summary refs log tree commit diff stats
path: root/results/classifier/118/none/1826175
diff options
context:
space:
mode:
Diffstat (limited to 'results/classifier/118/none/1826175')
-rw-r--r--results/classifier/118/none/182617580
1 files changed, 80 insertions, 0 deletions
diff --git a/results/classifier/118/none/1826175 b/results/classifier/118/none/1826175
new file mode 100644
index 00000000..ba75276f
--- /dev/null
+++ b/results/classifier/118/none/1826175
@@ -0,0 +1,80 @@
+graphic: 0.649
+semantic: 0.615
+permissions: 0.560
+mistranslation: 0.533
+ppc: 0.459
+device: 0.455
+performance: 0.422
+socket: 0.409
+network: 0.383
+user-level: 0.376
+vnc: 0.361
+VMM: 0.355
+hypervisor: 0.354
+assembly: 0.352
+kernel: 0.348
+arm: 0.332
+peripherals: 0.317
+PID: 0.315
+risc-v: 0.309
+architecture: 0.297
+register: 0.295
+files: 0.281
+virtual: 0.259
+TCG: 0.218
+KVM: 0.215
+debug: 0.211
+boot: 0.202
+i386: 0.167
+x86: 0.106
+
+Compilation on MSYS2/MinGW-w64 fails with error: "No rule to make target capstone.lib"
+
+I submitted this bug to Capstone directly but I figured it'd be useful to post it here too. The IS_MINGW check in the Makefile for Capstone fails under MSYS2 MinGW-w64 because cc --version doesn't have mingw in the output anymore:
+
+$ whereis cc
+cc: /mingw64/bin/cc.exe
+
+$ cc --version
+cc.exe (Rev2, Built by MSYS2 project) 8.3.0
+Copyright (C) 2018 Free Software Foundation, Inc.
+This is free software; see the source for copying conditions.  There is NO
+warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+
+Really simple patch:
+
+diff --git "a/Makefile" "b/Makefile"
+index 063f50db..1d9f042e 100644
+--- "a/Makefile"
++++ "b/Makefile"
+@@ -288,7 +288,7 @@ CFLAGS := $(CFLAGS:-fPIC=)
+ # On Windows we need the shared library to be executable
+ else
+ # mingw?
+-IS_MINGW := $(shell $(CC) --version | grep -i mingw | wc -l)
++IS_MINGW := $(shell $(CC) --version | grep -i msys2 | wc -l)
+ ifeq ($(IS_MINGW),1)
+ EXT = dll
+ AR_EXT = lib
+
+NB that patch will break Linux based Mingw builds, as they don't include "mys2" in their output from gcc, so it would need to match both.
+
+
+
+This change needs to go in via the capstone project. capstone is just an unmodified "submodule" in QEMU - we do not include any QEMU-only changes here.
+
+Re-opened, since although the fix needs to be done in capstone, QEMU also needs a commit to actually update the submodule hash to pull in the fix.
+
+I did submit a bug already to Capstone proper on their GitHub. Thanks for the heads-up about it needing to cover both cases for cross-compilers!
+
+https://github.com/aquynh/capstone/issues/1464 is the upstream bug -- thanks for reporting it there.
+
+
+The fix in upstream capstone has been merged: https://github.com/aquynh/capstone/commit/29893c63e34ee21846744d02c396ae3c801b936b
+
+I am still running into this issue when compiling QEMU on Windows using MSYS2. I can manually apply the fix per the Capstone commit above and compile just fine.
+
+For moving forward, how should this be handled in the codebase to get the MSYS2 build cleanly working? I imagine wholesale pulling the latest Capstone may break quite a bit. Is it possible to cherry-pick this one change into the submodule?
+
+As far as I can see we're using a capstone version now that contains the commit with the fix, so I'm closing this bug. If you are still having problems, please open again or file a new bug ticket.
+