles upgraded strings in the key. Previously +Setting C<%ENV> now properly handles upgraded strings in the key. Previously Perl sent the SV's internal PV directly to the OS; now it will handle keys as it has handled values since 5.18: attempt to downgrade the string first; if that fails then warn and use the utf8 form. =item * -Fix a memory leak in regcomp.c (L<GH #18604|https://github.com/Perl/perl5/issues/18604>) +Fix a memory leak in F<regcomp.c> (L<GH #18604|https://github.com/Perl/perl5/issues/18604>) =item * pack/unpack format 'D' now works on all systems that could support it @@ -1256,7 +1256,7 @@ now skipped for such op trees. This also addresses =item * -%g formatting broken on Ubuntu-18.04, NVSIZE == 8 +C<%g> formatting broken on Ubuntu-18.04, C<NVSIZE == 8> [L<GH #18170|https://github.com/Perl/perl5/issues/18170>] Buggy libc implementations of the C<gcvt> and C<qgcvt> functions @@ -1266,12 +1266,12 @@ C<sprintf> will be used in place of C<gcvt> and C<qgcvt>. Tests added as part of this fix also revealed related problems in some Windows builds. The makefiles for MINGW builds on Windows have -thus been adjusted to use USE_MINGW_ANSI_STDIO by default, ensuring +thus been adjusted to use C<USE_MINGW_ANSI_STDIO> by default, ensuring that they also provide correct C<(s)printf> formatting of numbers. =item * -op.c: croak on "my $_" when "use utf8" is in effect +F<op.c>: croak on C<my $_> when C<use utf8> is in effect [L<GH #18449|https://github.com/Perl/perl5/issues/18449>] The lexical topic feature experiment was removed in Perl v5.24 and @@ -1280,7 +1280,7 @@ still possible to make this declaration if C<use utf8> was in effect. =item * -regexec.c: Fix assertion failure +F<regexec.c>: Fix assertion failure [L<GH #18451|https://github.com/Perl/perl5/issues/18451>] Fuzzing triggered an assertion failure in the regexp engine when too many @@ -1295,13 +1295,13 @@ operations. =item * -semctl(), msgctl(), and shmctl() now attempt to downgrade the C<ARG> +C<semctl()>, C<msgctl()>, and C<shmctl()> now attempt to downgrade the C<ARG> parameter if its value is being used as input to C<IPC_SET> or C<SETALL> calls. A failed downgrade will thrown an exception. =item * -In cases where semctl(), msgctl() or shmctl() would treat the C<ARG> +In cases where C<semctl()>, C<msgctl()> or C<shmctl()> would treat the C<ARG> parameter as a pointer, an undefined value no longer generates a warning. In most such calls the pointer isn't used anyway and this allows you to supply C<undef> for a value not used by the underlying @@ -1345,7 +1345,7 @@ timeout. L<GH #16529|https://github.com/Perl/perl5/issues/16529> =item * -C<MARK> and hence C<items> are now correctly initialized in BOOT XSUBs. +C<MARK> and hence C<items> are now correctly initialized in C<BOOT> XSUBs. =item * |