summaryrefslogtreecommitdiff
path: root/indra/newview/llagent.h
diff options
context:
space:
mode:
authorNat Goodspeed <nat@lindenlab.com>2023-09-02 11:02:23 -0400
committerNat Goodspeed <nat@lindenlab.com>2023-09-02 11:02:23 -0400
commit3eea556c28f86d1c1334879ff4d7dfc36201485e (patch)
tree3efce2b58a1f17ca9c64b3d767eef512ba20cf18 /indra/newview/llagent.h
parentb13c1f77c7004ab991eb38637ccde01bc4f13b6f (diff)
SL-19243: Post xcarchive.zip instead of separate symbols tarball.
On Mac, in the CMake USE_BUGSPLAT logic, we created both xcarchive.zip (which is what BugSplat wants to see) and secondlife-symbols-darwin -64.tar.bz2 (which we don't think is used for anything). The tarball was posted to codeticket -- but why? If the point is to manually re-upload to BugSplat in case of failure, we'll do better saving xcarchive.zip to codeticket. For SL-19243, posting xcarchive.zip directly supports the goal of breaking out the upload to BugSplat as a separate step. Anyway, since xcarchive.zip is a superset of the tarball, the tarball can be recreated from the zip file, whereas the zip file can't be recreated from the tarball without opening the .dmg installer and extracting the viewer executable. If the xcarchive.zip file exists (that is, on Mac), post that to codeticket or GitHub, as applicable, instead of the tarball. In fact, in the USE_BUGSPLAT case, don't even bother creating the tarball since we're going to ignore it. Make the new build.sh logic that insists on BUGSPLAT_USER and BUGSPLAT_PASS conditional on BUGSPLAT_DB.
Diffstat (limited to 'indra/newview/llagent.h')
0 files changed, 0 insertions, 0 deletions