New Core Issues

ruff_hi

Live 4ever! Or die trying
Joined
Oct 24, 2005
Messages
9,135
Location
an Aussie in Boston
Should I start a new thread for this - sure ... why not.

I swap the new core stuff into a custom asset folder, fire up Civ4 and get an error I haven't seen before while trying to load a save ...

..\WARLORDS\ASSETS\PYTHON\SYSTEM\xmllib.py:9: DeprecationWarning: The xmllib module is obsolete. Use xml.sax instead.

Finally load the save and go to open the log ... CTD.
 
Reload and fiddle with the options - get lots of "Option change from None to False" etc. Hit end of turn and get ...

Code:
..\WARLORDS\ASSETS\PYTHON\SYSTEM\xmllib.py:9: DeprecationWarning: The xmllib module is obsolete.  Use xml.sax instead.
Traceback (most recent call last):

  File "CvEventInterface", line 30, in onEvent

  File "BugEventManager", line 167, in handleEvent

  File "BugEventManager", line 181, in _handleDefaultEvent

  File "UnitNameEventManager", line 256, in onUnitBuilt

  File "UnitNameEventManager", line 398, in getUnitNameConvFromIniFile

AttributeError: 'IniFile' object has no attribute 'getByEraAndClass'
ERR: Python function onEvent failed, module CvEventInterface

Oh - managed to open the log this time without a CTD.
 
..\WARLORDS\ASSETS\PYTHON\SYSTEM\xmllib.py:9: DeprecationWarning: The xmllib module is obsolete. Use xml.sax instead.

This warning is unavoidable yet harmless. I didn't want to try to package up a newer and larger XML parser into BUG given that the one that's in Civ works just fine for our needs.

Finally load the save and go to open the log ... CTD.

Hmm, this is disturbing. NikNaks tried it last night and his machine locked up. He had to hard-reset his PC. I, however, haven't experienced anything like these problems -- just Python exceptions.

Maybe you should tell your PC how much you love it more often. :mischief:
 
I've played a few more hours, and still only one outstanding bug. Unless something major happens, I'll definitely commit the new core this weekend and start focusing on a 3.1 release.
 
I feel like I'm talking to myself, but oh well. Marching on . . .

I've played nearly a full game with no issues now, and I'm pretty confident the new core is stable and ready for prime time. I'm finishing up some last tasks and will commit this week, maybe tomorrow or Wednesday if I'm very productive tonight.

This is a complete rewrite of how BUG initializes the game and accesses options. Anyone that has done their own merges will have some work to do the remerge.

The bright side is that merging should be a lot easier than it was previously, depending as always on the extent and nature of the changes in the mod you're merging. I'll be writing some docs soon, but I don't want to hold off the commit of this new code into the main codebase.

You've been duly warned: curves ahead. ;)
 
I feel like I'm talking to myself, but oh well. Marching on . . .
sorry

This is a complete rewrite of how BUG initializes the game and accesses options. Anyone that has done their own merges will have some work to do the remerge.
should this be a v4.0 release? Is it big enough for a new version number?
 
Maybe a 5.0 release, but not 4.0. I have my reasons, even if you think they're silly.:p

I too played through an almost complete game yesterday, and experienced no issues beside a slight slow down on init. Game play didn't have it. I didn't experience much of a delay selecting a unit stack in cities like you were talking about EF, but I didn't have any large stacks in cities that I selected, so that maybe why.
 
As I said, it wasn't the size of the stack (1 unit is noticeable), it's the fact that it's on a city. Clicking HUGE stacks or SMALL stacks on any tile without a city is instantaneous. So I don't see how it could be BUG.

Also, you didn't say anything about a slowdown on init yesterday. I'm the one that said that init is the only place I made any changes. I think it's psychosomatic because I've been launching saved games constantly during testing and haven't noticed it.

I'll try a non-BUG save just to be sure. In fact, I'll time the init to see how long it takes . . .

Edit: I timed it a few times, and the entire BUG initialization phase (read and parse XML files, create events, create and load options) takes 1/4 second.
 
I've updated BUG to use the new core version from SVN, and I get a bug in the option screen, because only the first tabs are drown; it seems that the problem is with the option for the PLE_Move_Highlighter, because it's the last thing displayed.

Here you can see the PytonErr:


Spoiler :
Code:
..\WARLORDS\ASSETS\PYTHON\SYSTEM\xmllib.py:9: DeprecationWarning: The xmllib module is obsolete.  Use xml.sax instead.
Traceback (most recent call last):

  File "CvEventInterface", line 30, in onEvent

  File "BugEventManager", line 221, in handleEvent

  File "BugEventManager", line 247, in _handleConsumableEvent

  File "BugOptionsEventManager", line 25, in onKbdEvent

  File "CvScreensInterface", line 112, in showBugOptionsScreen

  File "BugOptionsScreen", line 46, in interfaceScreen

  File "BugOptionsScreen", line 52, in createTabs

  File "BugPleOptionsTab", line 71, in create

  File "BugOptionsTab", line 211, in addCheckbox

  File "BugOptionsTab", line 50, in getOption

  File "BugOptions", line 165, in getOption

ConfigError: Missing option: PLE_Move_Highlighter
ERR: Python function onEvent failed, module CvEventInterface
 
it seems that the problem is with the option for the PLE_Move_Highlighter, because it's the last thing displayed.

Oops, merge error. Fixed and committed.
 
I've also noted that the ..\CustomAssets\Python\BUG\Options folder is still there even if it's empty...

Yes, I'm still pondering moving some files into that folder to organize the BUG folder better, and SVN isn't so keen on deleting and then readding the same folder. It will do it, but it can cause problems for people who update sometimes.
 
Yes, I'm still pondering moving some files into that folder to organize the BUG folder better, and SVN isn't so keen on deleting and then readding the same folder. It will do it, but it can cause problems for people who update sometimes.

Ok, no problem for me if you leave it where it is :)
Anyway, just for future knowledge, according to my experience those problems can be avoided adding/removing the folder(s) from the repo-browser, which seems to be the only way to perform basic operations on the file structure without problems.
 
Another issue: Specialist Stacker doesn't work anymore, regardless the selected option, it always uses the default display. No Python error in the log.
 
And in Unit Name, it put a couple of ' ' around the naming convention, breaking it.

That happened to me, but I just realized why it did. The default got mangled when I converted from Python to XML. Fixed.
 
That happened to me, but I just realized why it did. The default got mangled when I converted from Python to XML. Fixed.

Good, now it's ok :)

SpecialistStaker still not working, even deleting all the previous files and installing again from SVN last version.

On a side note, for Scoreboard, IMHO the default code list should be (as it was) the full one, with the same order used in the hover text, that is:
WSZC?EPTUNBDRAH*LO

No other iissue in 5 hours of play ^_^
 
Another issue: Specialist Stacker doesn't work anymore, regardless the selected option, it always uses the default display. No Python error in the log.

Fixed. Nice catch. This also fixes similar problems with other dropdown list options (PLE).

BTW, thank you for testing so thoroughly. :goodjob:
 
On a side note, for Scoreboard, IMHO the default code list should be (as it was) the full one, with the same order used in the hover text, that is:
WSZC?EPTUNBDRAH*LO

Done. 789A
 
Back
Top Bottom