29d94131eeFixed minor Boost Build error reporting bugs caused by module rules not being directly available using their full names when imported from inside a class rule - an error about rule errors.error or errors.user-error not being defined was getting reported instead of the desired error. This is due to the fact that rule name resolution does not look in the current module in that case but uses the class module instead.
Jurko Gospodnetić
2012-07-20 14:39:50 +0000
35fad6df6cFixed minor Boost Build error reporting bugs caused by module rules not being directly available using their full names when imported from inside a class rule - an error about rule errors.error or errors.user-error not being defined was getting reported instead of the desired error. This is due to the fact that rule name resolution does not look in the current module in that case but uses the class module instead.
Jurko Gospodnetić
2012-07-20 14:39:50 +0000
bae5e920acBoost Jam code cleanup - removed an old, untested and up to recently definitely non-compiling 'old Borland compiler' code branch in the filent.c module. The compilers in question themselves are Borlands free compiler tools predating the more modern and more standard compliant free 'Visual Studio Express' toolchain so there should be no need for anyone to depend on them any more.
Jurko Gospodnetić
2012-07-20 10:57:31 +0000
b95ccd82b9Boost Jam code cleanup - removed an old, untested and up to recently definitely non-compiling 'old Borland compiler' code branch in the filent.c module. The compilers in question themselves are Borlands free compiler tools predating the more modern and more standard compliant free 'Visual Studio Express' toolchain so there should be no need for anyone to depend on them any more.
Jurko Gospodnetić
2012-07-20 10:57:31 +0000
ae0e14f947Fixed a Boost Build bug causing scanner targets to be rebuilt every time if their base target has already been marked as including another target with a newer timestamp. For example, this occurred every time a single action generated both the scanned and the included target and did so slowly enough that the included target got a newer timestamp. Fix originally prepared by Steven Watanabe.
Jurko Gospodnetić
2012-07-19 14:32:31 +0000
81e8b88e8aFixed a Boost Build bug causing scanner targets to be rebuilt every time if their base target has already been marked as including another target with a newer timestamp. For example, this occurred every time a single action generated both the scanned and the included target and did so slowly enough that the included target got a newer timestamp. Fix originally prepared by Steven Watanabe.
Jurko Gospodnetić
2012-07-19 14:32:31 +0000
a80848695bCorrected the internal Boost Build library_order.py unit test - updated its test run parameter passing to use lists of strings instead of simple strings (was causing assertion failures with the latest internal Boost Build testing framework implementation), minor stylistic changes.
Jurko Gospodnetić
2012-07-19 12:30:48 +0000
523f030192Corrected the internal Boost Build library_order.py unit test - updated its test run parameter passing to use lists of strings instead of simple strings (was causing assertion failures with the latest internal Boost Build testing framework implementation), minor stylistic changes.
Jurko Gospodnetić
2012-07-19 12:30:48 +0000
0e4d14bfe6Added a new internal Boost Build test illustrating a bug in the current Boost Build/Jam implementation and causing targets to be rebuilt when not necessary in some cases.
Jurko Gospodnetić
2012-07-19 11:53:30 +0000
a3dde2e151Added a new internal Boost Build test illustrating a bug in the current Boost Build/Jam implementation and causing targets to be rebuilt when not necessary in some cases.
Jurko Gospodnetić
2012-07-19 11:53:30 +0000
e85732818dCorrected internal Boost Build testing framework bug causing it to report detected filesystem changes from a previous instead of the failed test run when the failure was caused by an unexpected exit code, unexpected stderr content, unexpected stdout content or exceeded maximal test duration.
Jurko Gospodnetić
2012-07-19 10:16:41 +0000
f4dd360dd7Corrected internal Boost Build testing framework bug causing it to report detected filesystem changes from a previous instead of the failed test run when the failure was caused by an unexpected exit code, unexpected stderr content, unexpected stdout content or exceeded maximal test duration.
Jurko Gospodnetić
2012-07-19 10:16:41 +0000
e0f7059b26Removed an unused variable from one of the internal Boost Build tests.
Jurko Gospodnetić
2012-07-19 04:39:42 +0000
d32d7ee1cbRemoved an unused variable from one of the internal Boost Build tests.
Jurko Gospodnetić
2012-07-19 04:39:42 +0000
3c935afa5fCleaned up Boost Build's internal 'parallel multi-file actions' tests a bit - removed some dead code and updated description comments.
Jurko Gospodnetić
2012-07-19 04:11:50 +0000
489d184636Cleaned up Boost Build's internal 'parallel multi-file actions' tests a bit - removed some dead code and updated description comments.
Jurko Gospodnetić
2012-07-19 04:11:50 +0000
2f64a35846Boost Build code cleanup - minor stylistic changes (trimmed trailing spaces, corrected a typo in the main module comment).
Jurko Gospodnetić
2012-07-19 03:47:31 +0000
776deb54b3Boost Build code cleanup - minor stylistic changes (trimmed trailing spaces, corrected a typo in the main module comment).
Jurko Gospodnetić
2012-07-19 03:47:31 +0000
681b6609afBoost Jam code cleanup - minor stylistic changes (added missing spaces around some parameters inside brackets in C code).
Jurko Gospodnetić
2012-07-18 15:06:44 +0000
727365fb20Boost Jam code cleanup - minor stylistic changes (added missing spaces around some parameters inside brackets in C code).
Jurko Gospodnetić
2012-07-18 15:06:44 +0000
cbb48b5d01Updated Boost Jam implementation comment related to it adding internal fake INCLUDES relationships between targets built by a single action in order to make sure none of them may get used until their action had successfully completed its work.
Jurko Gospodnetić
2012-07-18 14:50:53 +0000
f2c607bc0dUpdated Boost Jam implementation comment related to it adding internal fake INCLUDES relationships between targets built by a single action in order to make sure none of them may get used until their action had successfully completed its work.
Jurko Gospodnetić
2012-07-18 14:50:53 +0000
35883dacc0Boost Build build bug corrected when compiling using old Borland compilers. This has been in for a long time and has been noticed only as part of a quick code review. It could be there is no need to support this 'old Borland compilers' code branch.
Jurko Gospodnetić
2012-07-18 14:27:22 +0000
3462295cf2Boost Build build bug corrected when compiling using old Borland compilers. This has been in for a long time and has been noticed only as part of a quick code review. It could be there is no need to support this 'old Borland compilers' code branch.
Jurko Gospodnetić
2012-07-18 14:27:22 +0000
58beabb9dcBoost Jam code cleanup - moved code for creating new internal include targets into a single location (rules.c module) instead of having it done in two (rules.c & builtins.c modules).
Jurko Gospodnetić
2012-07-18 14:24:19 +0000
116439c266Boost Jam code cleanup - moved code for creating new internal include targets into a single location (rules.c module) instead of having it done in two (rules.c & builtins.c modules).
Jurko Gospodnetić
2012-07-18 14:24:19 +0000
1961065003Boost Build no longer needs to add a fake INCLUDES relationship between targets generated by a single Boost Build action (reapplying revision [79571]). This is already done by Boost Jam internally for a single Boost Jam action and in case when the Boost Build action is actually a rule registering multiple actions - this was not causing any desirable behaviour.
Jurko Gospodnetić
2012-07-18 14:08:35 +0000
5a2b9e404aBoost Build no longer needs to add a fake INCLUDES relationship between targets generated by a single Boost Build action (reapplying revision [79571]). This is already done by Boost Jam internally for a single Boost Jam action and in case when the Boost Build action is actually a rule registering multiple actions - this was not causing any desirable behaviour.
Jurko Gospodnetić
2012-07-18 14:08:35 +0000
ca78bc40b2Minor stylistic changes made to the Python Boost Build comment updated in the previous commit.
Jurko Gospodnetić
2012-07-18 13:26:02 +0000
1f8e7334eeMinor stylistic changes made to the Python Boost Build comment updated in the previous commit.
Jurko Gospodnetić
2012-07-18 13:26:02 +0000
714ed09cf4Updated Boost Build implementation note comment related to why it adds an INCLUDES relation between all sources registered for a single Boost Build action.
Jurko Gospodnetić
2012-07-18 13:24:40 +0000
fca2444fe2Updated Boost Build implementation note comment related to why it adds an INCLUDES relation between all sources registered for a single Boost Build action.
Jurko Gospodnetić
2012-07-18 13:24:40 +0000
e6a3faf35dCorrected the internal Boost Build generators_test.py unit test to correctly split its appender 'action' into separate appender1 actions (building only a single target each), and still have them all called correctly together or none at all. Old and too-simplistic solution registered called completely independent actions for each target, causing Boost Jam to build only some of the targets in some cases, unlike what it would have done had they all actually been built by a single action.
Jurko Gospodnetić
2012-07-18 12:23:18 +0000
a11cbcada3Corrected the internal Boost Build generators_test.py unit test to correctly split its appender 'action' into separate appender1 actions (building only a single target each), and still have them all called correctly together or none at all. Old and too-simplistic solution registered called completely independent actions for each target, causing Boost Jam to build only some of the targets in some cases, unlike what it would have done had they all actually been built by a single action.
Jurko Gospodnetić
2012-07-18 12:23:18 +0000
ddbdcd0c04Reverting revision [79571] since it broke the generators_test.py Boost Build test. It seems Boost Jam might not have the complete fix implemented to allow us to avoid this Boost Build code.
Jurko Gospodnetić
2012-07-17 09:29:02 +0000
0677deb372Reverting revision [79571] since it broke the generators_test.py Boost Build test. It seems Boost Jam might not have the complete fix implemented to allow us to avoid this Boost Build code.
Jurko Gospodnetić
2012-07-17 09:29:02 +0000
ef8a393163Boost Build code cleanup - Boost Build no longer needs to add a fake INCLUDES relationship between targets generated by a single action as this is already done by Boost Jam internally.
Jurko Gospodnetić
2012-07-17 09:08:10 +0000
1b73d419e7Boost Build code cleanup - Boost Build no longer needs to add a fake INCLUDES relationship between targets generated by a single action as this is already done by Boost Jam internally.
Jurko Gospodnetić
2012-07-17 09:08:10 +0000
21e5ae661aBoost Build code cleanup - scanner.install no longer taking an extra & never used vtarget parameter.
Jurko Gospodnetić
2012-07-17 09:03:38 +0000
d3f709b3afBoost Build code cleanup - scanner.install no longer taking an extra & never used vtarget parameter.
Jurko Gospodnetić
2012-07-17 09:03:38 +0000
c16ae1c68aBoost Jam/Build cleanup - minor stylistic changes (comment updates; line wrapping; removed some dead/unused function declarations, parameters & code; declared variables as const; reordered some #include directives alphabetically, removed some unnecessary ones and added several missing ones discovered by the reordering).
Jurko Gospodnetić
2012-07-17 08:56:49 +0000
fa43d8fa93Boost Jam/Build cleanup - minor stylistic changes (comment updates; line wrapping; removed some dead/unused function declarations, parameters & code; declared variables as const; reordered some #include directives alphabetically, removed some unnecessary ones and added several missing ones discovered by the reordering).
Jurko Gospodnetić
2012-07-17 08:56:49 +0000
145ce56f80Windows Boost Jam implementation now uses finer resolution than 1 second when creating timestamps based on the current system time. Since these are never (and should never be!) compared to file system timestamps - this causes no conflicts with file system timestamps still using 1 second resolution at best.
Jurko Gospodnetić
2012-07-14 18:55:06 +0000
1f4e39984bWindows Boost Jam implementation now uses finer resolution than 1 second when creating timestamps based on the current system time. Since these are never (and should never be!) compared to file system timestamps - this causes no conflicts with file system timestamps still using 1 second resolution at best.
Jurko Gospodnetić
2012-07-14 18:55:06 +0000
c2858c8b5dBoost Jam code cleanup - extracted code for getting a timestamp for the current system time to timestamp_current().
Jurko Gospodnetić
2012-07-14 17:28:50 +0000
0ff040d3a0Boost Jam code cleanup - extracted code for getting a timestamp for the current system time to timestamp_current().
Jurko Gospodnetić
2012-07-14 17:28:50 +0000
e548f836d9Boost Jam's internal timestamp data structure now holds a member for nanosecond information. This information is still always set to 0 but, if set, is used when needed. Updated Boost Jam's header cache file format to hold the new extended timestamp information (bumped format version up to 5).
Jurko Gospodnetić
2012-07-14 17:17:09 +0000
ace6747054Boost Jam's internal timestamp data structure now holds a member for nanosecond information. This information is still always set to 0 but, if set, is used when needed. Updated Boost Jam's header cache file format to hold the new extended timestamp information (bumped format version up to 5).
Jurko Gospodnetić
2012-07-14 17:17:09 +0000
f18abd8674Boost Jam now reports its timestamps using a bit more complete format (still ISO-8601 compliant) to include nanosecond information. The new format is 'YYYY-MM-DD HH:MM:SS.ZZZZZZZZZ +0000'. Currently the nanosecond part value is always 0.
Jurko Gospodnetić
2012-07-14 16:36:56 +0000
3ca9fc5911Boost Jam now reports its timestamps using a bit more complete format (still ISO-8601 compliant) to include nanosecond information. The new format is 'YYYY-MM-DD HH:MM:SS.ZZZZZZZZZ +0000'. Currently the nanosecond part value is always 0.
Jurko Gospodnetić
2012-07-14 16:36:56 +0000
3736a64fd6Boost Jam code cleanup - wrapped up timestamp values inside a timstamp structure instead of using a raw time_t value as another preparation step towards implementing support for timestamps with resolution finer than 1 second. All timestamp manipulation now done using the new timestamp_XXX() API.
Jurko Gospodnetić
2012-07-14 16:11:16 +0000
6383f7f161Boost Jam code cleanup - wrapped up timestamp values inside a timstamp structure instead of using a raw time_t value as another preparation step towards implementing support for timestamps with resolution finer than 1 second. All timestamp manipulation now done using the new timestamp_XXX() API.
Jurko Gospodnetić
2012-07-14 16:11:16 +0000
8202d97670Boost Jam code cleanup - extracted Windows specific FILETIME structure conversion functions from execnt.c to a lowel level filent.c Windows specific module so they may be reused in both. Renamed the functions to make their purpose clearer.
Jurko Gospodnetić
2012-07-14 13:23:45 +0000
e8febc705cBoost Jam code cleanup - extracted Windows specific FILETIME structure conversion functions from execnt.c to a lowel level filent.c Windows specific module so they may be reused in both. Renamed the functions to make their purpose clearer.
Jurko Gospodnetić
2012-07-14 13:23:45 +0000
eb8c859a26Boost Jam execnt.c module code cleanup. No longer attempts to perform some operations like closing alert windows or reading their output (before the code depended on some operations not doing anything if given invalid handles or process ids). No longer needs to find out the process id from a process handle (since it actually already had these process ids all along), thus avoiding hacks needed to support this on Windows versions prior to Windows XP SP1 (where there was no GetProcessId() Windows API). Minor stylistic changes.
Jurko Gospodnetić
2012-07-14 13:10:00 +0000
9f5192acffBoost Jam execnt.c module code cleanup. No longer attempts to perform some operations like closing alert windows or reading their output (before the code depended on some operations not doing anything if given invalid handles or process ids). No longer needs to find out the process id from a process handle (since it actually already had these process ids all along), thus avoiding hacks needed to support this on Windows versions prior to Windows XP SP1 (where there was no GetProcessId() Windows API). Minor stylistic changes.
Jurko Gospodnetić
2012-07-14 13:10:00 +0000
fe47abe424Boost Jam code cleanup - removed a redundant #include directive.
Jurko Gospodnetić
2012-07-14 11:49:52 +0000
c4229062d2Boost Jam code cleanup - removed a redundant #include directive.
Jurko Gospodnetić
2012-07-14 11:49:52 +0000
8e716991c8Corrected Boost Jam's Windows FILETIME to POSIX time_t conversion. New implementation lifted shamelessly from the CPython implementation. The old implementation did not work correctly in all cases depending on the user's regional (timezone & daylight saving time) settings as well as the TZ environment variable (affects some C runtime library routines). It also used the double type to represent some interim values during its calculation - sometimes causing off-by-one rounding errors.
Jurko Gospodnetić
2012-07-14 11:42:58 +0000
69f101b623Corrected Boost Jam's Windows FILETIME to POSIX time_t conversion. New implementation lifted shamelessly from the CPython implementation. The old implementation did not work correctly in all cases depending on the user's regional (timezone & daylight saving time) settings as well as the TZ environment variable (affects some C runtime library routines). It also used the double type to represent some interim values during its calculation - sometimes causing off-by-one rounding errors.
Jurko Gospodnetić
2012-07-14 11:42:58 +0000
743cc7ed4bBoost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-14 11:35:29 +0000
ce0b0668b2Boost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-14 11:35:29 +0000
0a656de095Boost Jam code cleanup - cleaned up the timestamp.c module identifiers and better commented its functions.
Jurko Gospodnetić
2012-07-13 23:14:40 +0000
140242a5f5Boost Jam code cleanup - cleaned up the timestamp.c module identifiers and better commented its functions.
Jurko Gospodnetić
2012-07-13 23:14:40 +0000
fb8fb8317eBoost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-13 23:11:45 +0000
d7a0d333e4Boost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-13 23:11:45 +0000
ada97b6599Boost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-13 22:48:58 +0000
d4a061618bBoost Jam code cleanup - minor stylistic code changes & comment updates.
Jurko Gospodnetić
2012-07-13 22:48:58 +0000
b9fa41fecdBoost Jam code cleanup - minor stylistic changes.
Jurko Gospodnetić
2012-07-13 20:15:00 +0000
c284fac693Boost Jam code cleanup - minor stylistic changes.
Jurko Gospodnetić
2012-07-13 20:15:00 +0000
678fe54e15Updated Boost Build tests to run the collect_debug_info.py test only when running the tests non-interactively, i.e. when collecting the test results in an xml output file. When running the tests interactively and observing the results on the console there is no need for this always-failing test to spoil the 'all tests passed' result - the user will generally know everything about his system and can still run this test directly if he wishes to see what it would have returned.
Jurko Gospodnetić
2012-07-13 19:17:04 +0000
0afcd1285bUpdated Boost Build tests to run the collect_debug_info.py test only when running the tests non-interactively, i.e. when collecting the test results in an xml output file. When running the tests interactively and observing the results on the console there is no need for this always-failing test to spoil the 'all tests passed' result - the user will generally know everything about his system and can still run this test directly if he wishes to see what it would have returned.
Jurko Gospodnetić
2012-07-13 19:17:04 +0000
53576e4073Boost Build testing system code cleanup. Mainly stylistic changes + stopped using old deprecated string module functions in the test_all.py module.
Jurko Gospodnetić
2012-07-13 19:03:54 +0000
43a3331dc4Boost Build testing system code cleanup. Mainly stylistic changes + stopped using old deprecated string module functions in the test_all.py module.
Jurko Gospodnetić
2012-07-13 19:03:54 +0000
8bc155d307Boost Build testing system cleanup - separated the concept of the 'time when a test run finished' from 'highest timestamp assigned by a test run'. This is a tiny preparation step towards making Boost Build support timestamp resolutions finer than 1 second.
Jurko Gospodnetić
2012-07-13 18:43:18 +0000
6ebe5f9cdbBoost Build testing system cleanup - separated the concept of the 'time when a test run finished' from 'highest timestamp assigned by a test run'. This is a tiny preparation step towards making Boost Build support timestamp resolutions finer than 1 second.
Jurko Gospodnetić
2012-07-13 18:43:18 +0000
8c067504b8Make annoying unconditional hcache message only appear with DEBUG_HEADER.
Steven Watanabe
2012-07-13 01:53:46 +0000
59fbc6c52eMake annoying unconditional hcache message only appear with DEBUG_HEADER.
Steven Watanabe
2012-07-13 01:53:46 +0000
8bef878ab5Boost Jam cleanup - minor stylistic changes.
Jurko Gospodnetić
2012-07-12 13:36:31 +0000
4728e87bf0Boost Jam cleanup - minor stylistic changes.
Jurko Gospodnetić
2012-07-12 13:36:31 +0000
9905deabeeFixed a Boost Jam error reporting bug where it would report the error as originating from an incorrect file & line in case the error occurred after parsing the final token in some file. It usually reported it as the file including the file in question or, if the file is the main project build script not included from anywhere, then it reported the error as originating from file '(builtin)' and line -1. Added a related internal Boost Build unit test (core_source_line_tracking.py).
Jurko Gospodnetić
2012-07-12 13:21:46 +0000
6001461227Fixed a Boost Jam error reporting bug where it would report the error as originating from an incorrect file & line in case the error occurred after parsing the final token in some file. It usually reported it as the file including the file in question or, if the file is the main project build script not included from anywhere, then it reported the error as originating from file '(builtin)' and line -1. Added a related internal Boost Build unit test (core_source_line_tracking.py).
Jurko Gospodnetić
2012-07-12 13:21:46 +0000
343b4d6532Updated the internal Boost Build testing system to use the Python subprocess module (introduced in Python 2.4) for running external processes instead of popen2 (deprecated since Python 2.6). We are already using Python 2.4 features in this codebase so there is no need to support Python releases older than 2.4.
Jurko Gospodnetić
2012-07-12 12:55:25 +0000