stephenbrooks.orgForumMuon1Generalv4.32c
Username: Password:
Search site:
Subscribe to thread via RSS
Stephen Brooks
2003-10-06 03:52:34
Here is the unofficial release of v4.32c before I modify the web page later this evening.  (ZIP file also available).

This may or may not have fixed a bug [OCAU]Badger and others have been having on some systems.  However it does have a better way of checking whether the FTP transfer has worked: it actually uploads and downloads a test file, then only deletes the sent results if the test file has come back intact.  Also you'll see that muon1_background has turned from a .BAT back into an .EXE stub, which runs Muon1 under windows with no console window involved.  The Win98 users who were having problems with it hiding should have no trouble now.

It doesn't make any sense: that's why they call it "virtual"
ZeonX[OCAU]
2003-10-06 04:54:31
Hey stephen, i tried to send some results but the ftp seemed to do the same as what it used to do, doesn't seem to do anything different.
It was a bogus results file so it didn't have anything in it but i wanted to see the new ftp function as i thought it might cause some trouble with the way some of the ftps permissions work.
ZeonX[OCAU]
2003-10-06 06:35:23
Ok stephen i've found a few problems.

1. I let some results build up and sent them but the manual send uploads it to the v4.32b dir not v4.32c

2. You forgot to update the version var in muon because it thinks its out of date as you have put the new results.dat in the ftps.

Apart from those two things i havn't noticed any difference from v4.32b.
Stephen Brooks
2003-10-06 07:27:24
Oh, that was just that I hadn't recompiled manualsend.exe since the last version.  I've done so now and the v4.32c manualsend is now packaged in those archives.

Maybe I ought to make manualsend.exe just be a stub like muon1_background that runs off the main module, then this won't happen again.  Part of the reason I unified the _cmdline and _background programs was of this compile-many-times for a release problem.

It doesn't make any sense: that's why they call it "virtual"
Henri Valta
2003-10-06 08:18:05
quote:
From the ftp set up page:
// Permissions

To protect other users' results, the Muon1 user should only be able to write files to the version directories.  Giving that user read access in these places would allow people to copy others' results and cheat. 


Does the new upload+download check work with these permissions, or do they need to be changed?
Stephen Brooks
2003-10-06 08:59:03
The upload/download test takes place in the root directory of the server where the signal.dat file is, and that has to be downloadable by muon users too, so it doesn't cause problems.

It doesn't make any sense: that's why they call it "virtual"
Henri Valta
2003-10-06 09:20:16
quote:
Originally posted by Stephen Brooks:
The upload/download test takes place in the root directory of the server where the signal.dat file is, and that has to be downloadable by muon users too, so it doesn't cause problems.



Well my ftp only had 'create directory' permission for users on root.  I'll add the write file permission there too.
AySz88
2003-10-06 10:07:39
Hey, Stephen.  I've been having the commandline window proplem here when running muon1 (with or without switches).  The strange thing is, I seem to be the only one with this problem that has Windows XP (unless I have a different problem than the one you mentioned in the first post).
Stephen Brooks
2003-10-06 12:36:21
You're getting the problem I had: if you start muon1 as "muon1 -b" from a commandline prompt, the prompt window won't go away because it wasn't muon1 that created it in the first place.  The new version with muon1_background.exe calls muon.exe with a -b switch "silently" via the Windows function WinExec, so no console appears.

It doesn't make any sense: that's why they call it "virtual"
[OCAU] badger
2003-10-06 16:39:09
Stephen - still getting the problem - instant crash with any type of start, here is the dump:
MUON1 executed an invalid instruction in
module MUON1.EXE at 0167:0040f38a.
Registers:
EAX=00000006 CS=0167 EIP=0040f38a EFLGS=00010246
EBX=0000000b SS=016f ESP=0068f7f8 EBP=0068f808
ECX=005603d4 DS=016f ESI=005603d4 FS=38ef
EDX=005603db ES=016f EDI=00000000 GS=0000
Bytes at CS:EIP:
0f 44 7d fc c6 04 3e 2e ff 75 0c 8d 44 37 01 50
Stack dump:
817139ea 8171dd88 00580000 00000006 0068f82c 00420610 0045018e 00450df0 817139ea 8171dd88 00580000 00450b59 00000000 0068fdb8 00425b53 0045018e

got a debug for 4.32c?

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
thormodr
2003-10-06 18:05:03
I'm the same as Badger...
Stephen Brooks
2003-10-07 02:55:06
This file has a lot of debugging in it.  It should actually spit out at least 2 files when you run it and it crashes.

It doesn't make any sense: that's why they call it "virtual"
thormodr
2003-10-07 03:09:55
Allocated.log says:

005704B4 allocated for c:\program files\lcc\include\rnd.c(312)
005704B4 allocated for d:\sjb39\muon1\textio.c(62)


debug.txt says:

----- setextension(00465B6E:config,00466828:txt) -----
ret=005704B4:config i=6
n=0 i=6

The dialog box says:
lcc runtime: GP fault.  Stack trace
0 setextension [9] d:\sjb39\muon1\muon1.c 67
1 readconfigfile [3] d:\sjb39\muon1\muon1.c 10
2 SBCLIB_main2 [6] d:\sjb39\muon1\muon1.c 58
3 WinMain [0] d:\sjb39\muon1\muon1.c 52

Current instruction: 0x4184f8
[OCAU] badger
2003-10-07 22:41:20
what he said...

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
thormodr
2003-10-11 18:21:11
Any progress on fixing the bug?  Or is it a too much work to fix a bug for too few?  Big Grin
Stephen Brooks
2003-10-12 06:47:32
The main problem is that the bug appears to happen for no particular reason - could be a compiler oddity.  If you re-download the v4.32c distribution, you'll find I've recompiled it to do that particular routine in an equivalent but slightly different way.  I have no way of telling if this will help the problem, though.

It doesn't make any sense: that's why they call it "virtual"
[OCAU] badger
2003-10-14 16:32:45
ok I downloaded 4.32c again, muon.exe from 9/10/03 9:19:34 am.
looks like no joy...


MUON1 executed an invalid instruction in
module MUON1.EXE at 0167:00420775.
Registers:
EAX=00000001 CS=0167 EIP=00420775 EFLGS=00010212
EBX=00000000 SS=016f ESP=0068f7f8 EBP=0068f80c
ECX=00560527 DS=016f ESI=00000000 FS=1247
EDX=7fc1d7e8 ES=016f EDI=7fc1d7e8 GS=0000
Bytes at CS:EIP:
df e9 dd d8 75 07 c6 05 4c fe 44 00 01 68 30 b0
Stack dump:
8170cc74 81714b84 00580000 00450b8e 6200002d 0068fdb8 00425b8e 004501ae 8170cc44 8170cc74 81714b84 00580000 8171673c bff741f7 bffc9490 bff8787a

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
: contact : - - -
E-mail: sbstrudel characterstephenbrooks.orgTwitter: stephenjbrooksMastodon: strudel charactersjbstrudel charactermstdn.io

Site has had 17125566 accesses.