aboutsummaryrefslogtreecommitdiff
path: root/src/util
Commit message (Collapse)AuthorAge
* staticChristian Grothoff2011-11-10
|
* fix #1885Christian Grothoff2011-11-07
|
* introducing GNUNET_abortChristian Grothoff2011-11-05
|
* treat empty string as NULLChristian Grothoff2011-11-05
|
* LRN: Add -lole32 (why had it worked without this untilChristian Grothoff2011-11-05
| | | | | now?)
* LRN: Add -lole32 (why had it worked without this untilChristian Grothoff2011-11-05
| | | | | now?)
* LRN: Mysterious patch. I don't know why calling executable with a dotChristian Grothoff2011-11-05
| | | | | | | | | forces it to create a new console window (which immediately pops up and grabs the focus, then disappears) for itself, but that is the empirical fact. Removing the dot fixes this VERY annoying W32 bug.
* LRN: Use regexps in logdefsChristian Grothoff2011-11-05
| | | | | | | | | | | | | Now "*" is not valid anymore (it's not like anyone had a lot of opportunities to use it though). Empty string means ".*", so it's still compatible. Matching is case-sensitive. ^ and $ match the beginning and the end of the text, newine matches a newline character. All regexps are considered non-extended. Regex compilation errors are not reported in any way (the parsing process just fails, and that's it).
* LRN: Add '-Wl,' prefix to -no-undefinedChristian Grothoff2011-11-05
| | | | | | | | | | -no-undefined is not a valid compiler option as of GCC 4.6.x (it IS valid linker option). Note that this REQUIRES an extra -no-undefined to be passed to libtool (this is done by GN_LIB_LDLFAGS and GN_PLUGIN_LDFLAGS), because libtool does not understand -Wl,-no-undefined.
* LRN: Fix loglevel testcase (disable redirection)Christian Grothoff2011-11-05
|
* use putenv instead of setenv for portabilityChristian Grothoff2011-11-05
|
* curly wars / auto-indentationChristian Grothoff2011-11-04
|
* cov fixChristian Grothoff2011-11-04
|
* do not translate, everChristian Grothoff2011-11-04
|
* fixChristian Grothoff2011-11-04
|
* nodebugChristian Grothoff2011-11-03
|
* implementing time and size parsers for #1875Christian Grothoff2011-11-03
|
* implementing plugin directory scanner for #1746Christian Grothoff2011-11-03
|
* increase timeout to fix #1872:Christian Grothoff2011-11-03
|
* implementing #1747Christian Grothoff2011-11-03
|
* fixChristian Grothoff2011-11-02
|
* setenv to fix testcase under other circumstancesChristian Grothoff2011-10-26
|
* getting rid of unnecessary ignore_shutdown callsChristian Grothoff2011-10-23
|
* fixChristian Grothoff2011-10-20
|
* fixChristian Grothoff2011-10-20
|
* bracketChristian Grothoff2011-10-19
|
* LRN: fix LOG definitionsChristian Grothoff2011-10-14
|
* LRN: Use GNUNET_strdup() instead of strdup()Christian Grothoff2011-10-14
|
* LRN loves slist: Use stack allocation for slist iteratorChristian Grothoff2011-10-14
|
* allow NULLChristian Grothoff2011-10-12
|
* LRN: Fix slist misuse in network.cChristian Grothoff2011-10-12
|
* check vspnrintf return value, stack-allocate log bufferChristian Grothoff2011-10-11
|
* reudce insanityChristian Grothoff2011-10-11
|
* converting to GNUNET_LOG_from*Christian Grothoff2011-10-11
|
* spaceChristian Grothoff2011-10-10
|
* use LOG macro in resolver_api.cChristian Grothoff2011-10-10
|
* use log fromChristian Grothoff2011-10-10
|
* LRN: Fix logdef processing logicChristian Grothoff2011-10-09
| | | | | | Now if logdef matches, it is used to both allow AND restrict the call.
* LRN: Small janitor fixesChristian Grothoff2011-10-07
| | | | | | | | * Prevent janitor from failing if a process can't be killed (it might be dead by the time janitor gets around to killing it). * Fix janitor messages about killing arm/non-arm processes
* compile errorMatthias Wachs2011-10-06
|
* ughChristian Grothoff2011-10-06
|
* change test to not require . in PATHChristian Grothoff2011-10-06
|
* vminko: new API to delete a TransmitContext (slightly modified patch from #1822)Christian Grothoff2011-10-06
|
* LRN: Fix runtime logging test, since pipes work as intendedChristian Grothoff2011-10-06
|
* LRN: Fix select() on pipes to trigger read callback on EPIPEChristian Grothoff2011-10-06
|
* testcase for runtime loglevels form LRNChristian Grothoff2011-10-06
|
* fixChristian Grothoff2011-10-05
|
* LRN: Use GNUNET_EXTRA_LOGGING to manage compile-time logging callsChristian Grothoff2011-10-04
| | | | | | (#1805).
* LRN's big logging rewrite (#1805):Christian Grothoff2011-10-04
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * GNUNET_BOTTOM_LOGLEVEL and GNUNET_TOP_LOGLEVEL set global levels Use bottom level to force logging to be more verbose than configured Use top level to force logging to be less verbose than configured Obviously, bottom <= top * GNUNET_LOG sets per-component levels GNUNET_LOG looks like this: name[/bottom[/top]]/... name starts with a non-digit character, must not include '/' bottom and top must consist only of digits, or be empty a description is only used if it matches the component exactly as a special exception (for now) the name '*' matches any component per-component loglevels override global loglevels global levels override whatever is given via arguments or in config Examples: test_client/8/8/ run test_client with DEBUG level (usually leads to a timeout, by the way) */2/2/core/8/8/transport/4/4 run everything with WARNING, core - with DEBUG, transport - with INFO *//1/peerinfo/4/ run everything with top loglevel ERROR, global/configured bottom loglevel, and peerinfo - with bottom loglevel INFO and global/configured top loglevel statistics/ does nothing * Added GNUNET_ERROR_TYPE_UNSPECIFIED enum value, to hold -1. Its corresponding string is NULL. * Changed the logger calls as Grothoff suggested - to use static int to hold the result of runtime evaluation of logability. Logging can be unconditionally disabled in advance by defining GNUNET_LOG_CALL_STATUS to 0, and enabled in advance by defining it to 1. * Added GNUNET_CULL_LOGGING, which, if defined, completely culls out all logging calls at compile time. * Log definition parsing is only done once, results are cached. * Changed definition format, now it looks like this: [component|*|];[file|*|];[function|*|];[from_line[-to_line]];level/[component...] All field separators are mandatory (but some fields could be empty or be '*'). Line definition must be either empty or "number" or "number-number" Level definition must not be empty, and is a string representation of the level (i.e. DEBUG, WARNING, INFO, etc). Definition entry must end with a slash, whether or not there's another entry after it. File name is matched to the end of __FILE__, which allows file name to match not only the base name, but also directories leading to it. * Removed default WARNING loglevel from program and service utility code. Now they default to NULL (UNSPECIFIED) level, which can be overriden by GNUNET_LOG definition, if no level is specified via config or commandline. Log levels from config or commandline are overriden by GNUNET_FORCE_LOG. If GNUNET_*LOG are undefined, and no levels came from config or commandline, logger internally defaults to WARNING level. Add --enable-logging configure option
* fix 1814Christian Grothoff2011-10-02
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-04-28 16:58:01 +0000