summaryrefslogtreecommitdiff
path: root/util.c
Commit message (Collapse)AuthorAgeFilesLines
* remove kludgey duplicate background error avoidance (causedGurusamy Sarathy1999-10-081-1/+0
| | | | | | "leaks"; %@ wasn't even user-visible under -Dusethreads); only repeats of most recent error are now avoided p4raw-id: //depot/perl@4316
* integrate cfgperl contents into mainlineGurusamy Sarathy1999-10-061-15/+2
|\ | | | | p4raw-id: //depot/perl@4303
| * Untangle the IV_IS_QUAD jungle by introdudingJarkko Hietaniemi1999-10-021-15/+2
| | | | | | | | | | | | macros to be used when doing formatted printing: IVdf, UVuf, UVxf, UVof. Also introduce Uid_t_SIGN. p4raw-id: //depot/cfgperl@4288
* | make die/warn and other diagnostics go to wherever STDERR happensGurusamy Sarathy1999-10-061-55/+67
|/ | | | | | to point at; change places that meant Perl_debug_log rather than PerlIO_stderr() p4raw-id: //depot/perl@4302
* off-by-one in fbm_compile() (spotted by John BleyGurusamy Sarathy1999-09-301-20/+25
| | | | | <jbley@cs.cmu.edu>); whitespace adjustments p4raw-id: //depot/perl@4253
* queue errors due to strictures rather than printing them asGurusamy Sarathy1999-09-201-8/+49
| | | | | | | | | | | warnings; symbols that violate strictures do *not* end up in the symbol table anyway, making multiple evals of the same piece of code produce the same errors; errors indicate all locations of a global symbol rather than just the first one; these changes make compile-time failures within evals reliably visible via the return value or contents of $@, and trappable using __DIE__ hooks p4raw-id: //depot/perl@4197
* dos-djgpp update (from Laszlo Molnar <laszlo.molnar@eth.ericsson.se>)Gurusamy Sarathy1999-09-101-15/+3
| | | p4raw-id: //depot/perl@4121
* Fix UV_SIZEOF to UVSIZE; change the overflow testsJarkko Hietaniemi1999-09-061-3/+3
| | | | | so that they overflow also on 64-bit platforms. p4raw-id: //depot/cfgperl@4093
* DEBUG_m() adjusted to internalize dTHXGurusamy Sarathy1999-08-201-4/+2
| | | p4raw-id: //depot/perl@4013
* integrate cfgperl contents into mainlineGurusamy Sarathy1999-08-201-25/+25
|\ | | | | p4raw-id: //depot/perl@4008
| * Integrate with Nick.Jarkko Hietaniemi1999-08-161-1/+5
| |\ | | | | | | p4raw-id: //depot/cfgperl@3999
| * | Use Pid_t.Jarkko Hietaniemi1999-08-101-7/+7
| | | | | | | | | p4raw-id: //depot/cfgperl@3950
| * | lexwarn maintenance: new warning class unsafeJarkko Hietaniemi1999-08-091-18/+18
| | | | | | | | | | | | | | | | | | | | | | | | subclasses 'overflow' and 'portable' created, used by the recent integer overflow warnings. Class syntax subclass 'octal' renamed to 'digit', binary and hexadecimal parsing errors also 'digit' warnings. p4raw-id: //depot/cfgperl@3942
* | | support USE_THREADS+MULTIPLICITY; source compat tweaks forGurusamy Sarathy1999-08-201-25/+10
| |/ |/| | | | | | | USE_THREADS and MULTIPLICITY; minor pod adjustments p4raw-id: //depot/perl@4007
* | New lightweight Carp has a require. If Carp is used in a __DIE__ handlerNick Ing-Simmons1999-08-141-1/+5
|/ | | | | | | | | | | this causes a POPSTACK panic. The problem seems to be that although die_where() has unwound the tail of perl_vdie() top_env setjmp has been set to resume execution there. Avoiding setting CATCH_SET(TRUE) in call_sv() avoids this. So invent a new G_NOCATCH flag to disable messing with CATCH_SET() in call_sv, use it in perl_vdie(). Add test to op/eval.t which will fail (panic) if bug comes back. >>> I AM NOT CONVINCED THIS IS CORRECT LONG TERM FIX <<< p4raw-id: //depot/perl@3988
* Integrate with Sarathy.Jarkko Hietaniemi1999-08-021-0/+1
|\ | | | | p4raw-id: //depot/cfgperl@3889
| * fix coredump under usethreads+debugGurusamy Sarathy1999-08-021-0/+1
| | | | | | p4raw-id: //depot/perl@3876
* | Fix some of the {IV_IS_QUAD,UV_IS_QUAD} cases so thatJarkko Hietaniemi1999-08-021-1/+2
|/ | | | | | emacs code indentation doesn't get wrong ideas-- in other words, introduce a couple of "redundant" if:s. p4raw-id: //depot/cfgperl@3887
* Simplicate duplicated code.Jarkko Hietaniemi1999-08-011-23/+0
| | | p4raw-id: //depot/cfgperl@3865
* Integrate with Sarathy. perl.h and util.c required manual resolving.Jarkko Hietaniemi1999-08-011-4/+27
| | | p4raw-id: //depot/cfgperl@3864
* 64-bit work. Now 32-bit platforms get a 100% make testJarkko Hietaniemi1999-08-011-0/+12
| | | | | | | | | | | | | | | with -Duse64bits (using long long). Tested in Solaris 2.6 sparc RH Linux 6.0 x86 (and Digital IX 4.0D, to get a true 64-bit opinion). Now e.g. 'print unpack "q", pack "q", 12345678901' should work on such 32-bit platforms. Still a lot of printf()s behind -D which wrongly assume that %ld/%lx and (long) are a good combination. Introducing a slew of new macros intended to be used in printf() format strings: e. g. PERL_PRId64 is the string to be used when printing an IV, printf("%" PERL_PRId64 "\n", iv). The PRI... naming follows the C9X naming of <inttypes.h> macros. p4raw-id: //depot/cfgperl@3861
* Integer overflow iteration.Jarkko Hietaniemi1999-07-311-4/+16
| | | p4raw-id: //depot/cfgperl@3843
* Hack the "integer overflow" code some more.Jarkko Hietaniemi1999-07-311-23/+6
| | | p4raw-id: //depot/cfgperl@3840
* Repent and make overly large integerishJarkko Hietaniemi1999-07-291-52/+121
| | | | | | | constants non-fatal. They are now promoted to NVs, accompanied by an overflow warning that is by default on. p4raw-id: //depot/cfgperl@3832
* Minuscule cleanup of the integer overflow patch.Jarkko Hietaniemi1999-07-271-4/+3
| | | p4raw-id: //depot/cfgperl@3801
* Integer constants (0x, 0[0-7], 0b) now overflow fatally,Jarkko Hietaniemi1999-07-271-37/+86
| | | | | | | they used to be just optional lexical warnings. Also, with warnings turned on, constants > 2**32-1 trigger a non-portability warning. p4raw-id: //depot/cfgperl@3798
* don't display tid from main thread (or testsuite breaks)Gurusamy Sarathy1999-07-191-1/+2
| | | p4raw-id: //depot/perl@3702
* Integrate with Sarathy.Jarkko Hietaniemi1999-07-181-1/+5
|\ | | | | p4raw-id: //depot/cfgperl@3698
| * display thread id in diagnostics (suggested by Dan Sugalski)Gurusamy Sarathy1999-07-181-0/+3
| | | | | | p4raw-id: //depot/perl@3696
| * File descriptor leak in do_exec3Ilya Zakharevich1999-07-171-1/+2
| | | | | | | | | | Message-Id: <199907090827.EAA03321@monk.mps.ohio-state.edu> p4raw-id: //depot/perl@3678
* | Integrate with Sarathy.Jarkko Hietaniemi1999-07-131-34/+61
|\ \ | |/ | | p4raw-id: //depot/cfgperl@3670
| * fixups for sundry warnings about function pointersGurusamy Sarathy1999-07-121-6/+6
| | | | | | p4raw-id: //depot/perl@3669
| * yet more cleanups of the PERL_OBJECT, MULTIPLICITY and USE_THREADSGurusamy Sarathy1999-07-121-28/+55
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | builds; passing the implicit context is unified among the three flavors; PERL_IMPLICIT_CONTEXT is auto-enabled under all three flavors (see the top of perl.h) for testing; all varargs functions foo() have a va_list-taking variant vfoo() for generating the context-free versions; the PERL_OBJECT build should now be hyper-compatible with CPAN extensions (C++ is totally out of the picture) result has only been tested on Windows TODO: write docs on the THX rationale and idiomatic usage of the Perl API p4raw-id: //depot/perl@3667
* | Change t/pragma/warn oct()/hex() overflow tests to use %ConfigJarkko Hietaniemi1999-07-101-2/+2
|/ | | | | | | to adapt to the underlying platform (the binary, 0b1..., test was broken in 64-bit platforms). Also change "hex" in the warning messages to "hexadecimal" to match "binary" and "octal". p4raw-id: //depot/cfgperl@3662
* more PERL_OBJECT cleanups (changes still untested on Unix!)Gurusamy Sarathy1999-07-081-3/+3
| | | p4raw-id: //depot/perl@3660
* fixes for logical bugs in the lexwarn patch; other tweaks to avoidGurusamy Sarathy1999-07-081-16/+17
| | | | | type mismatch problems p4raw-id: //depot/perl@3658
* lexical warnings update (warning.t fails one testPaul Marquess1999-07-071-7/+12
| | | | | | | due to leaked scalar, investigation pending) Message-ID: <5104D4DBC598D211B5FE0000F8FE7EB29C6C8E@mbtlipnt02.btlabs.bt.co.uk> Subject: [PATCH 5.005_57] Lexical Warnings - mandatory warning are now default warnings p4raw-id: //depot/perl@3640
* applied slightly tweaked version of suggested patch forIlya Zakharevich1999-07-061-0/+3
| | | | | | | improved RE API Message-Id: <199906092214.SAA14126@monk.mps.ohio-state.edu> Subject: [PATCH 5.005_57] REx engine rehash p4raw-id: //depot/perl@3606
* slightly tweaked version of suggested patchDan Sugalski1999-07-061-13/+14
| | | | | | Message-Id: <3.0.6.32.19990608140938.030f12e0@ous.edu> Subject: [PATCH 5.005_57]Use NV instead of double in the core p4raw-id: //depot/perl@3602
* AIX threaded build, plus few more on the side.Jarkko Hietaniemi1999-07-051-3/+3
| | | p4raw-id: //depot/cfgperl@3592
* Integrate with mainperl.Jarkko Hietaniemi1999-07-051-2/+43
|\ | | | | p4raw-id: //depot/cfgperl@3585
| * cygwin32 updateFifer, Eric1999-07-051-0/+36
| | | | | | | | | | | | Message-Id: <71E287AB0D94D111BBD600600849EC8185EDD9@POST> Subject: [ID 19990602.003] perl5.005_03 (CORE) cygwin32 port p4raw-id: //depot/perl@3582
| * NeXT doesn't have FD_CLOEXEC (suggested by Hans Mulder)Gurusamy Sarathy1999-07-051-2/+7
| | | | | | p4raw-id: //depot/perl@3581
* | EPOC port to Psion5.Olaf Flebbe1999-06-271-2/+2
|/ | | | | | | | To: Jarkko Hietaniemi <jhi@iki.fi> Cc: perl5-porters@perl.org Subject: Re: Psion5 Message-ID: <Pine.GHP.4.02.9906272048130.17736-100000@io.science-computing.de> p4raw-id: //depot/cfgperl@3555
* somewhat untested PERL_OBJECT cleanups (C++isms mostlyGurusamy Sarathy1999-06-271-2/+8
| | | | | | | | | gone from the public API); PERL_OBJECT builds again on windows TODO: namespace-clean the typedefs in iperlsys.h and elsewhere; remove C++ remnants from public headers p4raw-id: //depot/perl@3553
* Tidy up #3542 and #3543.Jarkko Hietaniemi1999-06-181-1/+1
| | | p4raw-id: //depot/cfgperl@3544
* Fixed two long-standing locale bugs.Jarkko Hietaniemi1999-06-171-1/+47
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Both problems were related to numeric locale which controls the radix character aka the decimal separator. (1) printf (and sprintf) were resetting the numeric locale to C. (2) Using locale-numerically formatted floating point numbers (e.g. "1,23") together with -w caused warnings about "isn't numeric". The operations were working fine, though, because atof() was using the local locale. Both problems reported by Stefan Vogtner. Introduced a wrapper for atof() that attempts to convert the string both ways. This helps Perl to understand numbers like this "4.56" even when using a local locale makes atof() understand only numbers like this "7,89". Remaining related problems, both of which existed before this patch and continue to exist after this patch: (a) The behaviour of print() is _not_ as documented by perllocale. Instead of always using the C locale, print() does use the local locale, just like the *printf() do. This may be fixable now that switching to-and-fro between locales has been made more consistent, but fixing print() would change existing behaviour. perllocale is not changed by this patch. (b) If a number has been stringified (say, via "$number") under a local locale, the cached string value persists even under "no locale". This may or may not be a problem: operations work fine because the original number is still there, but that the string form keeps its locale-ish outlook may be somewhat confusing. p4raw-id: //depot/cfgperl@3542
* win32 build fixesGurusamy Sarathy1999-06-101-18/+18
| | | p4raw-id: //depot/perl@3525
* more complete support for implicit thread/interpreter pointer,Gurusamy Sarathy1999-06-091-48/+124
| | | | | | | | | | | | | | | | | | | | | enabled via -DPERL_IMPLICIT_CONTEXT (all changes are noops without that enabled): - USE_THREADS now enables PERL_IMPLICIT_CONTEXT, so dTHR is a noop; tests pass on Solaris; should be faster now! - MULTIPLICITY has been tested with and without PERL_IMPLICIT_CONTEXT on Solaris - improved function database now merged with embed.pl - everything except the varargs functions have foo(a,b,c) macros to provide compatibility - varargs functions default to compatibility variants that get the context pointer using dTHX - there should be almost no source compatibility issues as a result of all this - dl_foo.xs changes other than dl_dlopen.xs untested - still needs documentation, fixups for win32 etc Next step: migrate most non-mutex variables from perlvars.h to intrpvar.h p4raw-id: //depot/perl@3524
* initial stub implementation of implicit thread/thisGurusamy Sarathy1999-06-071-97/+97
| | | | | | | | | | | | pointer argument; builds/tests on Solaris, win32 hasn't been fixed up yet; proto.h, global.sym and static function decls are now generated from a common database in proto.pl; some inconsistently named perl_foo() things are now Perl_foo(), compatibility #defines provided; perl_foo() (lowercase 'p') reserved for functions that take an explicit context argument; next step: generate #define foo(a,b) Perl_foo(aTHX_ a,b) p4raw-id: //depot/perl@3522
View Lorry Controller Status