Re: would like to break sys_select's dependency on sys_socket

Paul Eggert wrote: > > - The type fd_set and the FD_ZERO etc. macros are defined in . > > Before your patch they were included via #include which > > did > > #include ; now they are missing. > ... > Thanks for that analysis. Windows is not my strong suit, but does the > following littl

Re: [PATCH 06/10] * modules/pselect: Add select, LIBSOCKET.

On 07/06/11 11:48, Stefano Lattarini wrote: > If you've noticed a blunder in your > ChangeLog *before* pushing, why haven't you amended the commit the > blunder occurred into, instead of adding the fix in the following > commit? Well, partly because I was lazy. But there's a bigger issue here. I

Re: uuencode: multi-bytes char in remote file name contains bytes >0x80

Hi all: If there are characters in *decode_pathname* that are not in the portable > filename character set the results are unspecified. I think we could do some work to solve the non-ASCII file name problem according to the above feature. the implementation is not a problem, the point is whet

Re: [PATCH 06/10] * modules/pselect: Add select, LIBSOCKET.

On Wednesday 06 July 2011, Paul Eggert wrote: > On 07/06/11 03:22, Stefano Lattarini wrote: > > > Is it acceptable to edit the ChangeLog in this way? I've always thought > > that editing past ChangeLog entries was acceptable only to fix typos or > > other blunders. > > Yes, that's what I did, or

typo fixed in new pthread_sigmask module

* modules/pthread_sigmask (configure.ac): gl_SIGNAL_MODULE_INDICATOR and not gl_SYS_SELECT_MODULE_INDICATOR, fixing a typo exposed in an application that does not use the sys_select module. diff --git a/modules/pthread_sigmask b/modules/pthread_sigmask index f1126b6..d117b2d 100644 --- a/modules/pt

Re: uuencode: multi-bytes char in remote file name contains bytes >0x80

> Date: Tue, 05 Jul 2011 14:12:01 -0700 > From: Bruce Korb > CC: 張叁 , > Bruno Haible , > Eli Zaretskii , bug-gnulib@gnu.org, bug-gnu-ut...@gnu.org, > "\"Eric\" toe lin" > > Meanwhile, I'm still waiting for a good answer to, "Why do it at all?" Because not doing that would severely limit uu

Re: [PATCH] poll: do not return 0 on timeout=-1

On 07/06/2011 04:31 AM, Paolo Bonzini wrote: > On 07/06/2011 12:10 PM, Erik Faye-Lund wrote: >> * lib/poll.c: Loop with yield if no events occured >> --- >> OK, this should be a slightly more complete version. Added >> SwitchToThread and mailing as a proper git-patch (including >> a ChangeLog-entry

Re: ‘iconv’ fails to build on NetBSD 5.1

Hi Ludo, > The ‘iconv’ modules fails to build on NetBSD 5.1 (see gcc70 of the GCC > Compile Farm): Can you show the gnulib-tool invocations that you used when builting the tarball of your package? In particular, which of the modules 'iconv', 'iconv-h', 'iconv_open' are used, and which --avoid opt

[PATCH] pthread_sigmask: always replace when not using pthread

On mingw, when using win32 threading, the pthread_sigmask replacement function was not being installed. * m4/pthread_sigmask.m4 (gl_PTHREAD_SIGMASK): Force sigprocmask replacement when using some threading other than pthread. Fix logic bug. Signed-off-by: Eric Blake --- ChangeLog |

Re: [PATCH 06/10] * modules/pselect: Add select, LIBSOCKET.

On 07/06/11 03:22, Stefano Lattarini wrote: > Is it acceptable to edit the ChangeLog in this way? I've always thought > that editing past ChangeLog entries was acceptable only to fix typos or > other blunders. Yes, that's what I did, or at least what I thought I did; your other message was about

Re: uuencode: multi-bytes char in remote file name contains bytes >0x80

> > If there are characters in *decode_pathname* that are not in the portable > filename character set the results are _unspecified_. > 2011/7/6 Bruce Korb > On 07/05/11 13:13, Eric Blake wrote: > >> On 07/05/2011 12:12 PM, Bruce Korb wrote: >> >>> On 07/05/11 10:13, Eric Blake wrote: >>> >>>

‘iconv’ fails to build on NetBSD 5.1

Hi, The ‘iconv’ modules fails to build on NetBSD 5.1 (see gcc70 of the GCC Compile Farm): --8<---cut here---start->8--- depbase=`echo iconv.lo | sed 's|[^/]*$|.deps/&|;s|\.lo$||'`; /bin/ksh ../libtool --tag=CC--mode=compile gcc -DHAVE_CONFIG_H -I. -I../.

Re: [PATCH] poll: do not return 0 on timeout=-1

On 07/06/2011 12:10 PM, Erik Faye-Lund wrote: * lib/poll.c: Loop with yield if no events occured --- OK, this should be a slightly more complete version. Added SwitchToThread and mailing as a proper git-patch (including a ChangeLog-entry) that can be applied with git-am. Looks good to me. Paol

Re: [PATCH 06/10] * modules/pselect: Add select, LIBSOCKET.

Hello Paul. Again another nit, please bear with me ... On Wednesday 06 July 2011, Paul Eggert wrote: > --- > ChangeLog |6 -- > modules/pselect |2 ++ > 2 files changed, 6 insertions(+), 2 deletions(-) > > diff --git a/ChangeLog b/ChangeLog > index e815b29..ba7d0e4 100644 > --

Re: [PATCH 05/10] * lib/pselect.c (pselect): Use plain name, without "rpl_".

Hello Paul, and sorry if I chime in only to criticize... On Wednesday 06 July 2011, Paul Eggert wrote: > Don't #undef, since we don't need any underlying pselect. > * modules/pselect (Depends-on): Use our pselect.o if !HAVE_PSELECT. > Both changes suggested by Bruno Haible. > This commit message

[PATCH] poll: do not return 0 on timeout=-1

* lib/poll.c: Loop with yield if no events occured --- OK, this should be a slightly more complete version. Added SwitchToThread and mailing as a proper git-patch (including a ChangeLog-entry) that can be applied with git-am. ChangeLog |5 + lib/poll.c |7 +++ 2 files changed, 12

Re: [PATCH]: pselect: new module

On 06/07/11 08:52, Paul Eggert wrote: >> No serial number here? > > I still don't get why hand-maintained serial numbers are useful > in m4/*.m4 files (and nowhere else?!). Maybe they were useful > long ago, but their cost outweighs their benefit now, as far as > I can see. I got this answer whe

Re: lib/poll.c bug?

On 07/06/2011 10:23 AM, Erik Faye-Lund wrote: > Without proper kernel-level support for poll, we probably can't have a > poll on Windows that suspend the thread until an event happens; we'll > have to accept a bit of polling. But I believe that looking and > quacking like POSIX is better than

Re: lib/poll.c bug?

On Thu, Jun 30, 2011 at 8:40 PM, Erik Faye-Lund wrote: > On Thu, Jun 30, 2011 at 7:51 PM, Eric Blake wrote: >> On 06/29/2011 11:31 AM, Erik Faye-Lund wrote: >>> POSIX says the following about poll "A value of 0 indicates that the >>> call timed out and no file descriptors have been selected.". My

Re: would like to break sys_select's dependency on sys_socket

On 07/03/11 16:31, Bruno Haible wrote: > On native Windows, > - does not exist natively; gnulib provides it. > - The type fd_set and the FD_ZERO etc. macros are defined in . > Before your patch they were included via #include which did > #include ; now they are missing. > - The gnul

Re: [PATCH 3/5] dtotimespec: new module

Paul Eggert wrote: > * lib/timespec.h (dtotimespec): New decl. > * lib/dtotimespec.c, modules/dtotimespec: New files. > --- > ChangeLog |4 +++ > lib/dtotimespec.c | 69 > +++ > lib/timespec.h |2 + > modules/dtotimespec |

[PATCH 10/10] sys_select: define sigset_t more portably

* lib/sys_select.in.h: Always include , since we now need sigset_t and mingw defines it there. Include before split inclusion guard, to avoid mishaps on Solaris, whose eventually includes us. * m4/signal_h.m4 (gl_SIGNAL_H): Require gl_CHECK_TYPE_SIGSET_T. (gl_CHECK_TYPE_SIGSET_T): New macro, most

[PATCH 04/10] pselect: document better

* MODULES.html.sh (Support for systems lacking POSIX:2008): Add it. * doc/posix-functions/pselect.texi (pselect): Document new module. --- ChangeLog|4 MODULES.html.sh |1 + doc/posix-functions/pselect.texi |9 +++-- 3 files changed, 12

[PATCH 08/10] pselect: Use pthread_sigmask, not sigprocmask.

* lib/pselect.c (pselect): Use pthread_sigmask, as it supports multithreaded apps better than sigprocmask does. * modules/pselect (Depends-on): Depend on pthread_sigmask, not sigprocmask directly. --- ChangeLog |8 lib/pselect.c |4 ++-- modules/pselect |1 + 3 files c

[PATCH 01/10] * tests/test-sys_select.c (sigset_t): Test for it, too

Suggested by Bruno Haible. --- ChangeLog |5 + tests/test-sys_select.c |3 +++ 2 files changed, 8 insertions(+), 0 deletions(-) diff --git a/ChangeLog b/ChangeLog index 71c1560..aca6253 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,8 @@ +2011-07-05 Paul Eggert + +

[PATCH 09/10] * m4/sys_select_h.m4 (gl_HEADER_SYS_SELECT): Poison pselect.

Suggested by Bruno Haible. --- ChangeLog |3 +++ m4/sys_select_h.m4 |4 ++-- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index bea73df..7717ad0 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,5 +1,8 @@ 2011-07-06 Paul Eggert + * m4

[PATCH 07/10] modules/pselect: Add Bruno and Jim to maintainers.

--- modules/pselect |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/modules/pselect b/modules/pselect index 0e454d5..0ed5d8b 100644 --- a/modules/pselect +++ b/modules/pselect @@ -28,4 +28,4 @@ License: LGPLv2+ Maintainer: -Paul Eggert +Paul Eggert, Bruno Haible, Jim

[PATCH 06/10] * modules/pselect: Add select, LIBSOCKET.

--- ChangeLog |6 -- modules/pselect |2 ++ 2 files changed, 6 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index e815b29..ba7d0e4 100644 --- a/ChangeLog +++ b/ChangeLog @@ -2,8 +2,10 @@ * lib/pselect.c (pselect): Use plain name, without "rpl_".

[PATCH 05/10] * lib/pselect.c (pselect): Use plain name, without "rpl_".

Don't #undef, since we don't need any underlying pselect. * modules/pselect (Depends-on): Use our pselect.o if !HAVE_PSELECT. Both changes suggested by Bruno Haible. --- ChangeLog |5 + lib/pselect.c | 10 -- modules/pselect |2 +- 3 files changed, 10 insertions(+),

[PATCH 03/10] pthread_sigmask: new module

* MODULES.html.sh (Support for systems lacking POSIX:2008): Add it. * doc/posix-functions/pthread_sigmask.texi: Document new module. * lib/signal.in.h (pthread_sigmask): Arrange for replacement. This is done only as a macro; I don't know how well that'll work for C++. Move include before the incl

[PATCH 02/10] test-pselect: new module

* modules/pselect-tests, tests/test-pselect.c: New files. * tests/test-select.c, tests/test-sys_select-c++.cc: If TEST_PSELECT is defined, test pselect instead of testing select. --- ChangeLog|5 + modules/pselect-tests| 12 tests/test-pselect.c

Re: [PATCH]: pselect: new module

Thanks for your careful review, Bruno. I just now pushed some patches that I hope address most of your points, and will send out ten patch emails accordingly. The last one (patch 10/10) is the trickiest. I followed most of your advice, with the following quibbles. > If the emulation uses sigpro