summaryrefslogtreecommitdiff
path: root/devtools
Commit message (Collapse)AuthorAgeFilesLines
* identify_failing_build_options.py: pep8 cleanup.Gary E. Miller2018-07-261-0/+1
| | | | Note the program pep8 is now called pycodestyle
* ais.py: pep8 cleanup. No functional change.Gary E. Miller2018-07-261-3/+12
| | | | Note the program pep8 is now called pycodestyle
* tablegen.py: pep8 cleanup.Gary E. Miller2018-07-261-0/+1
| | | | Note the program pep8 is now called pycodestyle
* ais.py: Fix for PEP8Gary E. Miller2018-06-201-308/+379
|
* tablegen.py: Fix for PEP8Gary E. Miller2018-06-201-44/+61
|
* uninstall_clean.py: Fix for PEP8Gary E. Miller2018-06-201-3/+6
|
* test_json_calidity.py: Fix for PEP8Gary E. Miller2018-06-201-1/+2
|
* Adds comment to regress-builder about its current brokenness.Fred Wright2016-09-221-0/+8
| | | | | | | | | | | This is the last remaining included Python program that isn't compatible with Python 3, and it doesn't make sense to fix that unless someone is motivated to fix its general outdatedness. TESTED: No. Signed-off-by: Gary E. Miller <gem@rellim.com>
* Fixes cycle_analyzer for Python 3 (as well as other bugs).Fred Wright2016-04-162-35/+44
| | | | | | | | | | | | | | | | | | It appears that this program hadn't been used in a while, since it had bugs unrelated to Python 3. After fixing some bugs, it's able to run successfully on all the test logfiles, except that it chokes on the binary data in ac12_binary (due to a newline within the binary data), and on the known malformed sentence in triton400. In order to allow the utility functions in gps.misc to be imported, it adds a symlink from devtools/gps to gps. Due to some weirdness in Python 3 imports, there didn't seem to be any other way to make this work with uninstalled libraries, and in any case it allows the local libraries to take precendence over the system libraries. TESTED: Ran with all six supported Python versions on all daemon logfiles, with only the errors noted above.
* Fixes devtools/flocktest for Python 3.Fred Wright2016-04-161-23/+36
| | | | | | | | | These fixes are applied "mostly blind" due to the inability to run this "for real", but they're pretty straightforward. TESTED: Ran -? and (unsuccessful) -c with both Python 2 and Python 3. Otherwise untested.
* Fixes scons post-cleanup in devtools/sizes.Fred Wright2016-04-161-1/+1
| | | | | | | | | I'd missed that there were two places doing cleaning. This applies the same fix to the second as was previouly applied to the first. TESTED: Ran and verified that the .sconf_temp directory is no longer left around.
* Adds a tool for cleaning up after Python uninstall bug.Fred Wright2016-04-161-0/+154
| | | | | | | | | | | | | | Since the just-fixed SConstruct bug could have left empty Python package directories lying around in multiple Python installs, this provides a tool for cleaning them all up. It applies to every non-text python* file in the command PATH, and removes any empty 'gps' package directories. Nonempty directories are not removed. TESTED: Verified that directories left over by the uncorrected uninstall were removed, and that nonempty directories weren't. Tested with both Python 2 and Python 3.
* Fixes shebang line in devtools/striplog.Fred Wright2016-04-141-1/+1
| | | | | | | | This was using the less reliable non-env version of the Python shebang line. The fix was neglected in the previous commit. TESTED: Verified that it launches and passes text.
* Fixes more devtools programs for Python 3.Fred Wright2016-04-144-12/+38
| | | | | | | | | | | | | | This fixes the aivdmtable, regressdiff, sizes, and striplog programs to work with Python 3. The "striplog" program didn't require any actual changes, but it still adds the usual compatibility comment and future import. In the case of "sizes", it also adjusts the scons cleanup pattern to include the temporary directory, and adds explicit checking (with an exception) for build failures. TESTED: Ran all four programs with both Python 2 and Python 3.
* Fixes devtools/ais.py to use new-style classes.Fred Wright2016-04-111-4/+4
| | | | | | See previous commits for the rationale. Not tested.
* Fixes devtools/*.py for Python 3.Fred Wright2016-04-104-98/+139
| | | | | | | | | | | | | This applies the 2to3 fixes as well as integer division fixes to the four programs in devtools. Most of these transformations are similar to ones used elsewhere, with the notable exception of the reraise_with_traceback() function in ais.py. TESTED: Ran identify_failing_build_options.py with both Python 2 and Python 3 up through length 1, with identical results. Did *not* test the other three programs. Someone more familiar with their usage should do that.
* A few fixes and improvements for the broken-build checker.Fred Wright2016-03-261-6/+23
| | | | | | | | | | | | | | | | | | | | | 1) Adds the recently added 'gpsd' and 'gpsdclients' to the test list. 2) Uses 'build-all' rather than 'build', for more complete testing. 3) Fixes a potential ordering issue with the pre-clean. 4) Makes the removal of scons temporaries more complete. 5) Derives the number of jobs from the number of CPUs, rather than hard-coding it. 6) Applies parallelism to "check' as well as 'build-all'. 7) Makes the error-list file names more descriptive. TESTED: Runs and gets a couple of failures, which don't appear to be new problems.
* Eliminate duplicate code when running the build command.Beat Bolli2016-02-201-21/+13
| | | | | Signed-off-by: Beat Bolli <dev@drbeat.li> Signed-off-by: Jon Schlueter <jschlueter@redhat.com>
* Eliminate a temporary variable.Beat Bolli2016-02-201-2/+1
| | | | | | | Iterating over a generator function is preferred and saves memory. Signed-off-by: Beat Bolli <dev@drbeat.li> Signed-off-by: Jon Schlueter <jschlueter@redhat.com>
* Move loop invariants in front of the loop.Beat Bolli2016-02-201-2/+3
| | | | | Signed-off-by: Beat Bolli <dev@drbeat.li> Signed-off-by: Jon Schlueter <jschlueter@redhat.com>
* Simplify the preparation of the build parameters.Beat Bolli2016-02-201-11/+4
| | | | | Signed-off-by: Beat Bolli <dev@drbeat.li> Signed-off-by: Jon Schlueter <jschlueter@redhat.com>
* Add xgps to devtools/identify_failing_build_options.pyJon Schlueter2016-02-061-0/+1
|
* additional type 21 message with 368 bits long messageJon Schlueter2016-01-041-1/+1
| | | | | had to extend check to include extra size Also tweaked devtools/ais.py
* Move a test utility where it belongs, and document it.Eric S. Raymond2015-03-312-0/+28
|
* additional tweaks to identify_failing_build_optionsJon Schlueter2015-03-251-1/+1
| | | | | | | change guard for scons check to just require passing scons Now that we have some better guards on the individual check targets for what they require
* Add scons check testing for identify_failing_build_optionsJon Schlueter2015-03-251-0/+9
|
* Update identify_failing_build_options with build option changesJon Schlueter2015-02-241-1/+1
| | | | added ntp and removed oldstyle
* Fix bug reported on the dev list resulting from NTPSHMSEGS > MAX_DEVICES * 2.Eric S. Raymond2015-01-241-1/+1
| | | | | | As a side effect, simplify configuration slightly. All regression tests pass.
* Add identify_failing_build_options script to devtoolsJon Schlueter2015-01-232-0/+132
| | | | | | | | | | | | | | | | | | This script is a helper script to run through various combinations of build options as enumerated in the script looking for broken builds. This has helped identify several bugs and some other general cleanup in the code base. to run: > devtools/identify_failing_build_options.py it will run for as long as you let it trying to build different build configurations with minimal=on to find failing ones. It prints out hopefully useful status as well as stderr but discards stdout from the builds. Any failing build configurations are logged to failed_build_configs.txt and can be easily checked later to see why they failed and how to fix them.
* Add a script to generate test pairs for matrix inversion.Eric S. Raymond2015-01-082-0/+124
|
* Add do-build script to run tests.Greg Troxel2015-01-082-0/+24
| | | | | | | This script is almost too trivial to check in, but it leaves files named like CHECK.release-3.11-181-g2ab8763 with regression test output, which is helpful when trouble arises.
* Keep permissions 755.Bernd Zeimetz2014-01-271-0/+0
|
* Fix gpsd-debian-regressions.sh to recognize all new tests.Bernd Zeimetz2014-01-271-2/+2
|
* Add -j option to striplog.Eric S. Raymond2013-11-081-3/+16
|
* Test machinery improvements. raw-regress is now only partially broken.Eric S. Raymond2013-11-051-2/+8
|
* Test-machinery improvements. All regression tests pass. PPS is live.Eric S. Raymond2013-11-051-11/+19
|
* Document some conventions.Eric S. Raymond2013-10-061-0/+8
|
* Correct some code generation.Eric S. Raymond2013-10-061-7/+7
|
* Make vocabulary references in dump code.Eric S. Raymond2013-10-061-2/+20
|
* Generate structure name properly.Eric S. Raymond2013-10-061-0/+5
|
* Update a comment.Eric S. Raymond2013-10-061-1/+2
|
* Subtable parsing works.Eric S. Raymond2013-10-061-5/+22
|
* Modified table parsing - a step towards parsing vocabulary tables.Eric S. Raymond2013-10-061-13/+25
|
* Correctly generate JSON dumping code for trailing arrays.Eric S. Raymond2013-10-051-5/+8
|
* ind -> i as a loop variable name.Eric S. Raymond2013-10-051-4/+4
|
* Fixes to generation of JSON dump code for arrays.Eric S. Raymond2013-10-051-3/+3
|
* Fix a minor bug in boolean field generation.Eric S. Raymond2013-10-021-10/+10
|
* Derived-attribute generation for enumerated fields.Eric S. Raymond2013-10-021-0/+4
|
* Driver code generation for boolean fields.Eric S. Raymond2013-10-021-0/+2
|
* Change JSON-dumper code generation to match my previous commit.Eric S. Raymond2013-10-011-1/+5
|
View Lorry Controller Status