|
|
| Author |
Thread Statistics | Show CCP posts - 1 post(s) |
Xianthar
STK Scientific
The Initiative.
   |
Posted - 2009.08.10 16:41:00 -
[1]
Edited by: Xianthar on 10/08/2009 16:47:42
looks good, been waiting for a python version of EFT and got part way through one myself before getting too buisy, good work m8
however, there are a couple problems on the linux side:
1) any python script that should be executable should have #!/usr/bin/python as the first line.
2) i get an error from sqlite on launch:
xianthax@xian:~/Desktop/pyfa-0.1a$ python2.6 downloadData.py
All files up to date!
xianthax@xian:~/Desktop/pyfa-0.1a$ python2.6 launch.py
Traceback (most recent call last):
File "launch.py", line 21, in <module>
window = mainWindow.mainWindow()
File "/home/xianthax/Desktop/pyfa-0.1a/gui/mainWindow.py", line 61, in __init__
self.initGroupBrowser()
File "/home/xianthax/Desktop/pyfa-0.1a/gui/mainWindow.py", line 203, in initGroupBrowser
mGroupMods = marketGroup.getMarketGroup(name = "Ship Equipment", buildParent = False, buildChildren = True, deep = True)
File "/home/xianthax/Desktop/pyfa-0.1a/model/marketGroup.py", line 106, in getMarketGroup
c.execute(getMarketGroupByName, (name,))
sqlite3.OperationalError: no such table: invMarketGroups
looking in ~/home/ i have a .pyfa/ directory which contains a file named eve.db which is 0 bytes. when downloadData.py runs it completes almost instantly, clearly not downloading anything. A quick look at http://pyfa.sourceforge.net/files.ini in a browsers shows the proper text so i'm not sure where its breaking down.
EDIT: found your problem, the destination value in files.ini is "./" in downloadData.py on line 25 you call:
archive.extractall(dataFolder + destination)
as a result i also have a ~/.pyfa./ directory with the database file/icons, removing the extra . at the end fixes the issue and allows the program to run.
cheers
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.10 17:51:00 -
[2]
 Originally by: Sakari Orisi
Edited by: Sakari Orisi on 10/08/2009 17:35:48
I've fixed the pathing issues and updated the 0.1b stages. It should be updated on all sourceforge mirrors within a few hours. (in the meantime you can force it to the master mirror)
I'll fix the top lines of downloadData.py and launch.py shortly as well
EDIT: it also fixes a bug where fits & characters were saved in ~/.pyfafits and ~/.pyfacharacters instead of ~/.pyfa/fits and ~/.pyfa/characters respectivly, you'll have to move around the files when updating from a stage without that fix.
sweet i like double clicking icons :p
i'd also recommend you cross post this to the linux forum, i think you'll get a lot more peoples attention for testing / development over there.
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.10 17:59:00 -
[3]
Originally by: Sakari Orisi
Edited by: Sakari Orisi on 10/08/2009 17:35:48
EDIT: it also fixes a bug where fits & characters were saved in ~/.pyfafits and ~/.pyfacharacters instead of ~/.pyfa/fits and ~/.pyfa/characters respectivly, you'll have to move around the files when updating from a stage without that fix.
along this same line files.ini is currently stored as ~/.pyfafiles.ini instead of ~/.pyfa/files.ini
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.10 20:45:00 -
[4]
Originally by: Sakari Orisi
Edit: isn't cross-posting prohibited on the forums ?
technically yes, but the linux board seems to get more leeway in that regard due to specific nature of its content. Worst case a mod locks it, the additional attention i think you'll get is worth it 
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.10 21:23:00 -
[5]
using a snapshot of master git tree:
files.ini, character.ini, fits.ini properly appear in ~./pyfa as does the eve.db and icons/ directory.
the bash execution command at the top of downloadData.py is #!/usr/bin/python26 on ubuntu at least this needs to be #!/usr/bin/python2.6 as it is in launch.py. also on ubuntu 9.04 /usr/bin/python points to /usr/bin/python2.6 not sure which is safer across different distros.
also a show stopper bug:
setup a character using api import, skills seem to come in correctly.
creating a new fitting causes this error:
Traceback (most recent call last):
File "/home/xianthax/Desktop/pyfa/gui/mainWindow.py", line 523, in fittingChanged
fit.calculateModifiedAttributes()
File "/home/xianthax/Desktop/pyfa/model/fitting.py", line 393, in calculateModifiedAttributes
effect.func(effect, self, level = level)
File "/home/xianthax/Desktop/pyfa/model/effects/standardMissilesSkillBoostMissileVelocityBonus.py", line 6, in standardMissilesSkillBoostMissileVelocityBonus
self.item, extraMult = level)
File "/home/xianthax/Desktop/pyfa/model/effects/customEffects.py", line 97, in boostAmmoListByReq
for module, state, ammo in modList:
TypeError: 'item' object is not iterable
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.10 22:05:00 -
[6]
Originally by: Kurt Meyer
#!/usr/bin/env python
Isn't this the standard?
it certainly does take care of the issue of locating the python executable as this will search $PATH for 'python'.
The problem I was alluding to is that 'python' is nearly always a sym-link to some version of python be it python2.4, python2.6, python3.0 etc. i'm not really sure all distro's name these executables the same way and if you require a specific version you could get in trouble using either python or python2.6 in some situations.
When i've run into similar problems with java apps i've usually broken down and created either a bash script to check the VM version or on windows used JSmooth to create a wrapper exe(barf) to do all the checking.
One of the fun bits of working with interpreted languages.
|
Xianthar
STK Scientific
The Initiative.
|
Posted - 2009.08.11 19:30:00 -
[7]
Originally by: Sakari Orisi
Fixed shield boost amps and the missile projection skills (I don't have any missile skills so those were kinda poorly tested heh)
cheers, allowing me to play more now :) I also openned a couple of new tickets.
i have a general question also, how thread/process safe is the access to the fits/character files? I'm wondering the possibility of running 2 or more instances at the same time to be able to compare multiple setups or ships side by side. right now it appears that writes to disk only occur on application exit but this may be the OS caching in effect.
|
|
|
| |
|