qbismSuper8 alpha 061410 released

[mankrip]

A kinda old thread that I didn't see before.

This is really cool, I'll take a look on it.

[mrmmaclean]

Well, I take back my apology then!

This is pretty much a copy and paste from the qbism forum (which sadly doesn't get too much action) but I have a (possibly very newby) question: How do I use a skybox in this port???

In the Flash version, I put the 6 pcx files in gfx/env of pak0.pak and all I get from "loadsky <filename>" is "Couldn't load". I tried using "r_skyname <filename>" (only the main name, not the suffixes), and then "loadsky <filename>" (without quotes and angle brackets, of course) but the sky just goes black...

In the windows version, it simply crashes.

I was digging around in the source and comparing it to proquake for reference (I made a mini project of porting the ver 4.51 to Flash to learn some of the guts of the engine) and noticed that "svc_skybox" is commented out in qbismSuper8. Simply uncommenting that and tweaking some odds and ends didn't seem to help... what am I missing here?

[mankrip]

Use this PAK file, and type loadsky des_ in the console.

I got this skybox from Ajay's Bulldog Stadium mod, resized to 512x512, converted with dithering to Quake's 8-bit pallete and saved it as PCX files.

The underline character in their filenames isn't a requirement, I just used it in the names of these files to make them easier to read.

Anyway, adding console parameter autocompletion for skyboxes should help.

[mrmmaclean]

Still the same result, sadly. "Couldn't load" in the Flash trace and simply crashing the Windows version.

I did the same with some of the Kothic skyboxes, applying the quake palette and converting to pcx (shortening filenames when necessary just in case).

[mrmmaclean]

Solved! I simply replaced:

Code: Select all

sprintf (pathname, (const char *)sizeof(pathname), "gfx/env/%s%s.pcx\0", skyname, suf[r_skysideimage[i]]);

in r_sky.c with this:

Code: Select all

snprintf (pathname, sizeof(pathname), "gfx/env/%s%s.pcx", skyname, suf[r_skysideimage[i]]);

Seems to work fine and hopefully won't break anything! Thanks to referencing ProQuake 4.51 once again.
edit: seems that was the method used in Makaqu source as well. I wonder why qbism changed it?

[Sajt]

Code: Select all

sprintf (pathname, (const char *)sizeof(pathname), "gfx/env/%s%s.pcx\0", skyname, suf[r_skysideimage[i]]);

[mrmmaclean]

My thoughts exactly. Basically, the (const char *) is needed to stop Adobe Alchemy from throwing a warning but makes the function just not work... doesn't sound like a good trade off to me.

[qbism]

I think const char * was a partial misguided attempt to fix the issue. sprintf should not have the size argument at all. snprintf should, but it's C++, undefined when I compile Windows version. What it REALLY should be is Q_snprintfz

Code: Select all

Q_snprintfz (pathname, sizeof(pathname), "gfx/env/%s%s.pcx\0", skyname, suf[r_skysideimage[i]]);

Yeah, and I should check qbism site occaisionally

[mrmmaclean]

What it REALLY should be is Q_snprintfz

Code: Select all

Q_snprintfz (pathname, sizeof(pathname), "gfx/env/%s%s.pcx\0", skyname, suf[r_skysideimage[i]]);

Yeah, and I should check qbism site occaisionally

Nice, thanks! I'm pretty new to C/C++ in general and web reference tends to lump them together a lot. Alchemy can compile either as well, so since I'm working primarily with flash I don't notice, FML.

And on checking the site: Real life tends to make things like that kinda tough, so no worries! Thank you for the efforts so far with Super 8, I'm having a blast messing around with it, and thanks to mk for the work on Makaqu and the test skybox.

[Sajt]

My thoughts exactly. Basically, the (const char *) is needed to stop Adobe Alchemy from throwing a warning but makes the function just not work... doesn't sound like a good trade off to me.

The warning was given for a reason, you know... What you did was tell the compiler to just shut up and accept the wrong arguments, which end up as, effectively, undefined garbage (an integer passed as a string address, for example)! You have to tread lightly around the C language sometimes... (and the error/warning messages are not always helpful to those who aren't already familiar with how to "interpret" them)

[mrmmaclean]

The warning was given for a reason, you know... What you did was tell the compiler to just shut up and accept the wrong arguments, which end up as, effectively, undefined garbage (an integer passed as a string address, for example)!

I'm not sure if this is directed towards me since the (const char *) was in the qbismSuper8 source code and not my own.

That's why it was so hard to find the problem. The compiler was happy so I had to chase keyword finds through the source to get what was up.

That's when I found qbism's self proclaimed "misguided" solution. REMOVING that type declaration popped out a compiler warning, so that's when I used the sndprintf method instead. As a fix rather than just sending safe garbage. Though I wasn't aware that sndprintf was C++ rather than C. qbism came back to provide the C equivalent of my solution!

edit: spelling

[Baker]

snprintf isn't C++, it's C. It just isn't part of the ISO C-99 spec that some versions of Microsoft Visual Studio use [or maybe it isn't and Microsoft varies from the spec]; it has long been totally supported by gcc --- and as "FlashQuake" uses a gcc variant ... well ... [As Spike has noted, the MSVC 6.0 version of "_snprintf" does not null terminate strings ...]

Note: Q_snprintfz null terminates strings (adds the null character terminator, character 0 to the end), so doing a "\0" is entirely redundant. The "z" in "Q_snprintfz means that it null terminates it).

[mrmmaclean]

Ahhh, gotcha. Thanks Baker. I still have yet soooo much to learn!

[Baker]

Ahhh, gotcha. Thanks Baker. I still have yet soooo much to learn!

Who doesn't? I still have a lot to learn.

The main characteristic of gaining experience is figuring out the massive set of things you do not know. Believe me, I know what I don't know.

There is something to be said of knowing what you don't know; the word is "competence".

[qbism]

The "z" in "Q_snprintfz means that it null terminates it).

I did not know that.

If I read it correctly MicroSoft's _snprintf does not guarantee null termintation as gcc's snprintf does. http://msdn.microsoft.com/en-us/library/2ts7cx93.aspx