From c03c6055a2f75f36311ed0e576399a375aed048a Mon Sep 17 00:00:00 2001 From: Michael Kerrisk Date: Mon, 11 Aug 2008 08:58:32 +0000 Subject: [PATCH] sync --- Changes | 158 +++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 98 insertions(+), 60 deletions(-) diff --git a/Changes b/Changes index 99f5d7d2c..7cde85e57 100644 --- a/Changes +++ b/Changes @@ -9,13 +9,22 @@ Contributors The following people contributed notes, ideas, or patches that have been incorporated in changes in this release: +Alain Portal +Andries Brouwer Fabian Kreutz Filippo Santovito Gerrit Renker +Heikki Orsila +Khalil GHORBAL Lee Schermerhorn Maxin John Reuben Thomas +Samuel Thibault Sam Varshavchik +Soh Kam Yung +Stephane Chazelas +Pavel Heimlich +Reuben Thomas Apologies if I missed anyone! @@ -107,6 +116,12 @@ mbind.2 Now that cpuset man page exists, we can refer to it. Remove stale comment regarding lack thereof. + Lee Schermerhorn + Add brief discussion of mode flags. + Lee Schermerhorn + ttempt to clarify discussion of MPOL_DEFAULT. + mtk + Fix URI reference for libnuma. mprotect.2 mtk / Maxin John @@ -128,6 +143,15 @@ set_mempolicy.2 Now that we have a cpuset man page, we can refer to cpusets directly in the man page text. + + Lee Schermerhorn + Another attempt to rationalize description of MPOL_DEFAULT. + + Since ~2.6.25, the system default memory policy is "local allocation". + MPOL_DEFAULT itself is a request to remove any non-default policy and + "fall back" to the surrounding context. Try to say that without delving + into implementation details. + Lee Schermerhorn Add discussion of mempolicy mode flags to set_mempolicy(2). This adds another reason for EINVAL. @@ -184,11 +208,6 @@ clog2.3 AVAILABILITY: These functions are still not in glibc as at version 2.8. -cosh.3 - mtk - BUGS: In glibc 2.3.4 and earlier, an FE_OVERFLOW exception is not - raised when an overflow occurs. - ctime.3 mtk POSIX.1-2008 marks asctime(), asctime_r(), ctime(), and ctime_r() @@ -203,23 +222,6 @@ ecvt.3 CONFORMING TO: POSIX.1-2008 removes the specifications of ecvt() and fcvt(). -fenv.3 - mtk / Fabian Kreuz - Provide more detail in the description of rounding modes. - Add text describing FLT_ROUNDS (foermerly in fma.3). - Add BUGS section pointing out the FLT_ROUNDS does not reflect - changes by fesetround(). - -fma.3 - mtk - Remove text about FLT_ROUNDS, replacing with a cross-reference - to fenv(3). - -fpclassify.3 - mtk - CONFORMING TO: Note that the standards provide a weaker guarantee - for the return value of isinf(). - ftime.3 mtk CONFORMING TO: POSIX.1-2008 removes the specification of ftime(). @@ -268,15 +270,6 @@ isalpha.3 mtk CONFORMING TO: POSIX.1-2008 marks isalpha() as obsolete. -log.3 - mtk - BUGS: In glibc 2.5, log(nan("")) produces a bogus - FE_INVALID exception. - -lround.3 - mtk - Add reference to fenv(3) for discussion of current rounding mode. - makecontext.3 mtk CONFORMING TO: POSIX.1-2008 removes the specifications of @@ -295,29 +288,10 @@ opendir.3 mtk CONFORMING TO: POSIX.1-2008 specifies fdopendir(). -pow.3 - mtk - BUGS: described buggy NaN return when x is negative and y is large. - See also: http://sources.redhat.com/bugzilla/show_bug.cgi?id=3866. - BUGS: Note the bogus FE_INVALID exception that occurred in glibc - 2.3.2 and earlier on overflow and underflow. - rand.3 mtk CONFORMING TO: POSIX.1-2008 marks rand_r() as obsolete. -remainder.3 - mtk - Add reference to fenv(3) for discussion of current rounding mode. - -round.3 - mtk - Add reference to fenv(3) for discussion of current rounding mode. - -scalb.3 - mtk - CONFORMING TO: POSIX.1-2008 removes the specification of scalb(). - siginterrupt.3 mtk CONFORMING TO: POSIX.1-2008 marks siginterrupt() as obsolete. @@ -331,11 +305,6 @@ tempnam.3 mtk CONFORMING TO: POSIX.1-2008 marks tempnam() as obsolete. -tgamma.3 - mtk - BUGS: In glibc 2.3.3, tgamma(+-0) produced a domain error - instead of a pole error. - tmpnam.3 mtk CONFORMING TO: POSIX.1-2008 marks tmpnam() as obsolete. @@ -356,11 +325,6 @@ usleep.3 mtk CONFORMING TO: POSIX.1-2008 removes the specification of usleep(). -y0.3 - mtk - In glibc 2.3.2 and earlier, these functions do not raise an - FE_INVALID exception for a domain error. - standards.7 mtk Updated details for POSIX.1-2008, and noted that if @@ -371,3 +335,77 @@ standards.7 udp.7 mtk SEE ALSO: add udplite(7). + + +Changes to individual pages (math functions) +-------------------------------------------- + +Various changes here following on from the big update to the +math pages in the previous release. Test results going back +glibc 2.3.2 (so far) allowed updates to various pages to note +changes in historical behavior for error reporting by math +functions. Thanks to the following people for providing me +with test results on various distributions and glibc versions: +Alain Portal, Andries Brouwer, Fabian Kreutz, Heikki Orsila, +Khalil GHORBAL, Pavel Heimlich, Reuben Thomas, Samuel Thibault, +Soh Kam Yung, and Stephane Chazelas + +cosh.3 + mtk + BUGS: In glibc 2.3.4 and earlier, an FE_OVERFLOW exception is not + raised when an overflow occurs. + +fenv.3 + mtk / Fabian Kreuz + Provide more detail in the description of rounding modes. + Add text describing FLT_ROUNDS (foermerly in fma.3). + Add BUGS section pointing out the FLT_ROUNDS does not reflect + changes by fesetround(). + +fma.3 + mtk + Remove text about FLT_ROUNDS, replacing with a cross-reference + to fenv(3). + +fpclassify.3 + mtk + CONFORMING TO: Note that the standards provide a weaker guarantee + for the return value of isinf(). + +log.3 + mtk + BUGS: In glibc 2.5, log(nan("")) produces a bogus + FE_INVALID exception. + +lround.3 + mtk + Add reference to fenv(3) for discussion of current rounding mode. + +pow.3 + mtk + BUGS: described buggy NaN return when x is negative and y is large. + See also: http://sources.redhat.com/bugzilla/show_bug.cgi?id=3866. + BUGS: Note the bogus FE_INVALID exception that occurred in glibc + 2.3.2 and earlier on overflow and underflow. + +remainder.3 + mtk + Add reference to fenv(3) for discussion of current rounding mode. + +round.3 + mtk + Add reference to fenv(3) for discussion of current rounding mode. + +scalb.3 + mtk + CONFORMING TO: POSIX.1-2008 removes the specification of scalb(). + +tgamma.3 + mtk + BUGS: In glibc 2.3.3, tgamma(+-0) produced a domain error + instead of a pole error. + +y0.3 + mtk + In glibc 2.3.2 and earlier, these functions do not raise an + FE_INVALID exception for a domain error.