| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
| |
This is conceptually better (because each fragment should be
self-contained), and gives us better performance, as we no longer need to
load all explicit labels.
--
MOS_MIGRATED_REVID=118674470
|
|
|
|
|
|
|
|
|
|
| |
compilation inputs. As
of r264664, Clang doesn't require files in module maps to be present anymore if
the sources are instead available in the corresponding .pcm files.
--
MOS_MIGRATED_REVID=118670841
|
|
|
|
|
|
|
|
|
| |
This test has been flaky for a long time and does not test what it is supposed to test.
Fixes #248.
--
MOS_MIGRATED_REVID=118668487
|
|
|
|
|
|
|
| |
Constants.DEFAULT_PACKAGE_PATH.
--
MOS_MIGRATED_REVID=118666575
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118662172
|
|
|
|
|
|
|
| |
This is done by adding a default workspace suffix.
--
MOS_MIGRATED_REVID=118656979
|
|
|
|
|
|
|
| |
This check is awful and should be done somewhat differently, but since we are migrating to gRPC, there isn't much point in investing much in this.
--
MOS_MIGRATED_REVID=118656266
|
|
|
|
|
|
|
| |
parsing for it is not broken.
--
MOS_MIGRATED_REVID=118654985
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118654684
|
|
|
|
|
|
|
|
|
|
|
| |
Also remove a try-catch block, which was converting all configuration
creation errors to IllegalStateException.
This uncovered a bug in CppOptions, which wasn't adding host_grte_top to the
set of explicit labels.
--
MOS_MIGRATED_REVID=118654358
|
|
|
|
|
|
|
|
|
| |
map is generated.
R_FUTURE=cparsons
--
MOS_MIGRATED_REVID=118631953
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118623130
|
|
|
|
|
|
|
|
| |
Slight improvement, mirroring the updated java_library
'exports' explanation. Less deceptive.
--
MOS_MIGRATED_REVID=118608946
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118605600
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118598917
|
|
|
|
|
|
|
| |
DottedVersion and Platform types in Skylark.
--
MOS_MIGRATED_REVID=118593554
|
|
|
|
|
|
|
|
|
| |
modules are enabled.
Also warn when -fmodules-cache-path is explicitly set by the user (either in configuration or rule attributes)
--
MOS_MIGRATED_REVID=118590843
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
it more with respect to the new parse_headers feature, ideas I have about
configurable attributes etc.
I am trying to keep functional changes to a minimum. Functional changes
that couldn't really be easily avoided:
- Find correct rule for headers through filegroups
- Find correct target to build even if sources are missing. A proper error is
reported in the subsequent loading phase
Also remove SrcTargetUtil as it is no longer used.
--
MOS_MIGRATED_REVID=118589517
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Manual breaking lines was introduced in a49f0253 to ensure Bazel knows
the number of lines written, so the correct number of lines can be
erased when the progress bar is updated. As this is only an issue if curses
are in use, that wrapping as only added to the FancyTerminalEventHandler.
However, Bazel distinguishes between two uses of curses
- curses to add color to the output, and
- curses to control the cursor and erase outdated progress bars.
If the former is used, but not the latter, it is not necessary to wrap
lines; in fact, not doing so is more backwards compatible while still
keeping the line-eating bug fixed. So don't break lines in this case.
--
Change-Id: Ieef8ab9410a071c35609a395f45cd56653115426
Reviewed-on: https://bazel-review.googlesource.com/#/c/3170
MOS_MIGRATED_REVID=118587980
|
|
|
|
|
|
|
|
|
|
| |
In this way, we prevent the experimental UI from reporting
about the number of actions (actually 0) currently running.
--
Change-Id: Ic1c02d388f1928b8fe8d2d66dc14757b0213c43b
Reviewed-on: https://bazel-review.googlesource.com/#/c/3180
MOS_MIGRATED_REVID=118580580
|
|
|
|
|
|
|
|
| |
If a load statement is present in the WORKSPACE file, the
workspace name attribute will be lost from the WORKSPACE file.
--
MOS_MIGRATED_REVID=118577315
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118574071
|
|
|
|
|
|
|
| |
been migrated.
--
MOS_MIGRATED_REVID=118571367
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118568284
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118567315
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118565877
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118563271
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118561661
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118560010
|
|
|
|
|
|
|
| |
This is necessary because build-runfiles depends on the msys shared library, which (as per DLL lookup rules on Windows) is checked under the entries of PATH.
--
MOS_MIGRATED_REVID=118556684
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Broke bazel http://ci.bazel.io/job/Bazel/417/JAVA_VERSION=1.8,PLATFORM_NAME=linux-x86_64/console
by using a Guava 20 feature, while Gauva 20 is not yet released.
*** Original change description ***
Set -fmodules-cache-path to be rooted in the genfiles directory whenever modules are enabled.
Also warn when -fmodules-cache-path is explicitly set by the user (either in configuration or rule attributes)
--
MOS_MIGRATED_REVID=118555994
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bazel clean never runs any actions, hence any message
about running actions would confuse the user. Therefore,
signal that clean is a non-building command so that the
experimental UI knows not the report about progress of
building.
--
Change-Id: I193f199b052e990992359bc9efdbe34db397ac5f
Reviewed-on: https://bazel-review.googlesource.com/#/c/3125
MOS_MIGRATED_REVID=118554435
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118547283
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118532668
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118525960
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118523482
|
|
|
|
|
|
|
| |
RELNOTES:
--
MOS_MIGRATED_REVID=118518790
|
|
|
|
|
|
|
| |
--objc_generate_debug_symbols. These are delivered in distinct files, with an architecture identifier.
--
MOS_MIGRATED_REVID=118510432
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118500217
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118488877
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118488791
|
|
|
|
|
|
|
| |
Avoids spurious diagnostics about annotation processor generated class names not matching the file path.
--
MOS_MIGRATED_REVID=118488628
|
|
|
|
|
|
|
| |
JavaBuilder overwrites files when extracting source jars, silently ignoring duplicates. It it'd be better if it didn't, but until it's fixed turbine should have the same behaviour.
--
MOS_MIGRATED_REVID=118487939
|
|
|
|
|
|
|
| |
RELNOTES: Restrict file types allowed for "strings" attribute to ".strings".
--
MOS_MIGRATED_REVID=118479031
|
|
|
|
|
|
|
|
|
| |
modules are enabled.
Also warn when -fmodules-cache-path is explicitly set by the user (either in configuration or rule attributes)
--
MOS_MIGRATED_REVID=118478856
|
|
|
|
|
|
|
|
| |
This is so that we can access the build options during creation, which is a
step towards making the coverage module stateless.
--
MOS_MIGRATED_REVID=118475354
|
|
|
|
|
|
|
| |
This is mostly mechanical, cleaning up the vestigal marks of combining the value and key. The only mildly odd change is making the OverwritableConsumer and NonOverwritableConsumer package visible, for testing purposes.
--
MOS_MIGRATED_REVID=118474980
|
|
|
|
|
|
|
| |
":extra_entitlements".
--
MOS_MIGRATED_REVID=118473605
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=118473491
|
|
|
|
|
|
|
| |
This makes the late initialization of BuildView more obviously safe.
--
MOS_MIGRATED_REVID=118469655
|