Commit graph

10 commits

Author SHA1 Message Date
Simon Howard
ef816accb3 Bump up RAM requirement and reenable broken demos
These were failing because some WADs require more than a 16MiB heap.
2024-09-15 13:33:47 -04:00
Simon Howard
00450d0c85 Switch to new build system.
This new system works based on rules generation, rather than hand-
constructed make rules. This is derived from the makefile found in
the statcheck repository at chocolate-doom/statcheck@4d1182db1b.

Part of #3.
2024-09-15 12:45:45 -04:00
Simon Howard
b1bf88a8a0 We don't need the "touch" spam in make output. 2023-10-14 14:09:07 -04:00
Simon Howard
a82dd104b0 Don't specify both -q and -u.
FreeBSD make considers this an error.
2023-10-14 14:03:55 -04:00
Simon Howard
a24ca489c8 Don't use -DD option.
The Info-ZIP version of unzip supports this but the FreeBSD doesn't.
We can get the same result by touching the file after extract.
2023-10-14 13:51:32 -04:00
Simon Howard
0f380f7778 Only list the files that don't match.
There's no need to dump the entire diff to stdout.
2023-10-07 11:54:18 -04:00
Simon Howard
c8824b131a Rewrite to make better use of make.
Instead of having the test runner do the extraction, make can do it for
us. This allows for better parallelization. It also lets the WADs be
extracted separately (and potentially modified) before the tests are
run; I'm planning to reuse this for wadptr so that I can test WADs play
back as intended after being compressed.
2023-10-05 19:21:47 -04:00
Simon Howard
f093238324 Reverse comparison order.
Missing from generated output should produce a '-' not a '+'.
2019-11-11 19:11:56 -05:00
Simon Howard
4c3293bbba Don't load DEH lump for D2TWID.
This patch changes the par times and Crispy recognizes the PAR syntax.
However, it interferes with the statdump comparison. So just don't load
it.
2019-11-11 19:09:41 -05:00
Simon Howard
8018e1831e Add makefile and testrunner script. 2019-11-11 18:17:57 -05:00