| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
An expected failure was inadvertantly "fixed" by a5f2a7820f. Disable the
run environment setup for this autotest.
Change-Id: I34032ae0bdf51e6bc80ac1902df1ab8e31ed1464
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
git rev-parse --git-dir prints a relative or an absolute path, depending
on whether you call it from the repository base directory or not. Our
code now accounts for that.
Change-Id: Iac5eeb0fc55fe83d6c588efbcf97cab6786a4f89
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Ola Røer Thorsen <ola@silentwings.no>
|
|
|
|
|
|
|
|
|
|
| |
While we did set up the dependencies between multiplexed products
correctly, the *module* (i.e. Export item) that we pulled in was always
the same one and corresponded to the wrong product variant in most
cases.
Change-Id: If42cf7e946689259c7433d81fd35d8798919eb15
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
| |
The functionality of qtquickcompiler is now in qmlcachegen.
Change-Id: I70c45f9db72895e2146cbf4bb7902b8db294626c
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
|
|
|
|
|
|
|
|
| |
In Qt 5.11, the features/qtquickcompiler.prf file is always present,
even if the QtQuick compiler is not.
Task-number: QBS-1299
Change-Id: Ibc7f925f1b1753d870ddad939fd510637c3b7642
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When building multiple configurations in one go we create one
QScriptEngine per configuration, each in its own thread, but nearly at
the same time.
It turns out that QScriptEngine's creation/destruction is not
thread-safe, which leads to all kinds of interesting effects in the
multiple configurations case.
We fix this by guarding creation/destruction of QScriptEngine with a
mutex.
Task-number: QBS-1308
Change-Id: Ie01733d5943f1fb89e1d25344c4ffa998095e394
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
| |
These can be exceedingly difficult to debug otherwise.
Change-Id: Ie3bc42fbd60e11be105fc23624c77a623dc5f1aa
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Some autotests need to execute processes built earlier in the same test.
With mingw, these need to find libgcc at runtime. Make sure it is
available, either by using "qbs run" or by adding a dependency on the
cpp module in the product that runs the application.
Change-Id: I7f000c3d604a3412049a271b53b5fed4cc7fc54e
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
- We must activate these rules only on the product that is the actual
aggregator.
- The lipo command must only consider artifacts of the multiplexed
variants of the same product.
Change-Id: I25dc42cb53d84d85786b2d7ff96297b7fd58a90d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of always using the fallback value, we instead always use the
standard value unless we know we're running an older toolchain
which does not support it. This alleviates the potential for differing
behavior in newer versions of compilers which may attempt to remain
compatible with earlier drafts of corresponding standards.
[ChangeLog] Always use standard values for -std= when possible
Change-Id: I61ff3ecd863caa24cb8fb54500e62dc310a2af02
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
|
|
|
|
|
|
|
|
| |
It makes no sense to try to run a disabled product.
Task-number: QBS-1306
Change-Id: I07e36bc8f787d702ed9caa21972537e588db57e8
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
| |
These do not conceptually belong to the product.
Change-Id: Ic55cf246fe38ba7ad03563b5f67a8622810dad7d
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
| |
There is no guarantee that we find it in the environment.
Change-Id: I58281b5936e0ddad51cbdbe0833f62c4dbb42c1d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
| |
It did not know about qbs.toolchainType.
Change-Id: I00bcd5518f81a65fa49153b75f1fd2fb69d37fd3
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
| |
As first attempted in 7f6ec1cc27.
Change-Id: Ib76dbdfc0a60b27d38d93a22c9556d54f07a3dac
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
| |
Users should get feedback if they mistyped a product name, for instance.
Task-number: QBS-1288
Change-Id: Icb57490d896d21c5ff9ed3940b53d9b18fe00a30
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
... for the case that auxiliaryInputs or explicitlyDependsOn contains
"installable". They must be considered in addition to
inputsFromDependencies, because they also match target artifacts of
dependencies.
Change-Id: Ib960001a8e67034d6d1382bacc385bdcd9ddfb99
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Also add an undocumented property to add arbitrary options. This one
acts as an emergency fallback in case we miss a new option in the
future.
Task-number: QBS-1295
Change-Id: I0ba81325d6975d8bb25cbd397c9daaaa12f685c2
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Andreas Bacher <andreas.bacher@meon-medical.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
| |
It used to fail when executed from a directory that had "-c" in the
name.
Change-Id: Ib779082b55236843988c236aa8479265169bbec7
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
|
|
|
|
|
|
|
|
| |
The standard is out now.
Task-number: QBS-1020
Change-Id: Ie8249e1c860d5ebc2f9dadf02551544ddaaf2868
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
| |
Change-Id: I6807c55dd9115c9bb0be2aa4d5b7adf5c5e40853
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- tools.jar and rt.jar are gone
- The -bootclasspath option is not allowed when targeting Java 9
- JavaCompiler.isSupportedOption() (erroneously?) returns 1 instead of 0
for -Xlint
- version number related changes:
- major version has jumped to 9
- javac prints the version number to stdout rather than stderr
- version number no longer contains a forth component
- The -source and -target options take only the major version
Change-Id: I0b23c8868db7aeaa06785ee43c1efd9fcb1bc9d2
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|
|
|
|
|
|
|
| |
The default JDK on the system cannot be generally expected to support
all known language versions.
Change-Id: Ia224150977da651257874801d6bc9cba852521d6
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|\
| |
| |
| | |
Change-Id: I2d6cac6f24144c2c7e463ef7eed2e2c109f50bbf
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The SOFT ASSERT was triggered, because the Depends items in
Export/Module had the type ModuleInstance instead of ModuleParameters.
Move the assignment of that item type to ItemReaderASTVisitor instead of
adjusting it in the ModuleLoader, because this ensures that we don't
have to traverse prototypes in the adjustment procedure.
Task-number: QBS-1287
Change-Id: Ie1a1080d40b8bebe5c205f8b34298469cde5bd46
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Parameters of Depends items within Export items were ignored.
Project {
Product {
name: "foo"
Export {
Depends { name: "dep"; cpp.link: false } // dep
}
}
Product {
name: "bar"
Depends { name: "foo" }
}
}
The instance of foo in bar has an Depends child depinst. The item
depinst was created by createChildInstances and is an empty item with
the prototype dep, which is the Depends item in foo's Export item.
Thus, extractParameters must look up properties in the deepest
prototype (dep), not in the Depends item itself (depinst).
Unfortunately, this fix triggers
SOFT ASSERT: varValue in ../../../../1.10/src/lib/corelib/language/evaluatorscriptclass.cpp:574
which will be fixed separately.
Task-number: QBS-1287
Change-Id: I56b5464aad2578bb5ddab2e831fc607d6634ac8d
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I6a3222228f822b88d860b5036f459f1a732dc6ac
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
After adding an external file dependency, we would never remove it from
the global list again, meaning the look-up table would contain stale
entries if an external header file was moved around. Because we still
checked the status of the artifact-local list of dependencies, the
result was that in such a situation, the files including such a header
would get recompiled on every build.
[ChangeLog] Fixed constant rebuilding after moving an external header
file.
Task-number: QBS-1285
Change-Id: Id764da7485dd540f2ff64bcd890e43723f18e6cd
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
We failed to account for the carriage returns in the jar output.
Change-Id: I2061a8ea306c64ce44039472f2840f71898c2097
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|\|
| |
| |
| | |
Change-Id: Iae805216e202e582923a70733caa842e8070cc7e
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Dependencies must be merged before their dependents, because the merging
process for the former sets up data to be consumed by the merging
process for the latter.
Task-number: QBS-1275
Change-Id: Ide9146add79bf2b4fcf80d7ce95ac1d9d4ac20b2
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I68445c362386c6c310ab098ce5ea6780c1232e42
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
|\|
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/lib/qtprofilesetup/qtprofilesetup.cpp
src/lib/qtprofilesetup/templates/QtModule.qbs
src/lib/qtprofilesetup/templates/core.qbs
Change-Id: I342608a47f1a90b53eb12cc14fc3ff116063d15f
|
| |
| |
| |
| |
| |
| |
| |
| | |
Prefix headers should follow precompiled headers on the gcc command
line. Testcase added.
Change-Id: I3dfba8cae9c894932acce5c0a67f16d7d072c94b
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In the following example the name of the derived product should be
"derived", but it was "base".
---BaseProduct.qbs---
Product {
id: baseProduct
property string foo: "base"
name: baseProduct.foo
}
--derived.qbs--
BaseProduct {
foo: "derived"
}
The id baseProduct still pointed to the Product item originally imported
from BaseProduct.qbs.
Fix the item pointers in the base id scopes when handling item
inheritance.
Task-number: QBS-1016
Task-number: QBS-1262
Change-Id: Id11b98aa6780f8bbbda86fda5a7d11a0e006d889
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Such files are not to be matched against Rule.inputs. This part was
forgotten in 8ca922793f.
Change-Id: I0a01b3642f4a16c09d189e8c2228205fd0e9962b
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Do not try to execute the virtual binary.
Change-Id: I19b0a9e7600b33cb42c4ab60f689dd4c3329930b
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
... by using "qbs run" instead of manually executing the applications.
Change-Id: I548f7721acff3aac6644944ad7436ad8bcfa1867
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
We forgot to do this for alternatives, meaning that in Properties items
in Export items, "base" would not be properly set up.
Change-Id: Ieafe4933696074469aaf497d4e153ca92fa8f73b
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|\|
| |
| |
| | |
Change-Id: I29a74f96381a1b77c6cb1959547314786690baa4
|
| |
| |
| |
| |
| | |
Change-Id: I6a8ff3557bbaa5e38f3df0f4334e5d0331349565
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Apparently this is not declared in y.tab.h on all platforms
(such as Fedora).
Change-Id: I1ae92a8cd415afaa623a34ac7765eb44e5235320
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We only need to guarantee that the object file contents are equivalent
if we explicitly requested a reproducible build. This does NOT imply
that a non-reproducible build is *guaranteed* to produce different
object files. This is the case on Fedora systems as uncovered by that
platform's Qbs packagers.
Change-Id: I8e46610f08a8b0c2b59ff84da94e9529c030baba
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Users can write "qbs.xyz" instead of "modules.qbs.xyz", but we only
applied that during module prototype loading, meaning that using that
syntax (and only that syntax) would not override values set in project
files, which was obviously quite confusing to users.
Now the two variants behave exactly the same.
Task-number: QBS-1244
Change-Id: Ie9c5de1ac42e33205b80335566e0721d61718947
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Only the values in Depends items of products had a scope, and that one
did not allow to access module instances.
Now the scope always is the surrounding item, which should be a sensible
one in all contexts.
Task-number: QBS-1253
Change-Id: Iedca2d98f82f71887373579f8e4dc91f057eded3
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
There are several contexts in which we need to find out whether a
property value of a built-in item is the default one that was created by
us when initializing the item.
Rather than using heuristics for that, we now set a flag on such values,
so we can always get the correct information.
Change-Id: Ieb0d87a423d6c6836a2c1d60ccf503fa0b4f801b
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
This gives better compatibility with STL anyways.
Change-Id: Ie7f25b8ab8ccf4b50c7ad8e35f6dd4b5ca4cc4a2
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The comments around the code that skipped these checks made little sense
and were likely just outdated.
Change-Id: I8ff714a4d717359dd8dde39b55a41d233f84fd08
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|
|\|
| |
| |
| | |
Change-Id: Iecddc2722d22d7c15d4b32efd676fa8e7c00e99f
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The names of environment variables are case-insensitive on Windows.
Convert them to upper case when handing out the environment, so that
users can count on the names being canonical.
Change-Id: Iba2ce768b7cd22270dd2d240d15fc1a78915e3a7
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
|