error-report: Add info_report_once helper
PATCH 1/1 error-report: Add info_report_once helper
We already have error_report_once and warn_report_once, thus lets add info_report_once to complement. Actually I use this helper a lot so might be usefull for others.
--- include/qemu/error-report.h | 13 +++++++++++++ util/qemu-error.c | 20 ++++++++++++++++++++ 2 files changed, 33 insertions(+)
include/qemu/error-report.h +13 -0
@@ -47,6 +47,8 @@ bool error_report_once_cond(bool *printeGCC_FMT_ATTR (2 ,3 ); bool warn_report_once_cond(bool *printed, const char *fmt, ...) GCC_FMT_ATTR(2, 3); +bool info_report_once_cond(bool *printed, const char *fmt, ...) + GCC_FMT_ATTR(2, 3);void error_init (const char *argv0);
@@ -72,6 +74,17 @@ void error_init(const char *argv0); fmt, ##__VA_ARGS__); \ }) +/* +* Similar to info_report(), except it prints the message just once. +* Return true when it prints, false otherwise. +*/ +#define info_report_once (fmt, ...) \ + ({ \ + static bool print_once_; \ + info_report_once_cond(&print_once_, \ + fmt, ##__VA_ARGS__); \ + }) +const char *error_get_progname (void );extern bool enable_timestamp_msg;
util/qemu-error.c +20 -0
@@ -350,6 +350,26 @@ bool warn_report_once_cond(bool *printedreturn true; } +/* +* Like info_report(), except print just once. +* If *printed is false, print the message, and flip *printed to true. +* Return whether the message was printed. +*/ +bool info_report_once_cond (bool *printed,const char *fmt, ...) +{ +va_list ap; + +assert (printed); +if (*printed) { +return false; + } + *printed = true; +va_start (ap, fmt); +vreport (REPORT_TYPE_INFO , fmt, ap); +va_end (ap); +return true; +} +static char *qemu_glog_domains; static void qemu_log_func(const gchar *log_domain,
You neglected to cc: the maintainer (me). I spotted this anyway, more or less by chance. Next time, please use scripts/get_maintainer.pl to find people you might want to cc:.
Quoted 5 lines · Cyrill Gorcunov
We already have
error_report_onceandwarn_report_once, thus lets addinfo_report_onceto complement. Actually I use this helper a lot so might be usefull for others.Signed-off-by Cyrill Gorcunov <[email protected]>
Patch looks reasonable enough. However, the new info_report_once() and info_report_once_cond() aren't used. Your commit message suggests you have users in unpublished code. Is that correct?
Quoted 3 lines · Markus Armbruster
You neglected to cc: the maintainer (me). I spotted this anyway, more or less by chance. Next time, please use
scripts/get_maintainer.plto find people you might want to cc:.
Ah, sorry for that. Will do next time, thanks!
Quoted 11 lines
Quoted 5 lines · Cyrill Gorcunov
We already have
error_report_onceandwarn_report_once, thus lets addinfo_report_onceto complement. Actually I use this helper a lot so might be usefull for others.Signed-off-by Cyrill Gorcunov <[email protected]>Patch looks reasonable enough. However, the new
info_report_once()andinfo_report_once_cond()aren't used. Your commit message suggests you have users in unpublished code. Is that correct?
Exactly. I work on yet unpublished but open source code
(https://src.openvz.org/projects/UP/repos/qemu/browse) to be precise. But work is still in progress and hasn't been pushed out.
You know I posted the patch simply because it might be useful for someone and it looks a bit incomplete that we have "warn/error once" but no info_once. Up to you to decide if we need it in vanilla tree.