Anonymous | Login | Signup for a new account | 11-10-2008 11:13 PST |
Main | My View | View Issues | Change Log | Docs |
Viewing Issue Simple Details [ Jump to Notes ] | [ View Advanced ] [ Issue History ] [ Print ] | ||||||||
ID | Category | Severity | Reproducibility | Date Submitted | Last Update | ||||
0001279 | [buildroot] Architecture Specific | major | always | 03-20-07 15:19 | 03-22-07 10:47 | ||||
Reporter | Franck78 | View Status | public | ||||||
Assigned To | buildroot | ||||||||
Priority | normal | Resolution | fixed | ||||||
Status | closed | Product Version | |||||||
Summary | 0001279: bad compiler version discover in dependencies.sh | ||||||||
Description |
Hello, I don't know wheter or not $HOSTCC have to be defined before lauching 'make'. Anyway, when not configured, the following test fails: COMPILER=$(which $HOSTCC) 'which' gracefully respond with helps text.... that next $XSED is unable to understand ;) file: toolchain/dependencies/dependencies.sh; line Note: 0000152 Bye |
||||||||
Additional Information | |||||||||
Attached Files | |||||||||
|
Notes | |
(0002259) Franck78 03-20-07 15:37 |
same for $HOSTCXX, next test. Note also 'head' is used before testing presence by the script. |
(0002260) bernhardf 03-21-07 01:07 |
HOSTCC and HOSTCXX are optional. Please post the exact error you are seeing. I don't understand why you point to an unrelated comment? Please elaborate. |
(0002261) Franck78 03-21-07 02:55 |
Hello, Just log all the make programs output to see details of the error. Or ever simpler: go directly in dependencies subir and lauch dependencies.sh. It tests every needed support program presence (except for 'head') and it completly fails the compiler test version WHEN HOSTCC IS NOT SET (then HOSTCXX) Can't be more clear than that, sorry. Asking 'which something' returns pathto something Asking 'which' returns which's syntax usage. Guess what happens when HOSTSCC is empty.... Franck |
(0002262) Franck78 03-21-07 03:11 |
I understand why you don't understand the comment... The link-pointer was a real text entered by me and it is the mantis bug tracking system that replaced it. Note: 0000172=> was "~~ line 172 ~x172 |
(0002263) bernhardf 03-21-07 04:00 |
foo=$(which) does not print anything for me. what version of which do you see this with? i replaced the use of head with sed -e '1d', please svn up and let me know if that's better now. PS: If you don't provide a copy and paste of the error you are getting from dependencies.sh with HOSTCC and HOSTCXX unset, then i'll blissfully ignore this part of your report. Thanks for your understanding. |
(0002264) Franck78 03-21-07 04:36 edited on: 03-21-07 04:44 |
my which prog comes from 'gnu' repository. It does not print anything too, from command line. but inside the script, it is not the same !? Also, another thing changed in my system and now, dependencies.sh finds the rigth version of gcc. Previously, it was an error message with something 'not numeric'. Checking build system dependencies: CC clean: Ok CXX clean: Ok CPP clean: Ok CFLAGS clean: Ok INCLUDES clean: Ok CXXFLAGS clean: Ok LD_LIBRARY_PATH sane: Ok which installed: Ok ./dependencies.sh: line 108: toolchain/dependencies/check-host-sed.sh: Aucun fichier ou r |
(0002265) bernhardf 03-21-07 11:16 |
Please update. Better now? |
(0002269) Franck78 03-22-07 09:22 |
ok, no more junk in the log file and no more error about the first numeric error also (the one that returned bad version for compiler). The ">dev/null" doesn't explain why ;-) |
Copyright © 2000 - 2006 Mantis Group |