[M3devel] report on Windows XP builds/tests
Tony Hosking
hosking at cs.purdue.edu
Wed Sep 2 02:29:59 CEST 2009
Possibly something very broken in mutexes?
On 1 Sep 2009, at 20:09, Randy Coleburn wrote:
> Hi, I am back from vacation.
>
> I've updated my sandbox on WindowsXP to be current with the CVS head.
>
> I tried first to run Jay's "upgrade.py", but got the following error:
>
> C:\cm3\Sandbox\cm3\scripts\python>upgrade.py
> using c:\cm3\bin\cm3.exe
> PATH=c:\cm3\bin;%PATH%
> set CM3_TARGET=NT386
> set CM3_INSTALL=c:\cm3
> set M3CONFIG=C:\cm3\Sandbox\cm3\m3-sys\cminstall\src\config-no-
> install\NT386
> set CM3_ROOT=C:/cm3/Sandbox/cm3
> ERROR: errno.h not found in %INCLUDE%(C:\msdev\80\VC\include;C:
> \Program Files\Microsoft SDKs\Windows\v6.0A\Include;C:\Program Files
> \Microsoft Platform SDK for Windows Server 2003 R2\Include)
> C:\cm3\Sandbox\cm3\scripts\python>
>
> Next I tried my upgrade script and was successful. Then I used my
> "do-cm3.cmd" script to rebuild all packages and was successful,
> except for "m3-sys\m3cc" and "caltech-parser\parserlib\parserlib
> \test".
>
> In catching up on my email, I noticed Olaf had asked about mentor
> and juno on Windows, so I tried running these. Juno starts up and
> puts up a window, but quickly crashes. mentor crashes before any
> window comes up and I also get a firewall request from Windows
> asking whether to block the program--I suspect it is trying to
> access the network, hence the firewall request. Here are the
> runtime error reports:
>
> C:\cm3\bin>mentor
>
> ***
> *** runtime error:
> *** Attempt to reference an illegal memory location.
> *** pc = 0x849590 = LockMutex + 0x9c in ..\src\thread
> \WIN32\ThreadWin32.m3
> ***
>
> Stack trace:
> FP PC Procedure
> --------- --------- -------------------------------
> 0x2abf97c 0x846fdd SystemError + 0x66 in ..\src\runtime
> \NT386\RTSignal.m3
> 0x2abf9b8 0x849590 LockMutex + 0x9c in ..\src\thread
> \WIN32\ThreadWin32.m3
> 0x2abf9e8 0x1405c72 Be + 0x3e in ..\src\split\TextVBT.m3
> 0x2abfb24 0x1235764 pText + 0x6f4 in ..\src\FormsVBT.m3
> 0x2abfb84 0x12234b1 Item + 0x48b in ..\src\FormsVBT.m3
> 0x2abfbb0 0x124107a OneChild + 0xf8 in ..\src\FormsVBT.m3
> 0x2abfc90 0x1227c9a pMButton + 0x11c in ..\src\FormsVBT.m3
> 0x2abfcf0 0x12234b1 Item + 0x48b in ..\src\FormsVBT.m3
> 0x2abfd1c 0x124107a OneChild + 0xf8 in ..\src\FormsVBT.m3
> 0x2abfe24 0x1227520 pShape + 0x19c in ..\src\FormsVBT.m3
> ......... ......... ... more frames ...
>
>
> C:\cm3\bin>juno
>
> ***
> *** runtime error:
> *** <*ASSERT*> failed.
> *** file "..\src\runtime\common\RTCollector.m3", line 2284
> ***
>
> Stack trace:
> FP PC Procedure
> --------- --------- -------------------------------
> 0x210f9d8 0x5d4e43 CheckStoreTraced + 0xde in ..\src\runtime
> \common\RTCollector.m3
> 0x210fa38 0x10041c74 Unnest + 0x124 in ..\src\JunoCompileNF.m3
> 0x210fac8 0x10040a49 B1 + 0x3d0 in ..\src\JunoCompileNF.m3
> 0x210faf4 0x1003f9b6 Normalize + 0x46 in ..\src\JunoCompileNF.m3
> 0x210fb20 0x10016b9f Cmd.C2pp + 0x20 in ..\src\JunoCompile.m3
> 0x210fb70 0x10016ac3 Cmd.C2p + 0x618 in ..\src\JunoCompile.m3
> 0x210fbcc 0x1001594b Cmd.C2 + 0x9c8 in ..\src\JunoCompile.m3
> 0x210fc24 0x100156b8 Cmd.C2 + 0x735 in ..\src\JunoCompile.m3
> 0x210fc7c 0x100151a0 Cmd.C2 + 0x21d in ..\src\JunoCompile.m3
> 0x210fcc8 0x10016b43 Cmd.C2p + 0x698 in ..\src\JunoCompile.m3
> ......... ......... ... more frames ...
>
> C:\cm3\bin>
>
> Juno crashes on an ASSERT in the collector, while mentor seems to be
> trying to lock a mutex that isn't properly initialized. I have not
> looked at the source code to try and debug. Let me know if you want
> me to pursue further.
>
> I've rebuilt some of my own programs and run some very basic tests.
> So far, no problems detected.
>
> I will try to run some tests on Vista platform tomorrow.
>
> Maybe it would be prudent for me to compile and run some of the
> tests ya'll have been implementing. I think the scripts for these
> aren't native Windows, but if you can point me to some of these,
> I'll try to translate for use on Windows.
>
> Let me know how I can best assist with the release effort.
>
> Regards,
> Randy Coleburn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20090901/3e327c63/attachment-0002.html>
More information about the M3devel
mailing list