Well, could you "group" them under a define or something? In that way I'd just recompile with the define on and that's it.
Spike wrote: the current intended behaviour is for csqc to no longer receive CSQC_InputFrame calls when this cvar is active (DP lacks these calls too), but it also doesn't get tested much.
Whaaatt? And how can I handle input? Without CSQC_InputFrame I feel naked!
Well, I could use .think but it doesn't feel quite the right place for those stuff.. I (ab)use of it even on my only-csqc games!
Spike wrote:cl_loopbackprotocol controls the protocol used to connect the internal client to its internal server. when using my purecsqc mod the internal server is not meant to be in use at all (do NOT use the map command - purecsqc registers a csmap command that should be used instead), so this cvar will affect absolutely nothing in the specific case that you state.
Again I used words improperly!
When I said "purecsqc" I didn't mean your purecsqc mod but a pure CSQC installation starting from scratch (basically no qwprogs.dat, only csprogs.dat with CSQC_Ent_Update and CSQC_Parse* stuff off and pr_csqc_formenus 1 set in .cfg). Even for craFTEr I started from blank project but it has same issues with new version of FTE (v5143 and up)
Spike wrote:recent builds have focused on emulating DP's csqc a little better, primarily in order to run xonotic (requires a new default.cfg as well as a few other xonotic-specific fixups, so don't expect it to work out of the box without those).
probably I broke something in the wrong bit of code while trying to work around xonotic's flawed prediction logic (which is a bit of a nightmare tbh). if its flickering exactly every 1 in 64 frames then that'd probably be a simple off-by-one issue... other frequencies would be race conditions...
But, if I set cvars as I wrote above, will it work as previous versions? I mean, without problems? Or could it lead to crashes or weird behaviours?
LITTLE OFF TOPIC: are you still working on the NOLEGACY define? Because, that WOULD BE GREAT to add a NOLEGACYPURE define that takes in consideration only CSQC (so there wouldn't be need for the pr_csqc_formenus 1 at startup) and that don't even compile all the quake* stuff but only .iqm and skeletal, input, audio and render stuff (probably not even BSP)