I tried both off and on to see if that fixes it, no dice. 

Can you zip up your textures and the map and upload the thing? There could be something non-obvious going on here like an alpha channel or the lack of one, or even the draw order. I still need to log so I can see what your machine has. 

ericw's vcshift observation fixed --- http://quakeone.com/proquake/interims/mark_v_20150414c_windows.zip --- I rewired the command buffer to be more precise and to support multiple command buffers. Clearly, I need to completely mimick id1 with that and set missing arg pointers to an empty string or something.

@ericw -- thanks for the reminder and pointing out the limits. I had eyed the alloc block enhancement in Quakespasm a few times, but hadn't thought about it lately. 

Trying with your very latest build on my PC at home, there seem to be no issues with manual saves after all. I will verify this with the machine I use at work (where I had this glitch) tomorrow.

I intend to intensify testing this build with a complete Nehahra playthrough next to see if there are any major issues. I've been waiting for a long time for being able to play this addon with an up-to-date port. 

@fifth -- what is your volume and bgmvolume cvars set to?

In Mark V, if you turn down the volume, it also turns down the background music volume.

So if you have volume 0, the background music won't be making any noise.

I'm trying to emulate a no cd-drive scenario and the MP3 music is still playing. 

There is a link to my test here:

http://quakeone.com/forums/quake-help/quake-clients/11456-fitzquake-mark-v-fullbrights.html#post156965

Just unzip it somewhere and replace the execs with your latest ones.

I posted my console right below, at QuakeOne.

Cheers. 

dx8_mark_v.exe is loading the glow:

(...)
128 x 128 maps/dm3.bsp:tech05_1
16 x 128 maps/dm3.bsp:tech04_3
128 x 128 textures/sfloor4_2_glow (R)
128 x 128 textures/sfloor4_2 (R)
64 x 64 maps/dm3.bsp:sfloor1_2
(...)

mark_v.exe is not:

128 x 128 maps/dm3.bsp:tech05_1
16 x 128 maps/dm3.bsp:tech04_3
128 x 128 textures/sfloor4_2 (R)
64 x 64 maps/dm3.bsp:sfloor1_2

;) 

Check the filename ...

The current download, http://quakeone.com/proquake/interims/mark_v_20150414c_windows.zip, has the names are mark_v_c.exe and dx8_mark_v_c.exe

I want to be sure we are working with the right version. 

Something weird, though.

As I look through your texture list you posted at QuakeOne, I see this:

64 x 64 maps/dm3.bsp:sfloor4_4_glow
64 x 64 maps/dm3.bsp:sfloor4_4
64 x 64 maps/dm3.bsp:comp1_7
64 x 64 maps/dm3.bsp:comp1_3
64 x 64 maps/dm3.bsp:comp1_6_glow

Now the sfloor4_2_glow isn't there.

But I notice the other _glow textures are there.

I'm trying to think what circumstance could cause that particular texture to not load, but all the other _glow textures would load.

This is very inconsistent. 

However, only textures with an (R) next to them have been replaced with an external texture. 

Check to ensure that the exe name is, in fact, mark_v_c.exe and dx8_mark_v_c.exe. 

Here is pretty much the same zip you gave me but which only the current Mark V gl executable in it but includes the other needed files.

Download that, unzip and double click the mark_v_c.exe and type "map dm3" in the console.

http://quakeone.com/proquake/interims/adib.zip 

It works!

Sorry, I was running the Beta 6 from #667, not this latest. Both GL and DX versions are showing transparency and fullbrights.

Actually, DX version darkens the transparent parts of a "{" texture a little. But GL version is perfect and it's enough for me.

About what you say at #688, the glows are inside dm3.bsp, not external at the /textures folder:

64 x 64 maps/dm3.bsp:sfloor4_4_glow

Don't know how they got there, though, since iD doesn't _glow their textures. Maybe it's a texture generated in memory by your code to fullbright iD's stock textures.

Anyway, thanks for the attention and congrats for this engine. You have some great software here. 

I'm happy to see it is working for you now, that's the important part!

Now, if I can get MP3 working on Fifth's computer, all the known true "bugs" are resolved.

And then I can focus on tidy-ups like observations above about autosaves, map completion and such and rounding out the feature set for Rubicon Rumble support. 

Here's a bug report with the latest bundle of stuff. I didn't notice any issues with mark_v_c.exe, but mark_v_winquake_c.exe crashes and raises the "Mark V has stopped working" dialog if I set it to fullscreen 1080p (my monitor native resolution). Fullscreen 720p works though.

Let me know if you want any info or other tests from me. 

Interesting. Looks like one of the hard limits of vertical resolution built into the software renderer is being hit.

The crash is easy enough to reproduce. 

Savegames definitely fixed. Looks like it was an issue with the April 13 build only, then.

Regarding raising limits to play most recent custom map releases: I am definitely supporting the idea that it should be possible to play any map currently available (hopefully also future ones) without requiring additional launch parameters. That's something that should definitely be in the final release.

Since everything works fine for me now, I'll stress-test the latest build with a few old and very recent mods to see if there is any odd behavior. I am almost finished with a full playthrough of all the vanilla Quake episodes and found no reason to complain so far. 

When viewing teleporter surfaces through transparent water, the teleporters appear unfiltered, i.e. without the water layer effect.

To illustrate the effect, check this screenshot.

Using latest Mk V snapshot with following water settings:
gl_subdivide_size 16
r_oldwater 0
r_lavaalpha 0.8
r_slimealpha 0.7
r_wateralpha 0.6
r_waterquality 32
r_waterripple 3
r_waterwarp 1

Using .vis files to achieve water transparency (inside a PAK file). 

That looks more like bad depth sorting for alpha surfaces than anything else. 

new build and its predecessors -condebug is not working (qconsole.log) it remains empty

when i'm typing developer 1 the game crashes

can't run the travail pak. But it run just fine on my working PC

it seems the show_clock command doesnt work either (at least wnen DM mode is enable) it shows everytime despite enable/disable state
personally i don't like the position of the clock too. The only two(off four) of the scoreboard leaders are visible. 

Beta 7 - Windows Open GL | DX 8 | WinQuake
Source

1) Changed scr_clock to allow never drawing the clock even in deathmatch as an option. scr_clock 0 (never), scr_clock -1 (dm only, default), scr_clock 1 (always). If the clock is off in DM, will draw all 4 player slots. If the clock is on in DM will draw 2 player slots.

2) -condebug and -condebug + developer 1 are fixed. Thanks to spy for pointing this out.

3) Rewired the command buffer system to handle unexpected or uncoded arguments in the classic way that Quake handled them. My original plan was to audit every command in Quake, but I don't see an effective way to do that.

4) Increased 3 remaining limits for Rubicon Rumble that ericw says will make it playable. Have not implemented lightmap loading speed up. MAX_ENT_LEAFS was already 32 and already handles brushes in tons of visleafs about the same as Quakespasm (per MH from several months ago)

5) sv_autosave defaults off and doesn't save to config. Obviously needs 1 engine restart to clear. 

FitzQuake 0.85 draws all "liquid" surfaces in a single pass.

But I guess now that some "liquids" like water might be transparent and some "liquids" like teleporters or lava might not be ... the rendering needs a slight tweak.

Adjusted liquids to draw in 2 passes:
1) 1st pass: solid pass like teleporters
2) 2nd pass: alpha pass like r_wateralpha 0.5

Should render like how you would expect it.

Update: Teleporter + water rendering improved - Windows version
Source (may be 5-10 mins before I have it uploaded) 

can confirm all of the RRP maps load with the latest build 

@ericw -- do you happen to know the names of any of those "bad lightmap" 64-bit maps from last year? I want to experience one of those first hand.

(I can't do this with my Win32 build. But I'm hoping it crops up on the Mac build which is 64-bit.) 

jam2_mfx is a good one, just turn right from the start position and check for diagonal stripes to the right of the arch, there should be none. I posted some screenshots of that one and the fix as well on i3d: http://forums.inside3d.com/viewtopic.php?p=53977#p53977

Another is mfxsp17, just go straight ahead from the start position and check the wall panel behind the crate. It shouldn't have a black triangle in the upper left corner, here is a screenshot with the bug: https://www.dropbox.com/s/ey8htwsz37wdux9/mark_v_0000.png?dl=0