April 2013 - Asg - mailman3.cels.anl.gov

[aesop] #38: bracketed single-line if statements
by aesop 11 May '15

11 May '15

#38: bracketed single-line if statements --------------------------------+------------------------------------------- Reporter: gcdanie@… | Type: defect Status: new | Priority: don't know Component: don't know | Version: Keywords: | --------------------------------+------------------------------------------- When compiling a non-blocking function (e.g., without a __blocking identifier), should that function contain a single line if statement enclosed in brackets (as below), the aesop compiler will error (as below). if (condition) { return something; } aatree.c:79:1: error: stray ‘#’ in program aatree.c:79:7: warning: statement with no effect [-Wunused-value] aatree.c:79:10: error: expected ‘;’ before string constant aatree.c:81:31: error: ‘dir’ undeclared (first use in this function) aatree.c:81:31: note: each undeclared identifier is reported only once for each function it appears in aatree.c: In function ‘__ae_worker_treenode_traverse’: -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/38> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

2 3

[aesop] #46: parallelfor2 test failure
by aesop 20 Mar '15

20 Mar '15

#46: parallelfor2 test failure -------------------------+-------------------------------------------------- Reporter: carns | Owner: Type: defect | Status: new Priority: minor | Component: don't know Version: devel branch | Keywords: -------------------------+-------------------------------------------------- The parser/tests/blocking/parallelfor2 test program fails with the following assertion: {{{ parallelfor2: /home/pcarns/working/aesop-devel/build/include/aesop/ae- ctl.h:229: ae_ctl_done: Assertion `refcount >= 0' failed. Aborted (core dumped) }}} -- Ticket URL: <https://trac.mcs.anl.gov/projects/aesop/ticket/46> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 3

[aesop] #34: Initialization + declaration: invalid reorder
by aesop 30 Sep '14

30 Sep '14

#34: Initialization + declaration: invalid reorder ---------------------------+------------------------------------------------ Reporter: dkimpe | Owner: Type: defect | Status: new Priority: major | Component: code generator Version: git repository | Keywords: ---------------------------+------------------------------------------------ In the following (accepted by the compiler): {{{ ... init_a (); int my_var = call_b (); ... }}} my_var gets moved to the context structure (correct); However, it gets initialized at the beginning of the function, effectively moving {{{call_b}}} before {{{init_a}}} in the example. This can be a very hard to track down bug... Possible solution: disallow behavior in source translator or fix. -- Ticket URL: <https://trac.mcs.anl.gov/projects/aesop/ticket/34> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 2

[aesop] #26: OSX warnings during build
by aesop 12 Sep '14

12 Sep '14

#26: OSX warnings during build -------------------------------+-------------------------------------------- Reporter: rross@… | Type: defect Status: new | Priority: don't know Component: don't know | Version: Keywords: | -------------------------------+-------------------------------------------- Lots of warnings when building on my OSX box. -- Rob --- make -C src/c-utils GHC parser/ae-blocking-parser CC opcache.o CC aesop.o DEP lookup3.d DEP triton-string.d CC resource.o DEP libev/event.d In file included from ../resource.c:1: ../src/c-utils/libev/ev.h: In function ‘ev_default_loop_uc_’: ../src/c-utils/libev/ev.h:548: warning: nested extern declaration of ‘ev_default_loop_ptr’ DEP libev/ev.d ../resource.c: In function ‘ae_get_stack’: ../resource.c:738: warning: format not a string literal and no format arguments CC hints.o CC libev/ev.o CC libev/event.o CC triton-string.o CC lookup3.o CC ae-ctl.o CC ae-init.o RANLIB libc-utils.a Linking parser/ae-blocking-parser ... ld: warning: could not create compact unwind for _ffi_call_unix64: does not use RBP or RSP based frame AECC src/aefile/aefile.ae AECC src/socket/aesocket.ae AECC resources/resourcebuilder/resourcebuilder.hae AECC resources/thread/aethread.hae AECC resources/timer/timer.hae CC resources/resourcebuilder/resourcebuilder.o CC resources/thread/aethread.o CC resources/timer/timer.o In file included from ../resources/timer/timer.c:14: ../src/c-utils/libev/ev.h: In function ‘ev_default_loop_uc_’: ../src/c-utils/libev/ev.h:548: warning: nested extern declaration of ‘ev_default_loop_ptr’ /usr/include/assert.h:79: warning: previous declaration of ‘__assert_rtn’ was here /usr/include/assert.h:79: warning: previous declaration of ‘__assert_rtn’ was here ../src/aefile/aefile.hae:23: warning: previous declaration of ‘aefile_pwrite’ was here ../src/aefile/aefile.hae:24: warning: previous declaration of ‘aefile_pread’ was here ../src/aefile/aefile.hae:22: warning: previous declaration of ‘aefile_fsync’ was here ../src/aefile/aefile.hae:23: warning: previous declaration of ‘aefile_fdatasync’ was here ../src/aefile/aefile.ae: In function ‘__ae_worker_aefile_fdatasync’: ../src/aefile/aefile.ae:88: warning: nested extern declaration of ‘fdatasync’ ../src/aefile/aefile.ae: At top level: ../src/aefile/aefile.hae:25: warning: previous declaration of ‘aefile_ftruncate’ was here ../src/aefile/aefile.hae:25: warning: previous declaration of ‘aefile_unlink’ was here /usr/include/assert.h:79: warning: previous declaration of ‘__assert_rtn’ was here /usr/include/assert.h:79: warning: previous declaration of ‘__assert_rtn’ was here /usr/include/string.h:172: warning: previous declaration of ‘swab’ was here /usr/include/stdlib.h:218: warning: previous declaration of ‘mkstemp’ was here /usr/include/stdlib.h:217: warning: previous declaration of ‘mktemp’ was here /usr/include/stdlib.h:235: warning: previous declaration of ‘setkey’ was here /usr/include/stdlib.h:341: warning: previous declaration of ‘valloc’ was here /usr/include/stdlib.h:340: warning: previous declaration of ‘suboptarg’ was here /usr/include/stdlib.h:206: warning: previous declaration of ‘getsubopt’ was here /usr/include/assert.h:79: warning: previous declaration of ‘__assert_rtn’ was here /usr/include/pthread.h:400: warning: previous declaration of ‘pthread_kill’ was here /usr/include/pthread.h:406: warning: previous declaration of ‘pthread_sigmask’ was here /usr/include/unistd.h:710: warning: previous declaration of ‘mknod’ was here ../src/c-utils/libev/ev.h: In function ‘ev_default_loop_uc_’: ../src/c-utils/libev/ev.h:548: warning: nested extern declaration of ‘ev_default_loop_ptr’ ../src/socket/aesocket.ae: At top level: ../src/socket/aesocket.hae:79: warning: previous declaration of ‘aesocket_ready’ was here ../src/socket/aesocket.hae:45: warning: previous declaration of ‘aesocket_accept’ was here ../src/socket/aesocket.hae:56: warning: previous declaration of ‘aesocket_read’ was here ../src/socket/aesocket.hae:67: warning: previous declaration of ‘aesocket_write’ was here RANLIB lib/libaesop.a mcswl174:build rross$ -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/26> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 1

[aesop] #22: overriding CFLAGS in make or configure causes build to fail.
by aesop 12 Sep '14

12 Sep '14

#22: overriding CFLAGS in make or configure causes build to fail. ---------------------------+------------------------------------------------ Reporter: dkimpe | Owner: Type: defect | Status: new Priority: minor | Component: build system Version: git repository | Keywords: ---------------------------+------------------------------------------------ Configuring aesop using CFLAGS="-O2" (i.e. ./configure CFLAGS="-O2") or issuing make CFLAGS="-O2" causes the build to fail. One reason is that it overrides the -I needed to find aesop.h. This most likely is a bug in Makefile.in and/or aecc -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/22> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 2

[aesop] #48: set up automated Aesop tests in Jenkins
by aesop 23 Oct '13

23 Oct '13

#48: set up automated Aesop tests in Jenkins ----------------------------+----------------------------------------------- Reporter: carns | Owner: Type: feature request | Status: new Priority: critical | Component: build system Version: devel branch | Keywords: ----------------------------+----------------------------------------------- We should add Aesop to MCS's Jenkins test environment for automatic regression testing. The aesop tests are sequential and do not require any MPI or job scheduling environment. Testing can focus on the devel branch until it is merged back to trunk. -- Ticket URL: <https://trac.mcs.anl.gov/projects/aesop/ticket/48> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

2 2

[aesop] #20: Provide some debugging helper functions for aesop
by aesop 23 Oct '13

23 Oct '13

#20: Provide some debugging helper functions for aesop ---------------------------+------------------------------------------------ Reporter: dkimpe | Owner: Type: enhancement | Status: new Priority: minor | Component: aesop functions Version: git repository | Keywords: ---------------------------+------------------------------------------------ - would be nice if we had a function to poke at aesop and dump a list of what contexts are running, what function they are in, etc. - the problem right now is if you stop an aesop program in gdb you don't see anything helpful; it is usually in the event loop - could trigger with function call, signal, rpc, or whatever -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/20> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 1

[aesop] #43: Try compiling with clang and use static analyzer
by aesop 29 Jul '13

29 Jul '13

#43: Try compiling with clang and use static analyzer ------------------------+--------------------------------------------------- Reporter: dkimpe | Owner: Type: enhancement | Status: new Priority: minor | Component: build system Version: | Keywords: ------------------------+--------------------------------------------------- {{{ >-Phil > Oh, that reminds me! I haven't tried it in a while, but when we first split > aesop into its own tree and revised the aecc wrapper, one of the side > effects (and test cases) was to compile aesop using clang. I assume this > still works. Once Triton is fully building with the external aesop > implementation, this means (in theory) that we can build all of Triton with > clang as well. > We have other priorities right now, but we should revisit that once we > finish the triton conversion and have the various test programs working. I > think the clang static analyzer will turn up some interesting things to look > at in the code. It found a few things even when I ran it on the aesop tree > by itself, and it is pretty easy to use. }}} -- Ticket URL: <https://trac.mcs.anl.gov/projects/aesop/ticket/43> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 2

Draft schedule for meeting on 14th and 15th May
by Benafsh Husain 30 Apr '13

30 Apr '13

Dear all, I am attaching a draft for the agenda for the meeting at Clemson on 14th and 15th May. Please take a look and let me know if there are any suggestions. Also, we have allotted time for student presentations and breakout sessions. There are about 6 student presentation time slots available which can be increased based on demand. I will be giving one presentation. Could any other interested students please let me know their topic and I can include them on the agenda. Also, we have allocated time for breakout sessions and are open to ideas. I can include popular topics on the agenda. We look forward to seeing everyone at Clemson soon. Thanks, Benafsh

3 3

[aesop] #50: fsync memory leak
by aesop 26 Apr '13

26 Apr '13

#50: fsync memory leak --------------------------------+------------------------------------------- Reporter: gcdanie@… | Type: defect Status: new | Priority: don't know Component: aesop functions | Version: git repository Keywords: | --------------------------------+------------------------------------------- Whenever I call aefile_fsync on a valid file descriptor, the call leaks memory (details below). ==11238== 88 bytes in 1 blocks are possibly lost in loss record 57 of 106 ==11238== at 0x4C244E8: malloc (vg_replace_malloc.c:236) ==11238== by 0x447DAF: ae_opcache_get (opcache.c:329) ==11238== by 0x447C05: aethread_hint (aethread.c:177) ==11238== by 0x444F76: __ae_worker_aefile_fsync (aefile.ae:88) ==11238== by 0x445BB0: aefile_fsync (aefile.ae:84) -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/50> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 1