Re: MODULES.html.sh broken - unterminated substitute pattern

David Evans wrote: > Thanks for that. I've solved the problem more simply by installed GNU > sed on both FreeBSD and OS X, and it now works. I confirm that with MacOS X /usr/bin/sed, MODULES.html.sh spews error messages from 'sed' (with or without Paul's patch), whereas with GNU sed, MODULES.htm

Re: MODULES.html.sh broken - unterminated substitute pattern

In article <4eb05403.7060...@cs.ucla.edu>, Paul Eggert wrote: > On 11/01/11 11:12, David Evans wrote: > > > I will have to assume that Bash or sed on both FreeBSD and > > OS X is broken in some way. > > Quite possibly, but I've run into similar issues many times over > the years, with various

Re: MODULES.html.sh broken - unterminated substitute pattern

Paul Eggert wrote: > Replace foo=` ... sed -e COMPLICATED ... ` with > bar=COMPLICATED; foo=` ... sed -e "$bar" ... ` when COMPLICATED > contains (possibly-quoted) backslashes. Thanks, Paul. Nowadays I try to avoid backslashes inside backquote. Correctly parsing backslashes inside backquote (and p

Re: MODULES.html.sh broken - unterminated substitute pattern

On 11/01/11 11:12, David Evans wrote: > I will have to assume that Bash or sed on both FreeBSD and > OS X is broken in some way. Quite possibly, but I've run into similar issues many times over the years, with various shells, and it's fairly easy to work around these bugs. I installed the follow

Re: MODULES.html.sh broken - unterminated substitute pattern

I've now bisected this bug, and the first bad commit is fc367 Since this is many years ago, I find it hard to believe that this file, MODULES.html.sh, had been broken for this long. I will have to assume that Bash or sed on both FreeBSD and OS X is broken in some way.

split off module fchmodat from module openat

Hi Jim, The next split-off from module 'openat' is 'fchmodat'. Here's the proposed patch. Tested on a number of platforms. 2011-11-01 Bruno Haible New module 'fchmodat', split off from module 'openat'. * lib/openat.h (chmodat, lchmodat): Enable only if GNULIB_FCHMODAT is

Re: [PATCH] gitlog-to-changelog: provide a ChangeLog-repair mechanism

Eric Blake wrote: > On 11/01/2011 11:15 AM, Jim Meyering wrote: >> I've been thinking about this for years, ever since the unpleasantness >> of that first git commit that I mistakenly attributed to myself. >> >> With this new --amend=FILE option, you'll be able to maintain >> a list of pairs where

Re: [PATCH] gitlog-to-changelog: provide a ChangeLog-repair mechanism

On 11/01/2011 11:15 AM, Jim Meyering wrote: I've been thinking about this for years, ever since the unpleasantness of that first git commit that I mistakenly attributed to myself. With this new --amend=FILE option, you'll be able to maintain a list of pairs where SHA is a 40-byte SHA1 (alone on

Re: [PATCH] gitlog-to-changelog: provide a ChangeLog-repair mechanism

Jim Meyering wrote: > I've been thinking about this for years, ever since the unpleasantness > of that first git commit that I mistakenly attributed to myself. > > With this new --amend=FILE option, you'll be able to maintain > a list of pairs where SHA is a 40-byte SHA1 > (alone on a line) referr

[PATCH] gitlog-to-changelog: provide a ChangeLog-repair mechanism

I've been thinking about this for years, ever since the unpleasantness of that first git commit that I mistakenly attributed to myself. With this new --amend=FILE option, you'll be able to maintain a list of pairs where SHA is a 40-byte SHA1 (alone on a line) referring to a commit in the current

Re: [PATCH] stdalign: port better to MSVC and to Sun C 5.11

On 11/01/11 05:00, Bruno Haible wrote: > It was here: > Thanks, I added that to the ChangeLog. > 1) On mingw of 2009, the test fails > > test-stdalign.c:81: assertion failed > FAIL: test-stdalign.exe I don't use mingw, but I

Re: [PATCH] gitlog-to-changelog: fix git-log invocation

On Tue, Nov 01, 2011 at 03:19:43PM +0100, Jim Meyering wrote: > Running it on coreutils, I got this, ad infinitum: > > Use of uninitialized value $line[0] in pattern match (m//) at > build-aux/gitlog-to-changelog line 163, line 34755. Ouch. > >From bc2f20935ff445f1e55b9c7f7e90cd77ef4eb007

Re: [PATCH] gitlog-to-changelog: fix git-log invocation

Dmitry V. Levin wrote: > On Mon, Oct 31, 2011 at 05:11:28PM +0100, Jim Meyering wrote: >> Dmitry V. Levin wrote: >> > git-log mishandles date strings before 1970-01-01 UTC, and there is >> > no use to specify --since=1970-01-01 by default anyway. >> > * build-aux/gitlog-to-changelog: By default, wh

MODULES.html.sh broken - unterminated substitute pattern

While running MODULES.html.sh I get this repeated error: sed: 1: "s,\([^a-zA-Z_]\)\(FD_CL ...": unterminated substitute pattern ./gnulib-tool: line 818: printf: write error: Broken pipe caught signal 13 Uisng the trace facility in Bash, it seems to be happening at line 1527 Tested on OS X 10.6.

Re: [PATCH 2/2] gitlog-to-changelog: support 'tiny change' commits.

Gary V. Vaughan wrote: >> The notation requirement made more sense back each ChangeLog entry >> was not tied to the associated patch. > > Before modern VCS you mean? Right. > That's before my time then... even CVS still ties > the ChangeLog to the patch does it not? No. With CVS, it was not unc

Re: [PATCH 2/2] gitlog-to-changelog: support 'tiny change' commits.

Hi Jim, Karl, On 1 Nov 2011, at 20:13, Jim Meyering wrote: > Gary V. Vaughan wrote: >> >> diff --git a/ChangeLog b/ChangeLog >> index f370be6..d59d9f9 100644 >> --- a/ChangeLog >> +++ b/ChangeLog >> @@ -1,5 +1,14 @@ >> 2011-11-01 Gary V. Vaughan >> >> +gitlog-to-changelog: support `tiny c

Re: [PATCH 2/2] gitlog-to-changelog: support 'tiny change' commits.

Gary V. Vaughan wrote: > Here is another patch to solve a similar issue with ChangeLog `tiny change' > annotations, by interpreting a 'Copyright-paperwork-required: No' line in > the git log message to mean that the ChangeLog output requires the `tiny > change' annotation. > > Because the annotatio

Re: [PATCH 1/2] gitlog-to-changelog: support multi-author commits.

Gary V. Vaughan wrote: > More generally useful gnulib-local goodness from my Libtool `next' branch: > > I'm sure this is far from idiomatic Perl, but I'd very much like for this > patch or something similar to be pushed so that FSF projects have a means > to correctly track multiple patch authors w

[PATCH 2/2] gitlog-to-changelog: support 'tiny change' commits.

Here is another patch to solve a similar issue with ChangeLog `tiny change' annotations, by interpreting a 'Copyright-paperwork-required: No' line in the git log message to mean that the ChangeLog output requires the `tiny change' annotation. Because the annotation is added to the date_line, there

[PATCH 1/2] gitlog-to-changelog: support multi-author commits.

More generally useful gnulib-local goodness from my Libtool `next' branch: I'm sure this is far from idiomatic Perl, but I'd very much like for this patch or something similar to be pushed so that FSF projects have a means to correctly track multiple patch authors with a generated ChangeLog file.

Re: [PATCH] stdalign: port better to MSVC and to Sun C 5.11

Hi Paul, Thanks for the update. > I think these problems were reported by Bruno Haible, in email > that I've unfortunately misplaced. It was here: With this patch, the crashes on FreeBSD 6.4 and Cygwin 1.5.25 are gone. MSVC 9

Re: stdalign and alignof

Paul Eggert wrote on 2011-10-27: > none of this stuff affects the existing > alignof module. Perhaps the documentation for each should > refer to the other? Maybe alignof should use stdalign? The highest priority is to fix the conflict between the two modules, that now both define an alignof() m

Re: split off module fchownat from module openat

Bruno Haible wrote: > Jim Meyering wrote: >> I'll test via coreutils once it's pushed. > > OK, I've pushed it, with this additional change squashed in: ... Thanks. That works fine, so I've updated coreutils: >From 65527e1c1a69a9ec8201435a6361964bcf31a480 Mon Sep 17 00:00:00 2001 From: Jim Meyerin

Re: split off module fchownat from module openat

Jim Meyering wrote: > I'll test via coreutils once it's pushed. OK, I've pushed it, with this additional change squashed in: --- m4/fchownat.m4.orig Tue Nov 1 11:06:48 2011 +++ m4/fchownat.m4 Tue Nov 1 11:05:57 2011 @@ -15,7 +15,6 @@ AC_DEFUN([gl_FUNC_FCHOWNAT], [ AC_REQUIRE([gl_UNIS

Re: split off module fchownat from module openat

Bruno Haible wrote: > Hi Jim, Eric, > > Here is a proposed patch to split fchownat() off from the module 'openat'. > > The immediate motivation is that on MSVC, there is no and no > , therefore code that fiddles with owners and groups leads to > many more compilation problems than code that just d