Hi Edmund, first thanks for your help and the patch, and sorry for the late answer, but I spent all my time in trying to fix the gcc-5 compilation issue
I'm building right now a version with your patch, if everything goes smoothly I'll upload on experimental/unstable shortly, together with some cdecl fixes, to see how the builds on other arm platforms go. thanks, Gianfranco Il Sabato 30 Maggio 2015 12:06, Edmund Grimley Evans <[email protected]> ha scritto: To start off with, delete lines 106-109 and 126-129 in this file: https://sources.debian.net/src/casablanca/2.4.0-2/Release/src/utilities/base64.cpp/ They don't do anything useful on any architecture, as far as I can tell, and they cause a warning, treated as an error, when plain char is unsigned. With that change it got as far as running the tests on arm64, though some of them failed: The following tests FAILED: 1 - httpclient_test (Failed) 2 - httplistener_test (Failed) 3 - json_test (OTHER_FAULT) 5 - streams_test (Failed) 8 - websocketclient_test (Failed) It's possible some of those would pass in a different environment from the slightly weird one I was using, though I note that some of them also failed on hurd-i386 and kfreebsd-amd64, so probably not. I'm building right now a version with your patch, if everything goes smoothly I'll upload on experimental shortly http://debomatic-amd64.debian.net/distribution#experimental/casablanca/2.5.0-2/buildlog http://debomatic-armel.debian.net/distribution#experimental/casablanca/2.5.0-2/buildlog http://debomatic-armhf.debian.net/distribution#experimental/casablanca/2.5.0-2/buildlog http://debomatic-arm64.debian.net/distribution#experimental/casablanca/2.5.0-2/buildlog Il Sabato 30 Maggio 2015 12:06, Edmund Grimley Evans <[email protected]> ha scritto: To start off with, delete lines 106-109 and 126-129 in this file: https://sources.debian.net/src/casablanca/2.4.0-2/Release/src/utilities/base64.cpp/ They don't do anything useful on any architecture, as far as I can tell, and they cause a warning, treated as an error, when plain char is unsigned. With that change it got as far as running the tests on arm64, though some of them failed: The following tests FAILED: 1 - httpclient_test (Failed) 2 - httplistener_test (Failed) 3 - json_test (OTHER_FAULT) 5 - streams_test (Failed) 8 - websocketclient_test (Failed) It's possible some of those would pass in a different environment from the slightly weird one I was using, though I note that some of them also failed on hurd-i386 and kfreebsd-amd64, so probably not. -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
