Commits
181122-xiao_gu…
Name already in use
Commits on Dec 4, 2018
-
Allocator performance comparison test
This is a test comparing various alloocator strategies for a thread dispatch package
Commits on Nov 20, 2018
-
Call regfree even if regcomp fails
regcomp set regular expression even if it fails. Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Abstract pattern matching to reduce conditional compilation
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Commits on Nov 19, 2018
Commits on Nov 16, 2018
Commits on Nov 14, 2018
-
Add comment in the _activate functions
Make it clear that the recorder_ring_compare_exchange macro updates the head in place with the new value, because it uses the __atomic_compare_exchange_n primitive, which does, as documented in https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html This compares the contents of *ptr with the contents of *expected. If equal, the operation is a read-modify-write operation that writes desired into *ptr. If they are not equal, the operation is a read and the current contents of *ptr are written into *expected. Signed-off-by: Christophe de Dinechin <christophe@dinechin.org> Suggested-by: Frediano Ziglio <fziglio@redhat.com>
-
Just declared some lines above Signed-off-by: Frediano Ziglio <fziglio@redhat.com> Acked-by: Christophe de Dinechin <christophe@dinechin.org>
-
The temporary string is destroyed after the statement so cname will be a dandling pointer. Signed-off-by: Frediano Ziglio <fziglio@redhat.com> Acked-by: Christophe de Dinechin <christophe@dinechin.org>
-
No reason to export it. Signed-off-by: Frediano Ziglio <fziglio@redhat.com> Acked-by: Christophe de Dinechin <christophe@dinechin.org>
-
Use different option for basename
-s option is already used so currently cannot be used to specify the base name. Signed-off-by: Frediano Ziglio <fziglio@redhat.com> Acked-by: Christophe de Dinechin <christophe@dinechin.org>
Commits on Nov 9, 2018
-
Changes in this release: - License clarification (LGPLv3 for the library components) - Improved portability on Windows platforms (MinGW, MSYS and Cygwin) - Record scope has the ability to toggle display modes using keys - Various minor fixes and improvements - Welcoming a new contributor to the code (Frediano Ziglio)
-
-
-
Style change: lowercase
recordandrecord_fastThis makes these two look more like function calls. The existing macros still work for obvious compatibility reasons.
-
-
Signed-off-by: Christophe de Dinechin <dinechin@redhat.com> Suggested-by: Frediano Ziglio <fziglio@redhat.com>
-
Provides better drand48 replacement
Pretty easy to implement standard one using 64 bit integers. Modified to share MinGW implementation of drand48() between tests Signed-off-by: Frediano Ziglio <fziglio@redhat.com> Signed-off-by: Christophe de Dinechin <christophe@dinechin.org>
-
Source is using HAVE_DRAND48 so we need to include config.h Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Use array_size macro for recorder_info static allocated arrays
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Commits on Nov 8, 2018
-
The old carried-over headers still had remnants of GPL in them. Convert to LGPLv3 for what goes in the library, GPLv3 for the rest. Signed-off-by: Christophe de Dinechin <christophe@dinechin.org> Suggested-by: Frediano Ziglio <fziglio@redhat.com> NOTE: This may not be the final change. If this is to be used in LGPLv2.1 projects like SPICE, the current setup may not be correct, as indicated the license compatibility matrix at https://www.gnu.org/licenses/gpl-faq.html#AllCompatibility. The intent of the authors is to make the code license-compatible, and further licensing changes may be required to that effect.
-
Build a versioned shared library
Signed-off-by: Christophe de Dinechin <christophe@dinechin.org> Suggested-by: Frediano Ziglio <fziglio@redhat.com>
-
Rename old project variable to new name
INCPATH was renamed to INCLUDEPATH in QT4. Avoid a warning in newer qmake. Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
-
Avoid warning compiling recorder_trace_set if regex.h is not present
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Frediano Ziglio committedNov 8, 2018 -
Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Frediano Ziglio committedNov 8, 2018 -
Just declared some lines above Signed-off-by: Frediano Ziglio <fziglio@redhat.com>
Frediano Ziglio committedNov 8, 2018
Commits on Nov 4, 2018
-
Switch libregex to configuration
Some MinGW variants have it, some don't (MSYS2 MinGW64 environmnent)
-
Fix one remaining case of direct reference to regexec
This breaks on some MinGW variants with: recorder.c: In function "recorder_trace_set": recorder.c:2320:29: warning: implicit declaration of function "regexec" [-Wimplicit-function-declaration] if (regexec(&re, rec->name, 1, &rm, 0) == 0 && ^~~~~~~ recorder.c:2320:38: error: "re" undeclared (first use in this function) if (regexec(&re, rec->name, 1, &rm, 0) == 0 && ^~ recorder.c:2320:38: note: each undeclared identifier is reported only once for each function it appears in recorder.c:2320:57: error: "rm" undeclared (first use in this function) if (regexec(&re, rec->name, 1, &rm, 0) == 0 && ^~ recorder.c:2150:21: warning: unused variable "error" [-Wunused-variable] static char error[128]; ^~~~~ At top level: recorder.c:2150:21: warning: "error" defined but not used [-Wunused-variable] recorder.c:952:13: warning: "recorder_shans_file_extend" defined but not used [-Wunused-function] static bool recorder_shans_file_extend(int fd, off_t new_size) -
Add -lregex for versions of MinGW that require it
On MinGW, regex will be in /mingw32/lib/libregex.a. Apparently, it's also possible to use pcre2
-
Use drand48 autoconfiguration instead of CONFIG_MINGW
This is in case MinGW someday fixes their drand48 situation. Also, it happens that I was using drand48, which was not adjusted for.
-
Adjust to versions of MinGW that don't have <sys/mman.h>
I see this appear on an old Win7 machine
Commits on Oct 29, 2018
-
Avoid warning when toggling timing
Without precisely tracking if axis is already attached, you get: "Can not add axis. Axis already on the chart."