Michael Francis

There have always been efforts to improve WAD-locating, even in recent times. I would say it's still worth it, but it just happens to be a fairly sizable undertaking.

I would like if, from `d_main.c`, the call to `D_FindIWAD` always returns an allocated string (whether it was `strdup`'d from a constant or not, who cares), so that way a...

@AXDOOMER do you know what can be used in C to easily gather the OS limit?

Sounds like #660 and #975?

I've been using `--track-origins=yes` in place of `--show-leak-kinds=all`. Mind you, a lot of my own extra errors are coming from SDL and my GPU. Still, several memory errors of the...

This is with `valgrind --leak-check=full --track-origins=yes doomretro -iwad DOOM2.WAD` ``` ==24020== Conditional jump or move depends on uninitialised value(s) ==24020== at 0x5559CF9: strlen (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==24020== by 0x6301204: fputs (iofputs.c:33)...

Fixed. :+1:

`lumps_str` gets `malloc()`'d but then those bytes are lost by the later call to `commify()`. ```diff diff --git a/src/w_wad.c b/src/w_wad.c index de6cf3f2..4a80665a 100644 --- a/src/w_wad.c +++ b/src/w_wad.c @@ -146,14 +146,13...

What about taking the Crispy approach of still showing the message even if DSSECRET isn't present, using `sfx_itmbk` as a fallback?

Any update on this?