109 lines
3.6 KiB
Plaintext
109 lines
3.6 KiB
Plaintext
Part 1: Port to NetBSD
|
|
======================
|
|
Build system doesn't support NetBSD.
|
|
|
|
[X] Sent patches to upstream
|
|
[X] Imported proposed patchset from upstream
|
|
=> PKGREVISION not bumped, because binary package has not changed
|
|
[X] Renamed man page template to doomlegacy.1.in (suggested by wiz@)
|
|
=> PKGREVISION not bumped, because binary package has not changed
|
|
[X] Move man page to section 6 (it's a game)
|
|
=> Bump PKGREVISION
|
|
[X] Test building of the package again
|
|
=> Tested to work with SVN revision 1448
|
|
[X] Remove NetBSD specific patches after upstream integration
|
|
|
|
Part 1 finished.
|
|
|
|
|
|
Part 2: Fix large blockmaps on big endian machines
|
|
==================================================
|
|
On my Apple Power Mac G5 some modern (large) maps fail to load with errors like:
|
|
|
|
|
| Error: Blockmap offset[5556]= -32745, exceeds bounds.
|
|
|
|
This bug looks not security relevant, because the code catches it and gracefully
|
|
terminate the engine.
|
|
|
|
[X] Update to SVN revision 1453
|
|
=> Problem still exist
|
|
[X] Rework pkgsrc patches rejected by upstream (disable launcher and keyboard
|
|
setup +/- for automap zoom) for SVN revision 1453
|
|
[X] Add patch for (int16_t) integer overflow while loading Blockmap
|
|
Note: This should never happen on little endian machines
|
|
=> PKGREVISION not bumped (SVN based packages use build date in version)
|
|
[X] Sent Blockmap patch to upstream (Bug #653)
|
|
[X] Imported proposed patchset d42_blockmap_endian.diff from upstream
|
|
=> Drop former pkgsrc patch
|
|
=> PKGREVISION not bumped (SVN based packages use build date in version)
|
|
=> Works as expected (Reported in bugtracker for #653)
|
|
[X] Patches merged in SVN revision 1456
|
|
[X] Remove corresponding pkgsrc patches after upstream integration
|
|
|
|
Part 2 finished.
|
|
|
|
|
|
Part 3: Torches are displayed crippled with software renderer
|
|
=============================================================
|
|
This problem was not present in 1.47.2, maybe a side effect of OpenGL corona
|
|
changes since last release. Can be tested with:
|
|
|
|
|
| $ doomlegacy -iwad doom2.wad -warp 5
|
|
|
|
(walk through the opening and look at the torch in front of you)
|
|
|
|
Already reported upstream as Bug #652 (not present in OpenGL mode and really
|
|
a side effect of the corona code changes).
|
|
|
|
Suggested workaround: Deactivating coronas by going to:
|
|
|
|
"Options -> Effects Options -> Light Options"
|
|
|
|
Set "Corona" option to "Off" (verified to work on NetBSD).
|
|
|
|
Other workaround: It seems that only the option "Corona_Draw" set to "Additive"
|
|
is problematic. Coronas with other settings work.
|
|
|
|
[X] Test patch d41_draw8_alpha.diff from upstream
|
|
Patch not released yet
|
|
=> Patch merged in SVN revision 1455
|
|
=> Works as expected with SVN revision 1458
|
|
|
|
Part 3 finished.
|
|
|
|
|
|
Part 4: Problem with loading compressed blockmaps
|
|
=================================================
|
|
Some compressed blockmaps can interfere with the overflow detection.
|
|
|
|
[X] Reported upstream as Bug #654
|
|
[X] Fixed in SVN revision 1460
|
|
=> Works as expected (Reported in bugtracker for #654)
|
|
|
|
Part 4 finished.
|
|
|
|
|
|
Part 5: Rendering problems on big endian machine
|
|
================================================
|
|
On some maps the rendering is broken (only in parts of the map).
|
|
|
|
[X] Reported upstream as Bug #655
|
|
[X] Update to SVN revision 1473
|
|
=> Problem is only partly fixed (for some maps)
|
|
Bug #655 was a big endian problem, now fixed
|
|
=> Rendering problems with other map are unrelated
|
|
Ticket #657 opened for them
|
|
Looks like the bugs are in the level data, not in the engine
|
|
|
|
Part 5 finished.
|
|
|
|
|
|
Part 6: Import final release
|
|
============================
|
|
|
|
[ ] Import version 1.48
|
|
Not released yet
|
|
|
|
|
|
EOF
|