Ralf Wildenhues wrote:
a couple of nits, if I may:
The attached patch addresses all of Ralf's nits, fixes a few minor bugs,
and contains a few additional simplifications of the new functions. Any
objections?
Regards,
Derek
--
Derek R. Price
Solutions Architect
Ximbiot, LLC http://ximbiot.com
Ralf Wildenhues wrote:
* Derek Price wrote on Thu, Sep 25, 2008 at 05:10:31PM CEST:
--- gnulib-tool revision 1.307
+++ gnulib-tool working copy
+func_push ()
+{
+ var=$1
+ shift
+ for e; do
The more portable way really is
for e
do
not
for e; do
Okay, I've made
.
2008-09-25 Derek Price [EMAIL PROTECTED]
Bruno Haible [EMAIL PROTECTED]
* gnulib-tool (func_import): Report all license incompatibilities, not
just the first one.
Thanks!
Derek
--
Derek R. Price
Solutions Architect
Ximbiot, LLC http://ximbiot.com
Get CVS
Hey all,
Anyone mind if I commit the attached patch? It makes `gnulib-tool
--update --lgpl' list all incompatible modules instead of breaking after
the first one, like so:
$ ../gnulib/gnulib-tool --update --lgpl
gnulib-tool: *** incompatible license on modules:
gnulib-tool: ***
hosts with lines longer
than 2 GiB.
The problem is that the new SSIZE_MAX macro depends on SIZE_MAX, which
may not be defined. I've installed the attached patch, ripped verbatim
from serveral other modules.
2005-10-05 Derek Price [EMAIL PROTECTED]
* getdelim.c (SIZE_MAX): New macro
Simon Josefsson wrote:
Derek Price [EMAIL PROTECTED] writes:
may not be defined. I've installed the attached patch, ripped verbatim
from serveral other modules.
Shouldn't we use the size_max module instead?
I'm not sure. Some 18 modules currently don't, defining SIZE_MAX
Bruno Haible wrote:
The Woe32 socket API has three main differences w.r.t. the POSIX socket API:
- The type of a socket is 'SOCKET', not 'int'. A SOCKET cannot be used in
places where a file descriptor is used. This means in particular that
close(), select(), poll() don't work on
Ingolf Steinbach wrote:
I have done a select() emulation on Win32 once. It was horrible.
I don't suppose the implementation was GPL'd or LGPL'd or could be?
And maybe the situation is easier for CVS when the communication
can be reduced to pipes and sockets (e.g. no serial support).
Well,
Yoann Vandoorselaere wrote:
Bruno, is there any update concerning this issue ?
GnuLib CVS has not being fixed, and this problem break the compile on a
number of systems.
I sent in a patch for this previously:
http://lists.gnu.org/archive/html/bug-gnulib/2005-09/msg00055.html.
This patch
Excuse me, I misspoke - the patch has been working on half a dozen
systems for some weeks in nightly testing. I just finally got the other
dozen or so systems up only a few days ago, but it's been compiling
testing ok there since they came up.
Regards,
Derek
Derek Price wrote:
Yoann
Yoann Vandoorselaere wrote:
The mail you're pointing out is the patch I sent :-)
My question is rather whether it will be applied to the GnuLib CVS
version.
Er, that's what I meant. :) I did apply and test it on several systems
here, however, and it is working. :)
Derek
--
Derek R.
mbuiter.h makes the unportable assumption that mbsinit() is a working
function. Since it looks at first glance like mbiter.h has the same
problem, I wasn't sure whether adding AC_CHECK_FUNCS_ONCE(mbsinit) to
m4/mbiter.m4 and any other m4 macros for modules which use mbsinit(),
plus the following
netdb.h was already #included via getaddrinfo.h:
2005-09-20 Derek Price [EMAIL PROTECTED]
* getaddrinfo.c: Don't include netdb.h included from
getaddrinfo.h.
Regards,
Derek
--
Derek R. Price
CVS Solutions Architect
Ximbiot http://ximbiot.com
v: +1 717.579.6168
f: +1 717.234.3125
At least on Windows, where I guess strcasecmp is not declared in
string.h, compiling regex.c elicits the following warning:
lib\regcomp.c(852) : warning C4013: 'strcasecmp' undefined; assuming extern
returning int
I assume that, since strcase.h provides a prototype for strcasecmp,
other
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
http://lists.gnu.org/archive/html/bug-gnulib/2005-09/msg00055.html.
Bruno doesn't like AC_LIBSOURCES so I doubt whether he'll accept that
part of the patch. Surely you need only the AC_CHECK_HEADERS_ONCE and
the AC_LIBOBJ part
2005-09-19 Derek Price [EMAIL PROTECTED]
* srclist.txt: glibc's glob.h is now in lib/glob-libc.h.
Derek
--
Derek R. Price
CVS Solutions Architect
Ximbiot http://ximbiot.com
v: +1 717.579.6168
f: +1 717.234.3125
mailto:[EMAIL PROTECTED]
2005-09-16 Paul Eggert [EMAIL PROTECTED
Bruno Haible wrote:
Derek Price wrote:
The --tests-base option is no longer defaulting to tests as
`gnulib-tool --help' specifies, however:
gnulib-tool: *** missing --tests-base option
gnulib-tool: *** Stop.
Fixed. Thanks for reminding me.
Looks good. Thanks again!
Regards
Yoann Vandoorselaere wrote:
if (hints (hints-ai_flags ~AI_CANONNAME))
/* FIXME: Support more flags. */
return EAI_BADFLAGS;
If you say which flags your application needs to work, maybe Simon can
do something about it?
For my own use, I simply remove this check. This is
/strstr.c:46: parse error before `.'
...and so on. It looks like a pretty straightforward case of the strstr
definintion in string.h overwriting the one required by the strstr
module. The attached untested patch should fix this:
2005-09-15 Derek Price [EMAIL PROTECTED]
* lib/strstr.h: #include
like, * Incomplete - see
lib/getaddrinfo.c. would be better. The extra notes my patch adds to
modules/getaddrinfo could then be added to lib/getaddrinfo.c, then, but
I think they would be superfluous.
2005-09-15 Derek Price [EMAIL PROTECTED]
* modules/getaddrinfo: Document shortcomings
Bruno Haible wrote:
Thanks. I applied the appended patch, very similar to yours. (But move
the #include outside the extern C { ... }.)
Any reason why you left this inside?
+ #undef strstr
+ #define strstr rpl_strstr
Just because it didn't matter?
Thanks,
Derek
--
Derek R. Price
CVS
I'm seeing warning: `pure' attribute directive ignored warnings from
regex_internal.h regex_internal.c compiling with GCC 2.95.4 on Alpha
Linux 2.2.20. Similarly, I see warning: `always_inline' attribute
directive ignored warnings compiling regcomp.c regexec.c.
Is this something worth working
in regex_internal.h,
where I stumbled across a convenient #if __GNUC__ = 3 switch already
being used to protect usage of inline, and used that.
2005-09-15 Derek Price [EMAIL PROTECTED]
* regex_internal.h: Blank `pure' for GNUC 3.
* regex_internal.c: Ditto, using
Paul Eggert wrote:
I'd fix it if I were you. You can use the __GNUC_PREREQ macro of
md5.h.
I know I've already installed this fix but, for future reference, is
there a simple reference I could use to find what version of GCC an
attribute first appeared in? Something easier than fumbling
Jim Meyering wrote:
Derek Price [EMAIL PROTECTED] wrote:
I've installed the attached patch. It is almost identical to my
previous one, with a few extra portability and typo fixes.
2005-09-12 Derek Price [EMAIL PROTECTED]
* modules/canon-host: Add canon-host.h. Depend on getaddrinfo
Peter O'Gorman wrote:
Peter O'Gorman wrote:
Looks like the cvs folks need to update their gnulib.
Of course, they have done this, and I feel silly.
Yes, they have. :) That fix should be released with 1.12.13, which
shouldn't be very far away.
Regards,
Derek
--
Derek R. Price
CVS
Jim, all,
Is there a GNULIB standard for this yet? Paul Eggert just went through
my glob_.h and tweaked the cpp spacing in the other direction. I
assumed at the time this meant that double-include protection should be
ignored for the purposes of indenting compiler directives in headers,
but Jim
Jim Meyering wrote:
Personally, I've found it useful enough to have consistently cpp-indented
I like it too, but I was willing to go with the flow on GNULIB. :)
sources that I wrote cppi, and to use it in a commit-hook for the coreutils.
I don't know when it became an option, but
Paul Eggert wrote:
OK, but in that case shouldn't the AC_REQUIRE([AC_GNU_SOURCE]) be in
gl_GLOB rather than gl_PREREQ_GLOB?
I don't think so. gl_GLOB tests for the _GNU_GLOB_INTERFACE_VERSION
macro from gnu-versions.h a known bug in GNU glob's POSIX support.
Neither requires the GNU
Peter O'Gorman wrote:
As you can see from this little snippet of gcc configure, you can't
mmap /dev/zero on darwin.
AC_CACHE_CHECK([whether mmap from /dev/zero works],
gcc_cv_func_mmap_dev_zero,
[# Add a system to this blacklist if it has mmap() but /dev/zero
# does not exist, or
Peter O'Gorman wrote:
The test would have to be a run test, which would mean having a
cross-compile alternative switching on the system name (this is why
gcc switches on name).
All,
Generally, I choose to be pessimistic about test failures when
cross-compiling. Anyone have an opinion about
Okay, I've committed the glob-min-glibc-h-changes2.diff patch. The
glob.h-glibc-to-gnulib2.diff should be the new minimal patch for submission to
glibc.
2005-09-12 Derek Price [EMAIL PROTECTED]
Paul Eggert [EMAIL PROTECTED]
* modules/glob (Files): Add glob-libc.h
Is there any reason I can't just assume gl_GETADDRINFO ran and config.h
was included before getaddrinfo.h? The following test is always coming
up false on platforms without getaddrinfo (as of AC 2.59, at least,
AC_CHECK_FUNCS via AC_REPLACE_FUNCS leaves HAVE_GETADDRINFO undefined
when it is not
Re: http://sourceware.org/bugzilla/show_bug.cgi?id=1060, glibc
objected to the extent of our changes to an installed header (glob.h) to
bring the file into sync with GNULIB. (They did accept the glob.c
changes, though they have yet to apply them.)
It is true, as Roland says, that we could put
Paul Eggert wrote:
shortening the above test to: # if !HAVE_GETADDRINFO, and I'd
rather just simplify the header
Yes, that sounds right.
Okay, thanks, Paul. I've committed this to CVS CVS for more testing
before I finish importing it into GNULIB.
Cheers,
Derek
--
Derek R. Price
It's late, I'm tired. Patches actually attached now.
2005-09-08 Derek Price [EMAIL PROTECTED]
* m4/glob.m4 (gl_GLOB_SUBSTITUTE): AC_LIBSOURCE C files.
* lib/glob_.h: Move most code forked from glibc here, then include...
* lib/glob-glibc.h: ...this new file, which is the original
Bruno Haible wrote:
files. I'll change this to use join and sort instead of a double loop.
Did you do this? I'm still getting pretty slow operation with CVS:
$ time maint-aux/gnulib-update
real14m42.968s
user3m45.188s
sys 10m41.704s
$
Just a few weeks back, I think it was
Paul Eggert wrote:
By the way, did GCC really say In function รข? That is, an a with
a circumflex over it? If so, should this be a GCC bug that we should
be worrying about?
Oh, no, sorry. I'm accessing a Linux host via a Cygwin xterm. Some
sort of term incompatibility causes almost
Derek Price wrote:
Oh, no, sorry. I'm accessing a Linux host via a Cygwin xterm. Some
sort of term incompatibility causes almost everything quoted in all
program output (I assume with a backticks-singlequote combination,
though that's from memory and I haven't reexamined it since the problem
Yoann Vandoorselaere wrote:
On Thu, 2005-09-01 at 12:28 +0200, Yoann Vandoorselaere wrote:
Hi,
An OpenBSD Prelude user reported that GnuLib will fail to compile on
OpenBSD 3.7 due to the new dependencies of modules like strcase on
wctype.h and wchar.h headers.
I've got a second
on gethostbyname also implied that the
hostname would be resolved and a test program on 1 linux
NetBSD 1.6.1 showed that it was.
If we stumble across any systems where gethostbyname returns an IP
address when it is given a real hostname, this can be fixed then.
2005-09-05 Derek Price [EMAIL PROTECTED
Is the following line in lib/getpass.c correct?
#if TCSETATTR
A similar reference earlier uses `HAVE_TCGETATTR' for the get
counterpart to set. Also, there are no tests that I see which set
either macro and attempting to compile on HP-UX 11.00 yields the
following error message:
Hi all,
Would anyone object to a patch that caused canon-host to output warnings
via error (0, ...) when one of the functions it calls fails?
getaddrinfo returns errors differently than gethostbyname and
gethostbyaddr, making outputting a useful error message upon seeing a
simple NULL return
Jim Meyering wrote:
Personally, I don't mind if you add the code to emit warnings now,
as long as you agree to adjust the API later, (e.g., to add
a new enum parameter describing the error) if anyone complains.
Of course, it'd be better to keep it library-safe.
Hrm. A new enum parameter
Jim Meyering wrote:
I like your idea of keeping them separate.
How about passing either NULL or the address of a struct
containing a member for each error indicator?
Okay. Will do. Should I ignore single-threaded apps entirely, keep the
error data in a global to simplify for
2005-09-04 Derek Price [EMAIL PROTECTED]
* gnulib-tool: Fix reversed $symbolic logic.
Regards,
Derek
--
Derek R. Price
CVS Solutions Architect
Ximbiot http://ximbiot.com
v: +1 717.579.6168
f: +1 717.234.3125
mailto:[EMAIL PROTECTED]
Index: gnulib-tool
Jim Meyering wrote:
I don't thinks it's worthwhile to pander to single-threaded
applications for something like this.
In this case, pandering isn't very complicated, so I think I'll do it.
Besides, this way, if anyone else is using the canon-host module, they
won't need to update callers
Derek Price wrote:
Hrm. Why isn't canon-host dependant on getaddrinfo? It would
The alternative is that the ch_strerror_r function I've been working on
would need to handle ENOMEM too, which introduces a dependency on
strerror_r... I almost have the previously discussed canon-host code
done
Derek Price wrote:
This was about what I was thinking, though I was going to combine
canon_host canon_host_r in the canon-host module and let the caller
decide what to call. How about this API:
Hrm. Why isn't canon-host dependant on getaddrinfo? It would simplify
the canon-host code so
Hrm. The POSIX getaddrinfo specification is pretty clear about *not*
resolving a name for ai_canonname if name is in IP dot notation, yet
sometimes the canon-host implemetation does a reverse-lookup to get a
canonical name (for some odd condition where gethostbyname fills in the
h_name field of
Committed patch attached.
2005-07-03 Derek Price [EMAIL PROTECTED]
* getlogin_r (gl_GETLOGIN_R): Fix cut paste error.
From Larry Jones [EMAIL PROTECTED].
Derek
--
Derek R. Price
CVS Solutions Architect
Ximbiot http://ximbiot.com
v: +1 717.579.6168
f: +1 717.234.3125
mailto
Hi bug-gnulib,
I'd like to reopen a discussion on this list from last September:
http://lists.gnu.org/archive/html/bug-gnulib/2004-09/msg00010.html.
At the end of that discussion I didn't create the blocking-io module due
mostly to Paul Eggert's objections (and a lack of time on my part,
Simon Josefsson wrote:
+ if (!req_version || strverscmp (req_version, VERSION) 0)
+return VERSION;
Why did you choose to call the version string VERSION rather than
using the PACKAGE_VERSION string automatically defined by AC_INIT?
Regards,
Derek
Simon Josefsson wrote:
No particular reason. VERSION is defined automatically by AC_INIT
too, is it not? Is there a difference between PACKAGE_VERSION and
VERSION? I'd be happy to change, it doesn't matter for me.
The CVS config.h is only showing `PACKAGE_VERSION'.
Regards,
Derek
Paul Eggert wrote:
Please install it in gnulib; it looks good to me. Thanks for
following up on it.
Done:
2005-06-24 Derek Price [EMAIL PROTECTED]
and Bruno Haible [EMAIL PROTECTED]
Remove stat module update lstat.
* MODULES.html.sh (stat): Remove
Stepan Kasal wrote:
After that fix, we could also remove the AS_LITERAL_IF with m4_fatal.
I'm not sure what you were referring to here so I've attached a patch to
document AS_LITERAL_IF as you requested a few days ago. I will commit
it in a few days if there are no objections. It looks
Paul Eggert wrote:
I'm not sure about the documentation change. It's not yet clear to me
that we want to document AS_LITERAL_IF (the above example being one of
the gotchas). If we do want to document it we probably need to be
more systematic about it and its friends. So I left that alone for
Derek Price wrote:
Would anyone mind if I committed a patch with @msindex entries for the
37 or so undocumented m4_* macros in the `Redefined M4 Macros' node of
doc/autoconf.texi? I'm not bothering to include the patch in this email
because it is too simple but I don't want to waste the 15
Stepan Kasal wrote:
- We need to document also AS_LITERAL_IF and m4_fatal
(And you could also document m4_warning, when you are at it.)
I'll see about it after I get comments on the first round back.
- we have to document also the fact that AS_TR_SH AS_TR_CPP expand
to literal variable
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
What does this mean for Bruno's recent patch for stat lstat which
removed the SunOS 4.1.4 support in addition to some other fixes?
His patch made sense to me, but as far as I know nobody has taken the
time to integrate
Bruno Haible wrote:
I blindly searched for m4_pushdef in the manual and didn't find
it. (May I suggest that these renamed m4_* macros be added to the manual's
index, one by one?)
Autoconf folks:
Would anyone mind if I committed a patch with @msindex entries for the
37 or so undocumented
Bruno Haible wrote:
But AS_TR_* is actually undocumented, right?
Yes, AS_TR_SH AS_TR_CPP appear to be undocumented. I've submitted a
patch to autoconf-patches to remedy this and will commit it within a few
days unless there are objections there.
Cheers,
Derek
Paul Eggert wrote:
I wouldn't object to minor and obvious tweaks that are checked out by
a real user who actually needs findutils to run on SunOS 4.1.4. My
assumption in dropping support for SunOS 4.1.4 was partly predicated
on nobody really needing it any more (and computer museums don't count
Karl Berry wrote:
Is there a mailing list for CVS commit messages to GNULIB?
Yes. Do you want to be on it?
Yes, please.
Regards,
Derek
___
bug-gnulib mailing list
bug-gnulib@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-gnulib
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
1. Corrects an incorrect check for a successful return from
getlogin_r to assume only 0 means success, per the POSIX2 spec:
http://www.opengroup.org/onlinepubs/009695399/functions/getlogin_r.html.
2. Moves the check
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
submission part. Perhaps it would be smoother if someone already known
to the glibc team introduced me and this patch?
Yes, probably. I'm willing to have a go at it.
I suggest submitting two patches.
(1) the part that makes
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
Fair enough, but why undo the change to glob.m4? Shouldn't I just
change the target of the AC_DEFINE from MISSING_SYS_CDEFS_H to _SYS_CDEFS_H?
Yes, you're right.
Sorry, I'd forgotten the trick that I had suggested
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Conrad T. Pino wrote:
breaks the Windows build since Microsoft does NOT provide sys/cdefs.h
implementation.
Yes. Since we don't run configure on Windows, _SYS_CDEFS_H needed to be
defined to 1 in the windows-NT/config.h.in.in. I've done so and
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
I chose the _SYS_CDEFS_H route since it seemed simplest to me, though I
chose to name the macro `MISSING_SYS_CDEFS_H'.
Sorry, that's not right, since it fails in the following scenario:
#define MISSING_SYS_CDEFS_H 27
Bruno Haible wrote:
Derek Price [EMAIL PROTECTED] writes:
As near as I can tell, stat and lstat do not define names for their
replacements as many of the other GNULIB modules do.
Yes. The 'stat' and 'lstat' modules look incomplete. I think this should
be added to make them usable out
Larry Jones wrote:
Derek Price writes:
Larry, can you tell us if defining
_POSIX_PTHREAD_SEMANTICS would work to get the POSIX version of
getpwnam_r on Solaris?
It looks like it.
I've committed Paul's patch to the CVS CVS tree, as well as removing the
associated glob.c changes
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Sergey Poznyakoff wrote:
Derek Price [EMAIL PROTECTED] wrote:
At least as of Automake 1.9.5, and I thought for some time earlier,
automake will figure out of the above automatically since getdate.y is
AC_LIBSOURCE'd. This recent change appears
Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
+#ifdef LIMITS_H_HAS_MINMAX
+# include limits.h
+#elif SYS_PARAM_H_HAS_MINMAX
+# include sys/param.h
+#endif
This doesn't work if limits.h and sys/param.h both define MIN.
Hrm. Okay, I've fixed this, though I think
I've attached a minor patch that removes an unused variable in
lib/regex.c, thereby placating gcc -Wall.
2005-05-13 Derek R. Price [EMAIL PROTECTED]
* lib/regex.c: Remove unused variable.
Regards,
Derek
Index: lib/regex.c
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Karl Berry wrote:
Subject: [bug-gnulib] New GNULIB glob module?
Would it be possible to simply use the libc code as-is? I guess I mean,
with whatever changes are needed sent back to libc.
So much stuff in gnulib is 95% the same as libc. It doesn't
Matthias Kurz wrote:
On Thu, May 05, 2005, Paul Eggert wrote:
Derek Price [EMAIL PROTECTED] writes:
I prefer door #2. Trivial patch attached:
Thanks, but I'd rather use AC_CHECK_DECL, so I installed this instead,
into both coreutils and gnulib. Does it work?
2005-05-05
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Matthias Kurz wrote:
I do not think that this is a Solaris _bug_.
- From the point of view of the GNULIB getopt.m4 tests, if it doesn't
behave like GNU getopt, that is a bug. :)
I'd bet that every system that does not use GNU getopt will suffer
]points to the string -
|
/getopt/() returns -1 without changing /optind./
Cheers,
Derek
Derek Price wrote:
Regardless, since using an optind = 0 is not specified as supported by
POSIX, whereas optind = 1 is, and since using optind = 1 in place of
optind = 0 in CVS would avoid
Derek Price wrote:
2005-05-05 Paul Eggert [EMAIL PROTECTED]
* lib/getopt.m4 (gl_GETOPT): Check for Solaris 10 getopt, and
avoid needless checks.
Yes, this also works for me.
Okay, one more revision, to actually check if the -+ registers as an
option
79 matches
Mail list logo