On 11/26/2020 6:37 AM, Antoine Jacoutot wrote:
On Wed, Nov 25, 2020 at 07:14:39PM -0500, Brad Smith wrote:
On Wed, Nov 25, 2020 at 10:39:33PM +0100, Charlene Wendling wrote:
On Wed, 25 Nov 2020 16:20:38 -0500
Brad Smith wrote:
On Wed, Nov 25, 2020 at 11:39:49AM +0100, Charlene Wendling wrote:
Hi,
http://build-failures.rhaalovely.net/sparc64/2020-11-17/x11/gnome/gucharmap.log
(same thing on macppc)
This new version of gucharmap "requires" the -Bsymbolic-functions
linker flag. Bypassing the check allows to build gucharmap.
I didn't bump REVISION; this changes nothing on amd64 and it has
never been built on ld.bfd archs.
The runtime is fine on macppc.
Comments/feedback are welcome,
Charl??ne.
[0] https://bin.charlenew.xyz/gucharmap.log
I was looking at this yesterday. I just had not sent the diff out yet.
I was committing the fix while you sending that mail. I'm pretty
sure that upstream would prefer your fix to mine.
Just FYI this was part of a bigger diff I had to remove the workaround
for lld. I previously had a similar diff to the autoconf based ports for
gucharmap and then it switched from autoconf to meson. It also included
devel/vte but that was removed.
We got that diff but you never answered the question I asked about testing.
I had set aside the remainder of the diff after the the gucharmap update
went in thinking with Meson this no longer was an issue and I wasn't sure
how to meaningfully test the other bits. But it appears there is at least
a test suite for librsvg. Running the test suite before and after I see
the same
results. libcroco seems to be only used by librsvg with the C version of the
library and that appears to work Ok.
============================================================================
Testsuite summary for RSVG 2.40.21
============================================================================
# TOTAL: 71
# PASS: 66
# SKIP: 0
# XFAIL: 0
# FAIL: 4
# XPASS: 0
# ERROR: 1
============================================================================
