Getting 2.05-beta To Compile and Run Under Linux

[NightStalker]

SVN revision: 6735

Duzenko,

 

Your patch looks perfect and tested just as nicely! Thank you!

 

For the record, I built TDM from latest SVN (#6735), with Hamlet's published patch to 'idlib/geometry/RenderMatrix.cpp' (because I like his patch better than mine) and with my published patch to the DevIL library header files (because, for now, I like my patch well enough) using 32-bit Slackware 14.2. The video card is a GeForce 7600GT with the nVidia 32-bit 304.132 driver.

 

I used the same 2 missions that I've been using for testing this: "Closemouthed Shadows" and "The Outpost". Both are running fine, so I can say with confidence that the patch is, unsurprisingly, working as intended.

 

I have not (yet) applied the patch to my 2.05-beta tree and built from that because I'm still finalizing my testing of Baal's Linux crashes and I don't want to corrupt that (local) tree any more than I already have. So give me another day to apply your patch to 2.05-beta, re-test (with more missions too), then I will report back, at which time I would "ack" the patch, FWIW, for inclusion into the official 2.05 tree in SVN.

 

Thanks again for the patch! One less GL "time bomb" awaiting the unwary!

[NightStalker]

I have not (yet) applied the patch to my 2.05-beta tree and built from that because I'm still finalizing my testing of Baal's Linux crashes and I don't want to corrupt that (local) tree any more than I already have. So give me another day to apply your patch to 2.05-beta, re-test (with more missions too), then I will report back, at which time I would "ack" the patch, FWIW, for inclusion into the official 2.05 tree in SVN.

OK, I've finally had time to test this patch (from SVN #6735, as referred to in post #25 of this thread) against my actual 2.05-beta tree and run it (under 32-bit Slackware 14.2). And, again unsurprisingly, it works great (under "Closemouthed Shadows", "Coercion", "Flakebridge Monastery", and "WS1: In The North"). Sorry for the delay on this test, duzenko.

 

@grayman: Assuming you are the one applying changes to the 2.05 tree, at your next opportunity, please apply duzenko's patch since it fixes "breakage" introduced between 2.04 and 2.05 for users with video hardware+driver that does not support the OpenGL "pixel buffer object" extension.

[grayman]

SVN revision: 6735

 

Am I supposed to merge this to 2.05? (And 6736?)

[duzenko]

I think 6735 should be merged to 2.05 for the reasons listed in #27

6736 probably best left for 2.06

[grayman]

@NightStalker,

 

I added you to a PM between myself and duzenko from a few weeks back re: linux build errors I was seeing in the trunk.

 

I'm trying to build from the head of the trunk source again and am getting the errors noted in my Dec 15 5:10PM note in that PM, the one that begins with "I added RenderMatrix to Sconscript.idlib:".

 

Could you please coordinate duzenko's work on fixing those errors with the work you're doing?

 

Thanks.

[NightStalker]

I added you to a PM between myself and duzenko from a few weeks back re: linux build errors I was seeing in the trunk.

OK, good, thanks. Give me a chance to read it all and digest it a bit and I'll get back to you guys soon.

Could you please coordinate duzenko's work on fixing those errors with the work you're doing?

Gladly. Will report when I know more.

 

As for the 2.05 tree, I agree with everything duzenko said in his post #29. That is, I was only recommending that you add his #6735 commit to the 2.05 tree.

[NightStalker]

I'm trying to build from the head of the trunk source again and am getting the errors noted in my Dec 15 5:10PM note in that PM, the one that begins with "I added RenderMatrix to Sconscript.idlib:".

@Grayman: OK, I've properly digested that PM now.

 

Quick Answer: Please sit tight for a day or two. After that, I'll PM you asking you to re-attempt your Linux compile.

 

Longer Story:

 

One of Hamlet's patches should fix that part of your Linux-compile dilemma. His patch is essentially an elegant version of an unpublished quick hack I made weeks ago to get TDM to compile under Linux.

 

Hamlet and duzenko have already "weighed in", essentially approving my latest collection of Hamlet's patches and one of my own patches.

 

Grayman, if you are interested in the details and haven't already read it, you might want to read from this post onward.

 

So I expect to commit these patches to SVN 'trunk' within a day or two. I'm wanting to give all Linux compilers a chance to comment, but I may accelerate that process based on Hamlet's and duzenko's comments.

 

Hope that helps....

 

@Grayman: Is it possible, and if so, is it permitted for me to create a public branch in SVN for this patchset, as suggested by Hamlet in that thread? If the answer to either of those questions is "No", then I will simply commit the patchset to SVN 'trunk' reasonably soon.

Edited January 1, 2017 by NightStalker

[grayman]

Yes, you can creates branches for experiments.

 

Just give the experimenters the URL of the branch, i.e. "https://svn.thedarkmod.com/svn/darkmod_src/branches/linux_experiment"

[NightStalker]

OK, folks... I've finally committed the appropriate changes to SVN trunk. This includes 4 revisions, 6738 - 6741.

I believe this will make "latest SVN" compile-able for all GCC users (4.4.3 to 6.2.x), but if not, please let me know.

My thanks to all who tested this, especially duzenko for compiling it under Windows and Hamlet for supplying the bulk of the fixes.

I've PMed grayman to ask him to test this newest code under Linux.

I appreciate everyone's patience on this!