2009/5/12 <[email protected]>: > > Thanks again Erik. I tried removing it and it now crashes in src/ > objects-inl.h on line 1332. > > # > # Fatal error in src/objects-inl.h, line 1332 > # CHECK(descriptor_number < number_of_descriptors()) failed > #
That's reasonably strange. A stack trace certainly would be helpful. If you don't have gdb, but you do have glibc then perhaps you can use the backtrace function (as used in OS::StackWalk in platform-linux.cc) to hack up a quick stack-trace printer in V8_Fatal in checks.cc. It already tries to print a JS stack trace, but it doesn't attempt to print a C/C+ stack trace. > > I haven't been able to see what the values of these are yet, but I am > trying to track that down. > > On May 12, 1:58 pm, Erik Corry <[email protected]> wrote: >> Oops the assertion is wrong. Just delete the assertion. I'll fix it ASAP. >> >> Hopefully a more significant assertion will turn up if you delete that one. >> >> 2009/5/12 <[email protected]>: >> >> >> >> >> >> > Thanks Erik. >> >> > I did manage to continue tracking down the issue last night and it >> > turns out that there is an assertion occurring in assembler-asm.cc - >> > line 612, which is in the function void Assembler::addrmod5(...) >> > ----> ASSERT((instr & ~(CondMask | P | U | N | W | L)) == (B27 | >> > B26)); >> >> > Unfortunately, I am having trouble getting gdb to work properly on my >> > hardware platform, so I am trying some alternative debugging >> > methods. >> >> > Any ideas on the assertion? >> >> > Regards, >> > John >> >> > On May 11, 4:35 pm, Erik Corry <[email protected]> wrote: >> >> 2009/5/11 <[email protected]>: >> >> >> > Hi, >> >> >> > I have recently compiled the trunk version of the v8 javascript engine >> >> > for the ARM processor under Linux and have run into an issue when >> >> > executing the sample "shell" application from the command line. >> >> >> > I get the following output: >> >> > "native runtime.js:612: <error> >> >> >> > Fatal error in v8::Context::New() >> >> > Could not initialize environment >> >> >> Not a lot to go on here. I would recommend you compile with >> >> mode=debug on the scons command line to get all the slow assert code. >> >> It sucks for performance but it's great for debugging. If you get a >> >> failed assert, then run in gdb and use bt to see where it is >> >> happening. >> >> >> > " >> >> > ./shell --help works fine. >> >> >> > The target platform is an ARM11 with the armv6 instruction set. Linux >> >> > kernel is 2.6.14. gcc toolchain is 3.4.0. >> >> >> > To compile I did the following: >> >> > $ scons toolchain=gcc arch=arm library=static os=linux importenv=PATH >> >> > sample=shell >> >> > where: >> >> > CC="arm-linux-gcc" >> >> > CXX="arm-linux-g++" >> >> > RANLIB="arm-linux-ranlib" >> >> > AR="arm-linux-ar" >> >> >> > Thanks for your time and I look forward to hearing any ideas on how to >> >> > resolve this issue. >> >> >> > Regards, >> >> > John >> >> >> -- >> >> Erik Corry, Software Engineer >> >> Google Denmark ApS. CVR nr. 28 86 69 84 >> >> c/o Philip & Partners, 7 Vognmagergade, P.O. Box 2227, DK-1018 >> >> Copenhagen K, Denmark. >> >> -- >> Erik Corry, Software Engineer >> Google Denmark ApS. CVR nr. 28 86 69 84 >> c/o Philip & Partners, 7 Vognmagergade, P.O. Box 2227, DK-1018 >> Copenhagen K, Denmark. > > > -- Erik Corry, Software Engineer Google Denmark ApS. CVR nr. 28 86 69 84 c/o Philip & Partners, 7 Vognmagergade, P.O. Box 2227, DK-1018 Copenhagen K, Denmark. --~--~---------~--~----~------------~-------~--~----~ v8-users mailing list [email protected] http://groups.google.com/group/v8-users -~----------~----~----~----~------~----~------~--~---
