[ET Trac] [Einstein Toolkit] #1661: ExternalLibraries/PAPI does not build with gcc 4.8.2
Einstein Toolkit
trac-noreply at einsteintoolkit.org
Mon Sep 29 11:45:19 CDT 2014
#1661: ExternalLibraries/PAPI does not build with gcc 4.8.2
------------------------------------+---------------------------------------
Reporter: tbode | Owner: tbode
Type: defect | Status: new
Priority: major | Milestone: ET_2014_11
Component: EinsteinToolkit thorn | Version: development version
Resolution: | Keywords: PAPI
------------------------------------+---------------------------------------
Comment (by rhaas):
I'm not sure if using a hand-written search-and-replace is the best
solution to this. It seems to me that patches are fragile only because we
tended to include in them the name of the output directory (so papi-5.3 or
so appears in the patch file). However that can be avoided if instead of
{{{
patch -p0 <somepatch.patch
}}}
we use
{{{
pushd ${NAME}
patch -p1 <somepatch.patch
}}}
as Tanja's proposed fix does.
The major disadvantage with using perl to me seems that it is actually too
flexible. Patch files have a fairly standard syntax and it is
straightforward to see what is changed. With a potentially complex perl
script it is much harder to see what is changed (and in which files). It
also requires that one reads all of configure.sh to see what changes are
done.
There's also the usual issue that {{{find -print | xargs foo}}} does not
deal with file names containing spaces (or more exotic things like
newlines) but given the "portable" solutions mentioned on
http://www.unix.com/man-page/POSIX/0/xargs/ I am inclined to wait for it
to actually cause problems before even attempting to implement those
solutions.
Please note: I fully agree that the patches as currently used are fragile
and require lots of work maintaining (and Erik shoulders most of the
burden of maintaining most of the ExternalLibraries) and that the perl
script proposed will very likely be more robust against trivial changes in
the source files.
--
Ticket URL: <https://trac.einsteintoolkit.org/ticket/1661#comment:3>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit
More information about the Trac
mailing list