On Wed, Dec 06 2017, Ævar Arnfjörð Bjarmason jotted:
> On Wed, Dec 6, 2017 at 7:56 PM, Daniel Jacques wrote:
>> On Wed, Dec 6, 2017 at 1:47 PM, Junio C Hamano wrote:
>>>
>>> Johannes Sixt writes:
>>>
>>> > The updated series works for me now. Nevertheless, I suggest to squash
>>> > in the foll
On Wed, Dec 6, 2017 at 7:56 PM, Daniel Jacques wrote:
> On Wed, Dec 6, 2017 at 1:47 PM, Junio C Hamano wrote:
>>
>> Johannes Sixt writes:
>>
>> > The updated series works for me now. Nevertheless, I suggest to squash
>> > in the following change to protect against IFS and globbing characters in
On Wed, Dec 6, 2017 at 1:47 PM, Junio C Hamano wrote:
>
> Johannes Sixt writes:
>
> > The updated series works for me now. Nevertheless, I suggest to squash
> > in the following change to protect against IFS and globbing characters in
> > $INSTLIBDIR.
>
> Yeah, that is very sensible.
>
> > diff -
Johannes Sixt writes:
> The updated series works for me now. Nevertheless, I suggest to squash
> in the following change to protect against IFS and globbing characters in
> $INSTLIBDIR.
Yeah, that is very sensible.
> diff --git a/Makefile b/Makefile
> index 7ac4458f11..08c78a1a63 100644
> --- a
Am 05.12.2017 um 22:35 schrieb Junio C Hamano:
> Dan Jacques writes:
>
>> Thanks for checking! The patch that you quoted above looks like it's from
>> this "v4" thread; however, the patch that you are diffing against in your
>> latest reply seems like it is from an earlier version.
>>
>> I believ
Dan Jacques writes:
> Thanks for checking! The patch that you quoted above looks like it's from
> this "v4" thread; however, the patch that you are diffing against in your
> latest reply seems like it is from an earlier version.
>
> I believe that the $(pathsep) changes in your proposed patch are
Johannes Sixt wrote:
> I don't know what I tested last week; most likely not the version of the
> patch I quoted above.
>
> Today's version, with the tip at 5d7f59c391ce, is definitely bogus
> with its quoting. It needs the patch below, otherwise an unquoted
> semicolon may be expanded from $(path
Johannes Sixt writes:
> Today's version, with the tip at 5d7f59c391ce, is definitely bogus
> with its quoting. It needs the patch below, otherwise an unquoted
> semicolon may be expanded from $(pathsep). This would terminate the sed
> command, of course.
Of course ;-)
Somehow I was lucky that t
Am 01.12.2017 um 18:25 schrieb Johannes Sixt:
> Am 01.12.2017 um 18:13 schrieb Johannes Schindelin:
>> Hi Hannes,
>>
>> On Fri, 1 Dec 2017, Johannes Sixt wrote:
>>
>>> Am 29.11.2017 um 16:56 schrieb Dan Jacques:
@@ -1989,6 +1986,15 @@ GIT-PERL-DEFINES: FORCE
echo "$$FLAGS" >$@
On Sun, Dec 03 2017, Junio C. Hamano jotted:
> Johannes Sixt writes:
>
>>> + sed -e 's=@@PATHSEP@@=$(pathsep)=g' \
>>
>> This doesn't work, unfortunately. When $(pathsep) is ';', we get an
>> incomplete sed expression because ';' is also a command separator in
>> the sed language.
>
> It is co
Johannes Sixt writes:
>> +sed -e 's=@@PATHSEP@@=$(pathsep)=g' \
>
> This doesn't work, unfortunately. When $(pathsep) is ';', we get an
> incomplete sed expression because ';' is also a command separator in
> the sed language.
It is correct that ';' can be and does get used in place of LF wh
On Dez 01 2017, Dan Jacques wrote:
> I am not a `sed` wizard, but perhaps the tool is ignoring the semicolon
> because
> it's in the middle of the "s" expression? The shell may similarly be ignoring
> it
> because it's nested in between single-quotes?
As far as POSIX is concerned, semicolons a
On Fri, 1 Dec 2017, Johannes Sixt wrote:
>>> This doesn't work, unfortunately. When $(pathsep) is ';', we get an
>>> incomplete sed expression because ';' is also a command separator in the
>>> sed language.
>>
>> Funny, I tried this also with ';' as pathsep, and it worked in the Git for
>> Windo
Am 01.12.2017 um 18:13 schrieb Johannes Schindelin:
Hi Hannes,
On Fri, 1 Dec 2017, Johannes Sixt wrote:
Am 29.11.2017 um 16:56 schrieb Dan Jacques:
@@ -1989,6 +1986,15 @@ GIT-PERL-DEFINES: FORCE
echo "$$FLAGS" >$@; \
fi
+GIT-PERL-HEADER: $(PERL_HEADER_TEMPLATE) GIT-
Hi Hannes,
On Fri, 1 Dec 2017, Johannes Sixt wrote:
> Am 29.11.2017 um 16:56 schrieb Dan Jacques:
> > @@ -1989,6 +1986,15 @@ GIT-PERL-DEFINES: FORCE
> > echo "$$FLAGS" >$@; \
> >fi
> > +GIT-PERL-HEADER: $(PERL_HEADER_TEMPLATE) GIT-PERL-DEFINES perl/perl.mak
> > Makefile
> >
Am 29.11.2017 um 16:56 schrieb Dan Jacques:
@@ -1989,6 +1986,15 @@ GIT-PERL-DEFINES: FORCE
echo "$$FLAGS" >$@; \
fi
+GIT-PERL-HEADER: $(PERL_HEADER_TEMPLATE) GIT-PERL-DEFINES perl/perl.mak Makefile
+ $(QUIET_GEN)$(RM) $@ && \
+ INSTLIBDIR=`MAKEFLAGS= $(
Currently, the generated Perl script headers are emitted by commands in
the Makefile. This mechanism restricts options to introduce alternative
header content, needed by Perl runtime prefix support, and obscures the
origin of the Perl script header.
Change the Makefile to generate a header by proc
17 matches
Mail list logo
