Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
Rune Morling
@ermo
(boost is only for one to two cores)
Has compilation become slower than it used to be? Things look and feel ... different.
or maybe it's just because I'm used to compiling on a 'slightly' faster box.
(twice the frequency, twice the cores)
Rune Morling
@ermo
@BenjamenMeyer My impression is that the PIE flag in cmake doesn't actually add/enable the correct CXX flags when you turn it on
I needed to manually add -fPIE to the CXXFLAGS_<something> field after toggling to advanced mode in ccmake
(I don't use the script to build, so maybe that's why)
dear me, it takes around 21 minutes of wall time to build here D:
Rune Morling
@ermo
it was CMAKE_CXX_FLAGS
(and now it builds)
ermo@solmate:~/repos/Vega-Strike-Engine-Source0.8.x
$ ./bin/vegasettings --target ../privateer_wcu
Set data directory to ../privateer_wcu
Error: Failed to find Version.txt anywhere.
ermo@solmate:~/repos/Vega-Strike-Engine-Source0.8.x
$ ls -la ../privateer_wcu/Version.txt 
-rw-rw-r-- 1 ermo ermo 4 Jun  9 23:20 ../privateer_wcu/Version.txt
Benjamen Meyer
@BenjamenMeyer
@ermo latest master should set -pie and -fpie correctly; that should be in 0.8.x too
if not, we need to backport it
Rune Morling
@ermo
I literally just compiled latest 0.8.x and it didn't add -fPIE when enabling the ccmake option as described above, which made c++ barf all over itself.
Next thing is that Version.txt in PWCU isn't being recognised despite the fact that it exists as shown in the output above.
So from a consumer standpoint, this feels like "pre-alpha" xD
Rune Morling
@ermo
For now, I've symlinked .vegastrike/ to .pu/ and copied the default privateer_wcu vegastrike.config file there. I just managed getting a load screen to pop up, but my system dumped core with a floating point exception, so rebuilding with a different config.
this is "fun"
(for a certain definition of "fun" -- YMMV)
Rune Morling
@ermo
Still getting a floating point exception; pasting here in case anybody has a clue:
Thread 1 "vegastrike-engi" received signal SIGFPE, Arithmetic exception.
0x0000555555b77566 in Movable::Movable (this=0x7fffffffc1b0)
    at /home/ermo/repos/Vega-Strike-Engine-Source/engine/src/cmd/movable.cpp:74
74            sec = GameConfig::GetVariable( "graphics", "insys_jump_ani_second_ahead", 4) / (GameConfig::GetVariable( "physics", "game_speed", 1) * GameConfig::GetVariable( "physics", "game_accel", 1));
(gdb) bt
#0  0x0000555555b77566 in Movable::Movable (this=0x7fffffffc1b0)
    at /home/ermo/repos/Vega-Strike-Engine-Source/engine/src/cmd/movable.cpp:74
#1  0x0000555555b25064 in Unit::Unit (this=0x555557842710)
    at /home/ermo/repos/Vega-Strike-Engine-Source/engine/src/cmd/unit_generic.cpp:391
#2  0x0000555555a2442c in GameUnit::GameUnit (this=0x555557842710)
    at /home/ermo/repos/Vega-Strike-Engine-Source/engine/src/cmd/unit.cpp:85
#3  0x0000555555733fa2 in main (argc=2, argv=0x7fffffffe658)
    at /home/ermo/repos/Vega-Strike-Engine-Source/engine/src/main.cpp:445
(gdb)
heading to bed
Feels like bringing up pwcu on 0.8.x is going to be a bit of a bear xD
Hm. Could it be a divide-by-zero?
Rune Morling
@ermo
game_speed is 0.2 and game_accel is 1 in vegastrike.config for pwcu.
insys_jump_any_second_ahead doesn't exist, but should default to 4 unless I'm misreading things?
Benjamen Meyer
@BenjamenMeyer
@ermo let's file a bug :)
Rune Morling
@ermo
well, still investigating...
Benjamen Meyer
@BenjamenMeyer
@ermo I can reproduce with Py3...trying Py2 to see if it's a regression in moving from Py2 to Py3
waiting on the build
@ermo filed vegastrike/Vega-Strike-Engine-Source#474 to help track on the VS side
Rune Morling
@ermo
Looking at the GCC manual, there's a difference between -fpie and -fPIE?
my compiler complained that the default (-fpie) wasn't enough and that I needed to supply -fPIE (which I duly did and it compiled)
@BenjamenMeyer fair warning that the PWCU resources are py3-only
Is the goal to see if using a py2 engine will make it further along in the loading of PWCU assets?
Benjamen Meyer
@BenjamenMeyer
apparently complimentary/versioning:
-fpie and -fPIE both define the macros "__pie__" and "__PIE__".  The macros have the value 1 for -fpie and 2 for -fPIE.
from man gcc
Rune Morling
@ermo
right, heading to bed for real now
Stephen G Tuggy
@stephengtuggy

@Loki1950 @nabaco @stephengtuggy @royfalk I updated https://github.com/vegastrike/Vega-Strike-Engine-Source/wiki/Pull-Request-Validation with some additional parameters around testing - mostly I've noticed a lot of my own PRs have had to deviate because of not being things that the play test would cover. The intent is that such cases need to be documented in the PR as an alternative to the play test and improve our PR testability. Please review the updated wiki for feedback/rewording/etc. All feed back welcome.

I think it's reasonable. I admit it's very hard to test certain things using play-testing and the lack of unit tests is hurting us too.

Looks good to me. Just a couple of typos. forms is misspelled as forums in a couple of places. And cover is missing the c: over

Rune Morling
@ermo

Hm. Could it be a divide-by-zero?

@BenjamenMeyer I've updated vegastrike/Vega-Strike-Engine-Source#474 with my findings and a patch that fixes the issue for me.

It does not appear to be a PWCU asset problem per se, more a case of PWCU assets using a default configuration value that exposes an underlying issue/wrong assumption in the VS engine code?
Benjamen Meyer
@BenjamenMeyer
@ermo cool; yeah gives us a place to start; it'd be great to announce a release of PWCU with the 0.8.x release
Rune Morling
@ermo
Release? That's probably ambitious. More like "playable" ;)
(with lots of niggles)
Roy Falk
@royfalk
@BenjamenMeyer a gentle reminder about vegastrike/Vega-Strike-Engine-Source#459 . I reverted and you said you may know an incantation. If you don't have the time or patience, I can reopen PR for another branch and just add a pointer to it (or was this what I was supposed to do and I missed it).
Now I'm not sure.
Benjamen Meyer
@BenjamenMeyer
@royfalk thanks for the reminder
Benjamen Meyer
@BenjamenMeyer
@royfalk I did send you a PM about my initial findings
Benjamen Meyer
@BenjamenMeyer
Just FYI - things are going to be busy around the house for me the next few days; so I'll response on chat and in github (to what I can) but may not be able to do much per code.
Roy Falk
@royfalk
If you don’t mind I’ll try an finish this PR with @stephengtuggy by branching and doing another pr.
Benjamen Meyer
@temporalbeing:matrix.org
[m]
👍