[Bug general/27450] Update Polish translation
https://sourceware.org/bugzilla/show_bug.cgi?id=27450 --- Comment #13 from Piotr Drąg --- (In reply to Mark Wielaard from comment #11) > I added a ChangeLog entry to both and pushed them. Thanks. Thank you! (In reply to Mark Wielaard from comment #12) > Could/should we use one of these techniques when doing make check (in > --enable-maintainer-mode) to catch missing files? You could use intltool if you’re OK with a brittle, unmaintained Perl script in your release process. -- You are receiving this mail because: You are on the CC list for the bug.
[RFC PATCH 6/6] Support systemd watchdog
Signed-off-by: Victor Westerhuis
---
debuginfod/debuginfod.cxx | 35 +--
1 file changed, 33 insertions(+), 2 deletions(-)
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index 35b5fc18..b31c77d3 100644
--- a/debuginfod/debuginfod.cxx
+++ b/debuginfod/debuginfod.cxx
@@ -94,6 +94,7 @@ using namespace std;
#include
#ifdef ENABLE_SYSTEMD
+#include
#include
#endif
@@ -3600,13 +3601,43 @@ main (int argc, char *argv[])
}
}
+#ifdef ENABLE_SYSTEMD
+ uint64_t timeout_usec;
+ int watchdog_enabled = sd_watchdog_enabled (true, &timeout_usec);
+ if (watchdog_enabled < 0)
+report_error (0, -watchdog_enabled,
+ "warning: cannot determine if watchdog is enabled");
+#endif
+
/* Trivial main loop! */
#ifdef ENABLE_SYSTEMD
(void) sd_notify (false, "READY=1");
#endif
set_metric("ready", 1);
- while (! interrupted)
-pause ();
+#ifdef ENABLE_SYSTEMD
+ if (watchdog_enabled > 0)
+{
+ timeout_usec /= 2;
+ struct timespec timeout;
+ timeout.tv_sec = timeout_usec / 100;
+ timeout.tv_nsec = timeout_usec % 100 * 1000;
+
+ rc = 0;
+ struct timespec remaining_timeout;
+ while (! interrupted)
+{
+ if (rc == 0)
+{
+ (void) sd_notify (false, "WATCHDOG=1");
+ remaining_timeout = timeout;
+}
+ rc = nanosleep (&remaining_timeout, &remaining_timeout);
+}
+}
+ else
+#endif
+while (! interrupted)
+ pause ();
scanq.nuke(); // wake up any remaining scanq-related threads, let them die
set_metric("ready", 0);
#ifdef ENABLE_SYSTEMD
--
2.30.1
[RFC PATCH 5/6] Report error exit reason to systemd
Signed-off-by: Victor Westerhuis
---
debuginfod/debuginfod.cxx | 74 ---
1 file changed, 46 insertions(+), 28 deletions(-)
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index 9542c5e2..35b5fc18 100644
--- a/debuginfod/debuginfod.cxx
+++ b/debuginfod/debuginfod.cxx
@@ -120,6 +120,24 @@ using namespace std;
#endif
+#ifdef ENABLE_SYSTEMD
+#define report_error(status, errnum, format, ...) \
+ do \
+{ \
+ int s = (status), e = (errnum); \
+ if (s != 0 && e != 0) \
+{ \
+ (void) sd_notifyf (false, "ERRNO=%d", e); \
+ (void) sd_notifyf (false, "STATUS=" format, ##__VA_ARGS__); \
+} \
+ error (s, e, format, ##__VA_ARGS__); \
+} \
+ while (false)
+#else
+#define report_error(...) error(__VA_ARGS__)
+#endif
+
+
inline bool
string_endswith(const string& haystack, const string& needle)
{
@@ -3316,10 +3334,10 @@ main (int argc, char *argv[])
#endif
int rc = regcomp (& file_include_regex, ".*", REG_EXTENDED|REG_NOSUB); //
match everything
if (rc != 0)
-error (EXIT_FAILURE, 0, "regcomp failure: %d", rc);
+report_error (EXIT_FAILURE, 0, "regcomp failure: %d", rc);
rc = regcomp (& file_exclude_regex, "^$", REG_EXTENDED|REG_NOSUB); // match
nothing
if (rc != 0)
-error (EXIT_FAILURE, 0, "regcomp failure: %d", rc);
+report_error (EXIT_FAILURE, 0, "regcomp failure: %d", rc);
// default parameters for fdcache are computed from system stats
struct statfs sfs;
@@ -3337,8 +3355,8 @@ main (int argc, char *argv[])
argp_program_version_hook = print_version; // this works
(void) argp_parse (&argp, argc, argv, ARGP_IN_ORDER, &remaining, NULL);
if (remaining != argc)
- error (EXIT_FAILURE, 0,
- "unexpected argument: %s", argv[remaining]);
+ report_error (EXIT_FAILURE, 0,
+"unexpected argument: %s", argv[remaining]);
if (scan_archives.size()==0 && !scan_files && source_paths.size()>0)
obatched(clog) << "warning: without -F -R -U -Z, ignoring PATHs" << endl;
@@ -3362,13 +3380,13 @@ main (int argc, char *argv[])
if (rc == SQLITE_CORRUPT)
{
(void) unlink (db_path.c_str());
- error (EXIT_FAILURE, 0,
- "cannot open %s, deleted database: %s", db_path.c_str(),
sqlite3_errmsg(db));
+ report_error (EXIT_FAILURE, 0,
+"cannot open %s, deleted database: %s", db_path.c_str(),
sqlite3_errmsg(db));
}
else if (rc)
{
- error (EXIT_FAILURE, 0,
- "cannot open %s, consider deleting database: %s",
db_path.c_str(), sqlite3_errmsg(db));
+ report_error (EXIT_FAILURE, 0,
+"cannot open %s, consider deleting database: %s",
db_path.c_str(), sqlite3_errmsg(db));
}
// open the readonly query variant
@@ -3381,8 +3399,8 @@ main (int argc, char *argv[])
NULL);
if (rc)
{
- error (EXIT_FAILURE, 0,
- "cannot open %s, consider deleting database: %s",
db_path.c_str(), sqlite3_errmsg(dbq));
+ report_error (EXIT_FAILURE, 0,
+"cannot open %s, consider deleting database: %s",
db_path.c_str(), sqlite3_errmsg(dbq));
}
@@ -3393,16 +3411,16 @@ main (int argc, char *argv[])
rc = sqlite3_create_function(dbq, "sharedprefix", 2, SQLITE_UTF8, NULL,
& sqlite3_sharedprefix_fn, NULL, NULL);
if (rc != SQLITE_OK)
-error (EXIT_FAILURE, 0,
- "cannot create sharedprefix function: %s", sqlite3_errmsg(dbq));
+report_error (EXIT_FAILURE, 0,
+ "cannot create sharedprefix function: %s",
sqlite3_errmsg(dbq));
if (verbose > 3)
obatched(clog) << "ddl: " << DEBUGINFOD_SQLITE_DDL << endl;
rc = sqlite3_exec (db, DEBUGINFOD_SQLITE_DDL, NULL, NULL, NULL);
if (rc != SQLITE_OK)
{
- error (EXIT_FAILURE, 0,
- "cannot run database schema ddl: %s", sqlite3_errmsg(db));
+ report_error (EXIT_FAILURE, 0,
+"cannot run database schema ddl: %s", sqlite3_errmsg(db));
}
vector daemons;
@@ -3413,7 +3431,7 @@ main (int argc, char *argv[])
if (fds <= 0)
{
close_databases ();
- error (EXIT_FAILURE, -fds, "cannot get fds from systemd");
+ report_error (EXIT_FAILURE, -fds, "cannot get fds from systemd");
}
for (int i = 0; i < fds; ++i)
@@ -3424,13 +3442,13 @@ main (int argc, char *argv[])
if (rc < 0)
{
close_databases ();
- error (EXIT_FAILURE, -rc, "unable to determine fd type");
+ report_error (EXIT_FAILURE, -rc, "unable to determine fd type");
}
else if (rc == 0)
{
close_databases ();
- error (EXIT_FAILURE, 0,
- "fd %d is not of the correct socket type", fd);
+ report_error (EXIT_FAILURE, 0,
+"fd %d i
[RFC PATCH 3/6] Notify systemd when ready and stopping
Signed-off-by: Victor Westerhuis
---
config/debuginfod.service | 1 +
debuginfod/debuginfod.cxx | 6 ++
2 files changed, 7 insertions(+)
diff --git a/config/debuginfod.service b/config/debuginfod.service
index b64d8cb9..6c434705 100644
--- a/config/debuginfod.service
+++ b/config/debuginfod.service
@@ -4,6 +4,7 @@ Documentation=http://elfutils.org/
After=network.target
[Service]
+Type=Notify
EnvironmentFile=/etc/sysconfig/debuginfod
User=debuginfod
Group=debuginfod
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index caced48c..4dece371 100644
--- a/debuginfod/debuginfod.cxx
+++ b/debuginfod/debuginfod.cxx
@@ -3576,11 +3576,17 @@ main (int argc, char *argv[])
}
/* Trivial main loop! */
+#ifdef ENABLE_SYSTEMD
+ (void) sd_notify (false, "READY=1");
+#endif
set_metric("ready", 1);
while (! interrupted)
pause ();
scanq.nuke(); // wake up any remaining scanq-related threads, let them die
set_metric("ready", 0);
+#ifdef ENABLE_SYSTEMD
+ (void) sd_notify (false, "STOPPING=1");
+#endif
if (verbose)
obatched(clog) << "stopping" << endl;
--
2.30.1
[RFC PATCH 2/6] Allow socket activation by systemd
Signed-off-by: Victor Westerhuis
---
debuginfod/debuginfod.cxx | 161 --
1 file changed, 121 insertions(+), 40 deletions(-)
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index 2aecc049..caced48c 100644
--- a/debuginfod/debuginfod.cxx
+++ b/debuginfod/debuginfod.cxx
@@ -93,6 +93,10 @@ using namespace std;
#include
+#ifdef ENABLE_SYSTEMD
+#include
+#endif
+
#if MHD_VERSION >= 0x00097002
// libmicrohttpd 0.9.71 broke API
#define MHD_RESULT enum MHD_Result
@@ -367,6 +371,10 @@ static const struct argp_option options[] =
{ "fdcache-prefetch", ARGP_KEY_FDCACHE_PREFETCH, "NUM", 0, "Number of
archive files to prefetch into fdcache.", 0 },
#define ARGP_KEY_FDCACHE_MINTMP 0x1004
{ "fdcache-mintmp", ARGP_KEY_FDCACHE_MINTMP, "NUM", 0, "Minimum free space%
on tmpdir.", 0 },
+#ifdef ENABLE_SYSTEMD
+#define ARGP_KEY_SYSTEMD 0x1005
+ { "systemd", ARGP_KEY_SYSTEMD, NULL, 0, "Use systemd socket activation and
cache directory.", 0 },
+#endif
{ NULL, 0, NULL, 0, NULL, 0 }
};
@@ -412,6 +420,9 @@ static long fdcache_mbs;
static long fdcache_prefetch;
static long fdcache_mintmp;
static string tmpdir;
+#ifdef ENABLE_SYSTEMD
+static bool systemd;
+#endif
static void set_metric(const string& key, double value);
// static void inc_metric(const string& key);
@@ -541,6 +552,11 @@ parse_opt (int key, char *arg,
case ARGP_KEY_ARG:
source_paths.insert(string(arg));
break;
+#ifdef ENABLE_SYSTEMD
+case ARGP_KEY_SYSTEMD:
+ systemd = true;
+ break;
+#endif
// case 'h': argp_state_help (state, stderr,
ARGP_HELP_LONG|ARGP_HELP_EXIT_OK);
default: return ARGP_ERR_UNKNOWN;
}
@@ -3267,6 +3283,16 @@ static void sqlite3_sharedprefix_fn (sqlite3_context* c,
int argc, sqlite3_value
}
+static void close_databases ()
+{
+ sqlite3 *database = db;
+ sqlite3 *databaseq = dbq;
+ db = dbq = 0; // for signal_handler not to freak
+ (void) sqlite3_close (databaseq);
+ (void) sqlite3_close (database);
+}
+
+
int
main (int argc, char *argv[])
{
@@ -3372,46 +3398,105 @@ main (int argc, char *argv[])
"cannot run database schema ddl: %s", sqlite3_errmsg(db));
}
- // Start httpd server threads. Separate pool for IPv4 and IPv6, in
- // case the host only has one protocol stack.
- MHD_Daemon *d4 = MHD_start_daemon (MHD_USE_THREAD_PER_CONNECTION
+ vector daemons;
+#ifdef ENABLE_SYSTEMD
+ if (systemd)
+{
+ int fds = sd_listen_fds (true);
+ if (fds <= 0)
+{
+ close_databases ();
+ error (EXIT_FAILURE, -fds, "cannot get fds from systemd");
+}
+
+ for (int i = 0; i < fds; ++i)
+{
+ int fd = SD_LISTEN_FDS_START + i;
+
+ rc = sd_is_socket_inet (fd, AF_UNSPEC, SOCK_STREAM, true, 0);
+ if (rc < 0)
+{
+ close_databases ();
+ error (EXIT_FAILURE, -rc, "unable to determine fd type");
+}
+ else if (rc == 0)
+{
+ close_databases ();
+ error (EXIT_FAILURE, 0,
+ "fd %d is not of the correct socket type", fd);
+}
+
+ MHD_Daemon *d = MHD_start_daemon (MHD_USE_THREAD_PER_CONNECTION
#if MHD_VERSION >= 0x00095300
- | MHD_USE_INTERNAL_POLLING_THREAD
+| MHD_USE_INTERNAL_POLLING_THREAD
#else
- | MHD_USE_SELECT_INTERNALLY
+| MHD_USE_SELECT_INTERNALLY
#endif
- | MHD_USE_DEBUG, /* report errors to
stderr */
- http_port,
- NULL, NULL, /* default accept policy */
- handler_cb, NULL, /* handler callback */
- MHD_OPTION_END);
- MHD_Daemon *d6 = MHD_start_daemon (MHD_USE_THREAD_PER_CONNECTION
+| MHD_USE_DEBUG /* report errors
to stderr */
+| MHD_USE_ITC, /* don't close
socket on shutdown */
+0,
+NULL, NULL, /* default accept
policy */
+handler_cb, NULL, /* handler
callback */
+MHD_OPTION_LISTEN_SOCKET,
+fd,
+MHD_OPTION_END);
+ if (d == NULL)
+{
+ close_databases ();
+ error (EXIT_FAILURE, 0,
+"cannot start http server on fd %d", fd);
+}
+
+ obatched(clog) << "started http server on fd " << fd << endl;
+
+ daemons.push_back (d);
+}
+}
+ else
+#endif
+{
+ // Start httpd server threads. S
[RFC PATCH 0/6] Allow systemd activation for debuginfod
As a developer I have a few self-built Debian packages in a small repo on my own laptop. I like having debuginfod running when debugging them, but it does not have to be running all the time. This patch series is a first try at integrating systemd socket activation in debuginfod. I've tried to split the series into different ways of integration. I'm not sure if watchdog support is really useful, since the main loop doesn't do any work. If it isn't, I can drop it. Victor Westerhuis (6): Find and link libsystemd Allow socket activation by systemd Notify systemd when ready and stopping Use cache directory from systemd Report error exit reason to systemd Support systemd watchdog config/debuginfod.service | 3 +- configure.ac | 12 ++ debuginfod/Makefile.am| 7 + debuginfod/debuginfod.cxx | 267 +- 4 files changed, 226 insertions(+), 63 deletions(-) -- 2.30.1
[RFC PATCH 1/6] Find and link libsystemd
Signed-off-by: Victor Westerhuis
---
configure.ac | 12
debuginfod/Makefile.am | 7 +++
2 files changed, 19 insertions(+)
diff --git a/configure.ac b/configure.ac
index aa8439e8..8d52b666 100644
--- a/configure.ac
+++ b/configure.ac
@@ -773,6 +773,17 @@ AS_IF([test "x$enable_debuginfod" != "xno"], [
if test "x$enable_debuginfod" = "xno"; then
AC_MSG_ERROR([dependencies not found, use --disable-debuginfod to
disable.])
fi
+
+AC_ARG_ENABLE([systemd],AC_HELP_STRING([--enable-systemd], [Enable systemd
support in debuginfod]))
+AS_IF([test "x$enable_systemd" != "xno"], [
+PKG_CHECK_MODULES([libsystemd],[libsystemd],[enable_systemd=yes],[
+AS_IF([test "x$enable_systemd" = "xyes"], [
+AC_MSG_ERROR([need libsystemd, use --disable-systemd to
disable.])
+], [enable_systemd=no])
+])
+])
+AS_IF([test "x$enable_systemd" =
"xyes"],AC_DEFINE([ENABLE_SYSTEMD],[1],[Enable systemd support]))
+AM_CONDITIONAL([SYSTEMD],[test "x$enable_systemd" = "xyes"])
])
AS_IF([test "x$enable_debuginfod" !=
"xno"],AC_DEFINE([ENABLE_DEBUGINFOD],[1],[Build debuginfod]))
@@ -823,6 +834,7 @@ AC_MSG_NOTICE([
Extra Valgrind annotations : ${use_vg_annotations}
libdebuginfod client support : ${enable_libdebuginfod}
Debuginfod server support : ${enable_debuginfod}
+Debuginfod systemd support : ${enable_systemd}
Default DEBUGINFOD_URLS: ${default_debuginfod_urls}
EXTRA TEST FEATURES (used with make check)
diff --git a/debuginfod/Makefile.am b/debuginfod/Makefile.am
index 3adb2755..e7528c3a 100644
--- a/debuginfod/Makefile.am
+++ b/debuginfod/Makefile.am
@@ -35,6 +35,10 @@ AM_CPPFLAGS += -I$(srcdir) -I$(srcdir)/../libelf
-I$(srcdir)/../libebl \
$(libmicrohttpd_CFLAGS) $(libcurl_CFLAGS) $(sqlite3_CFLAGS) \
$(libarchive_CFLAGS)
+if SYSTEMD
+AM_CPPFLAGS += $(libsystemd_CFLAGS)
+endif
+
# Disable eu- prefixing for artifacts (binaries & man pages) in this
# directory, since they do not conflict with binutils tools.
program_prefix=
@@ -71,6 +75,9 @@ endif
debuginfod_SOURCES = debuginfod.cxx
debuginfod_LDADD = $(libdw) $(libelf) $(libeu) $(libdebuginfod) $(argp_LDADD)
$(fts_LIBS) $(libmicrohttpd_LIBS) $(sqlite3_LIBS) $(libarchive_LIBS) -lpthread
-ldl
+if SYSTEMD
+debuginfod_LDADD += $(libsystemd_LIBS)
+endif
debuginfod_find_SOURCES = debuginfod-find.c
debuginfod_find_LDADD = $(libdw) $(libelf) $(libeu) $(libdebuginfod)
$(argp_LDADD) $(fts_LIBS)
--
2.30.1
[RFC PATCH 4/6] Use cache directory from systemd
Signed-off-by: Victor Westerhuis
---
config/debuginfod.service | 2 +-
debuginfod/debuginfod.cxx | 7 +++
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/config/debuginfod.service b/config/debuginfod.service
index 6c434705..c99b5406 100644
--- a/config/debuginfod.service
+++ b/config/debuginfod.service
@@ -8,7 +8,7 @@ Type=Notify
EnvironmentFile=/etc/sysconfig/debuginfod
User=debuginfod
Group=debuginfod
-#CacheDirectory=debuginfod
+CacheDirectory=debuginfod
ExecStart=/usr/bin/debuginfod -d /var/cache/debuginfod/debuginfod.sqlite -p
$DEBUGINFOD_PORT $DEBUGINFOD_VERBOSE $DEBUGINFOD_PRAGMAS $DEBUGINFOD_PATHS
# Stopping can take a long time if scanning of large archives is in progress
TimeoutStopSec=60
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index 4dece371..9542c5e2 100644
--- a/debuginfod/debuginfod.cxx
+++ b/debuginfod/debuginfod.cxx
@@ -3307,6 +3307,13 @@ main (int argc, char *argv[])
/* Set computed default values. */
db_path = string(getenv("HOME") ?: "/") + string("/.debuginfod.sqlite"); /*
XDG? */
+#ifdef ENABLE_SYSTEMD
+ if (getenv("CACHE_DIRECTORY"))
+{
+ db_path = string(getenv("CACHE_DIRECTORY")) +
string("/debuginfod.sqlite");
+ setenv (DEBUGINFOD_CACHE_PATH_ENV_VAR, getenv("CACHE_DIRECTORY"), false);
+}
+#endif
int rc = regcomp (& file_include_regex, ".*", REG_EXTENDED|REG_NOSUB); //
match everything
if (rc != 0)
error (EXIT_FAILURE, 0, "regcomp failure: %d", rc);
--
2.30.1
Re: [RFC PATCH 0/6] Allow systemd activation for debuginfod
Hi - > As a developer I have a few self-built Debian packages in a small repo > on my own laptop. I like having debuginfod running when debugging them, > but it does not have to be running all the time. What about the grooming/scanning cycles? > This patch series is a first try at integrating systemd socket > activation in debuginfod. I've tried to split the series into different > ways of integration. [...] Thanks, the code is fairly readable in this breakdown. The mhd initialization looks rather worse, and if we decided to do this, I might focus on simplifying that. (Maybe I'd use the classic inetd socket-stdin/stdout style rather than systemd-specific APIs.) But I'm not sure a web-api-only functionality makes much sense - unless there is also a sibling debuginfod to do the scanning/grooming, in which case ... how does this help? - FChE
Re: [RFC PATCH 0/6] Allow systemd activation for debuginfod
Hi, On Sun, 2021-03-07 at 09:54 -0500, Frank Ch. Eigler via Elfutils-devel wrote: > As a developer I have a few self-built Debian packages in a small > > repo > > on my own laptop. I like having debuginfod running when debugging them, > > but it does not have to be running all the time. > > What about the grooming/scanning cycles? I was thinking the same thing. This patch series takes care of the web- socket activation part. systemd also provides path-based activation of units [1]. That could maybe be used for the scanning cycles. systemd would then trigger activation when something is dropped into (or removed from) the search directory. Cheers, Mark [1] https://www.freedesktop.org/software/systemd/man/systemd.path.html
Re: [PATCH] debuginfod-client: Don't compare a double to a long
Hi - Merged (along with Jakub's suggestions). - FChE
Re: [RFC PATCH 0/6] Allow systemd activation for debuginfod
Hi - > > What about the grooming/scanning cycles? > > I was thinking the same thing. This patch series takes care of the web- > socket activation part. systemd also provides path-based activation of > units [1]. That could maybe be used for the scanning cycles. systemd > would then trigger activation when something is dropped into (or > removed from) the search directory. yeah but it's not recursive, not symlink-following-aware, doesn't work for all filesystems (NFS!). - FChE
uncontroversial process patch
Hi - Is it time? diff --git a/CONTRIBUTING b/CONTRIBUTING index bb48975bac33..2be49cd9cc84 100644 --- a/CONTRIBUTING +++ b/CONTRIBUTING @@ -50,8 +50,9 @@ using your real name (sorry, no pseudonyms or anonymous contributions.) git commit --signoff will add such a Signed-off-by line at the end of the commit log message for you. -The ideal patch contains a ChangeLog entry and a test case for the -bug fixed or feature added. +The ideal patch contains a ChangeLog entry in the git commit log message, +(ChangeLog files being deprecated), and a test case for the bug fixed or +feature added. The testsuite (make check) is expected to have zero failing tests. Do not knowingly add tests that FAIL. If there are architectures or
Re: [RFC PATCH 0/6] Allow systemd activation for debuginfod
Hi - > But I'm not sure a web-api-only functionality makes much sense - > unless there is also a sibling debuginfod to do the scanning/grooming, > in which case ... how does this help? Another angle to consider: metrics/monitoring. The code goes to some effort to let an operator know how the system running: how fast sqlite queries are, what sort of network throughput to expect. If instead we have short-lived processes serving this or that purpose, we lose monitoring capability completely. - FChE
Buildbot failure in Wildebeest Builder on whole buildset
The Buildbot has detected a failed build on builder whole buildset while building elfutils. Full details are available at: https://builder.wildebeest.org/buildbot/#builders/4/builds/742 Buildbot URL: https://builder.wildebeest.org/buildbot/ Worker for this Build: debian-i386 Build Reason: Blamelist: Timm Bäder BUILD FAILED: failed test (failure) Sincerely, -The Buildbot
[Bug debuginfod/27531] Support retry of failed downloads
https://sourceware.org/bugzilla/show_bug.cgi?id=27531 --- Comment #3 from Sergio Durigan Junior --- (In reply to Frank Ch. Eigler from comment #1) > Can you collect more information about the nature of the timeouts and > what diagnostics if any libcurl/debuginfod-client returned? The default > timeout imposed by the debuginfod-client code is 90s to start returning > file content, as governed by the $DEBUGINFOD_TIMEOUT environment variable. > Would these folks like us to retry *beyond* that timeout? First of all: yes, I can try to collect more info from them. I've just emailed one of the guys who brought this to my attention, and asked him if he can provide more info and leave a comment here. With that out of the way, I think it's worth mentioning that the retry idea did not come from the reports I've received. It is something that I figured would be nice to have in order to mitigate the issues. We are talking about two distinct (albeit related) things here: having a big timeout does not necessarily mean that the download will succeeded, therefore having the possibility of retrying makes a lot of sense. > Or is the timeout at some earlier stage? We'd need the diagnostics or > packet trace or something like that. I understand the intention to address the timeout problem, but there are things out of debuginfod's control here. One of the emails I received was from a person who lives in China and therefore has to cope with their sub-optimal international links. As far as I understood from what they explained, a connection between China and Europe is prone to suffer from the instability that was reported. -- You are receiving this mail because: You are on the CC list for the bug.
[Bug gdb/27532] GDB should ask the user for permission before downloading debuginfod files
https://sourceware.org/bugzilla/show_bug.cgi?id=27532 Sergio Durigan Junior changed: What|Removed |Added Resolution|WONTFIX |--- Product|elfutils|gdb Summary|debuginfod should ask the |GDB should ask the user for |user for permission before |permission before |downloading files |downloading debuginfod ||files Status|RESOLVED|REOPENED Version|unspecified |HEAD Component|debuginfod |gdb --- Comment #4 from Sergio Durigan Junior --- Instead of WONTFIX for debuginfod, I am reopening this against GDB because I still think it is a good idea for the client to implement this feature. -- You are receiving this mail because: You are on the CC list for the bug.
[Bug debuginfod/27531] Support retry of failed downloads
https://sourceware.org/bugzilla/show_bug.cgi?id=27531 --- Comment #4 from Frank Ch. Eigler --- > having a big timeout does not necessarily mean that the download will > succeeded, therefore having the possibility of retrying makes a lot of sense. OK, retrying for an outright aborted connection within the overall timeout limit is something we could do. Retrying AFTER a timeout, probably not obvious; we'd need yet another timeout to limit the post-timeout timeout, yo dawg. -- You are receiving this mail because: You are on the CC list for the bug.
