Hi Damon
On 07/02/2013 10:53 PM, Damon Courtney wrote:
> When building Apache, Tcl and Rivet from source, I always have to
> build Rivet with:
>
> ./configure --with-apr-config=/usr/local/apr/bin/apr-1-config
>
> Else the Rivet build fails because it can't find apr.h. I built and
> installed the APR package in its default place, so this is where it
> wants to live. Can we make Rivet's configure script check for this?
>
I thought it did
AC_DEFUN([APR_HANDLING],[
AC_MSG_CHECKING(for Apache apr)
AC_ARG_WITH(
apr_config,
[ --with-apr-config=FILE apr portable apr-1-config path],
[if test -x "${with_apr_config}"; then
apr_found="${with_apr_config}"
fi]
,
[if test -x "${apache_base}/bin/apr-1-config" ; then
....
])
why didn't it work for you?
> Also, just a note, but the tclrivet package needs an update to the
> parser to grok the new <?= ?> construct. Not a big deal, but I just
> noticed it.
>
Again I thought it did
Adding branch 2.1
------------------------------------------------------------------------
r1337938 | mxmanghi | 2012-05-13 19:17:12 +0200 (Sun, 13 May 2012) | 9 lines
* rivet/packages/tclrivet/tclrivetparser.tcl: Tcl parser implements
new shorthand echo syntax (closes bug #53217)
in fact
proc tclrivetparser::parse { data outbufvar } {
.....
while {$i < $len} {
incr i
set cur $next
set next [string index $data $i]
if { $inside == 0 } {
# Outside the delimiting tags.
if { $cur == [string index $starttag $p] } {
incr p
if { $p == [string length $starttag] } {
if {$next == "="} {
# puts stderr "shorthand begin detected"
append outbuf "\"\n $outputcmd "
set shorthand 1
incr i
set next [string index $data $i]
} else {
append outbuf "\"\n"
}
....
IIRC it should be working, but admittedly there is no test for
tclrivetparser::parse in the test suite, only the C parser is tested.
> So, with 2.1.2, we don't install packages at all by default? The
> user has to explicitly call 'make install-packages' to get them?
> What was the reason behind this? Call me crazy, but the point of
> most installations today is to install all of the packages but load
> on demand, which is what we were already doing.
>
I understand, my choice is arguable and I would have liked feedbacks
when I proposed it
http://mail-archives.apache.org/mod_mbox/tcl-rivet-dev/201306.mbox/%3C51B5E2F6.2030109%40unipr.it%3E
A better approach would have been to preserve 'make install' as an
'install everything' target and split it into 2 depends on
'install-packages' and 'install-binaries' targets. I think we could go
that way quite safely and I will get the blame for not having thought of
it on time and for having gotten out a lame version of rivet.
The proposal came up from my extra engagement as rivet package
maintainer for Debian/Ubuntu. More specifically: Ubuntu is going a
multi-architecture organization which, needless to say, means
arch-independent files can be shared between arch-specific
libraries/executables. Debian itself is stressing out the separation
between arch-independent and arch-dependent builds, I presume to
partially relieve their hardware from the burden of building the same
packages to the same extent for each supported architecture.
--
-- Massimo Manghi
Dipartimento di Neuroscienze
Unità di Biofisica e Fisica Sanitaria
via Volturno 39
43125 Parma
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
