Strange! I tried sever times running with -c fetch. It always produced wrong and same checksum. Then I edited recipe for those wrong checksums. Then do_unpack crashed - not in gzip fomat. This same could be verified manually,it did not open manually, too.
I loaded that file manually with browser. It had correct md5 sum. I placed it in downloads and voila, recipe went through fine. How does this happen? I have faced lot's of these checksum problems - especially on angstrom_2010.x branch. Therefore I try t cope with 2008.1. Could e.g. my http fetcher be broken somehow? -Matti 2011/7/24 matti kaasinen <matti.kaasi...@gmail.com>: > Hi, > I tried to build openjdk-6 with following > Build Configuration: > BB_VERSION = "1.12.0" > METADATA_BRANCH = "2011.03-maintenance" > METADATA_REVISION = "a1f31c7" > TARGET_ARCH = "arm" > TARGET_OS = "linux-gnueabi" > MACHINE = "at91sam9260ek" > DISTRO = "angstrom" > DISTRO_VERSION = "2011.03" > TARGET_FPU = "soft" > > Build crashes when fetching: > http://www.apache.org/dist/excalibur/avalon-framework/source/avalon-framework-api-4.3-src.tar.gz > MD5: expected "d4cffb4ba1d07bdc517ac6e322636495", got > "1406ff3b7b035d3da7b54fb6ae002c99" > SHA256: expected > "a4d56a053609ddfc77f6a42c3f15a11708d5e0eb29ffc60a40b87e4cc7331d47", > got "0566f3786a8ed3d34e9a310caa0f454fd4d32609b3b08c97aaa14d768d6ce907" > > Do anyone know if this file has really (logically) changed or would it > be OK just to change checksums from recipe? > > BTW I got much more of these problems when I tried to build > console-image with master branches. What's happened with Ångström > distribution? > As a rule: would it be ok just fix these checksums. I don't have > "original" revisions to compare with? > > -Matti > _______________________________________________ Angstrom-distro-users mailing list Angstrom-distro-users@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-users
