Consider the following source tree:
foo/narf.cpp
bar/narf.c
bar/gnampf.cpp
The .pro file has
SOURCES += foo/narf.cpp bar/gnampf.cpp
The file bar/narf.c is not supposed to be built for whatever reason.
QMake's nmake Makefile generator generates inference rules of the form
{.\foo}.cpp{debug\}.obj::
...
for every source subdirectory and every source file extension.
Thus, we have
{.\foo}.cpp{debug\}.obj::
{.\bar}.cpp{debug\}.obj::
{.\bar}.c{debug\}.obj::
Depending on the exact execution order of the inference rules (which
depends on the names of the files) the latter rule might get picked,
and we're erronously compiling bar/narf.c even though it's not
referenced in the .pro file.
Conclusion: QMake's detection of conflicting source files must
consider the base names of source files, and not the exact file names.
Fixes: QTBUG-72059
Change-Id: I50c2725ae2a7421053369a10680230f571af00ea
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
'=' cannot be handled in the same manner as other "critical" characters
as no amount of backslashes will escape it. Use a variable instead.
The documentation for nmake suggests that '=' in file names is not among
the "Special Characters in a Makefile". Therefore, we assume nmake can
handle it and don't escape it.
Fixes: QTBUG-67262
Change-Id: Ib60f808d7d4e981c98f7d8bf2075d55b2b7f3b7d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Michael Brasser <michael.brasser@live.com>
The read from a QHash needs to be protected too if other threads are
writing.
sync-up with qtc, no actual effect on qmake itself.
Fixes: QTCREATORBUG-21416
Change-Id: I75e5634e11b10056d6dbb6fdceef482ca2222ca1
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
(cherry picked from qtcreator/5f79b5d2e5e33321cdcd00362f0d6d9442a73ec2)
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Allows a qt build to be configured to target arm64 desktop apps cross
platform and build them with nmake.
Change-Id: I99fed12047b45a504a1644201bcc19b18c69f3e6
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
The assert still happens on MSVC 2015 64 bit when running
qmake -tp vc -r.
This reverts commit f4169a633b.
Fixes: QTBUG-71228
Change-Id: I05bd3e0677414edb970f07e0555cdc95ce32f592
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Also clean up QTextCodec usage in qmake build and some includes
of qtextcodec.h.
Change-Id: I0475b82690024054add4e85a8724c8ea3adcf62a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The correct name is c++1z. Anyhow, this is easy enough to get wrong,
so make sure CONFIG += c++17 works as well.
Task-number: QTBUG-67527
Change-Id: Iea26b18824b38b1b5170f85987cf5c750b8e10ab
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: André Hartmann <aha_1980@gmx.de>
it must end up in front of QMAKE_LIBS{,_PRIVATE}, but not of
LIBS{,_PRIVATE} (which are preceded by QMAKE_LIBDIR).
Task-number: QTBUG-61982
Started-by: Liang Qi <liang.qi@qt.io>
Change-Id: Id3de01ee0e9b66af02f79949aeb5a0eabd55363f
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
the early merging of LIBS* into QMAKE_LIBS* meant that we could not
interleave them properly. defer the merging until the points of use.
Task-number: QTBUG-70779
Started-by: BogDan Vatra <bogdan@kdab.com>
Change-Id: I890f98016c3721396a1f0f6f149a9e2b37d56d8e
Reviewed-by: Liang Qi <liang.qi@qt.io>
it's unused now, and just complicates matters. its interaction with
LIBS_PRIVATE & co. has always been a bit shaky. google produces no
public hits outside qt itself, so let's assume it really remained
internal.
Change-Id: I6606bbabd44f1b76d84e97219e155e38d6f1b3a6
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
this was introduced in 2002 supposedly for qnx4, but doesn't appear to
have actually been used ever. remove it, as it's in the way now.
Change-Id: I54dcabb61e1d3609a1e7a9fa4ff4b25509cfdb7a
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
lld for coff/mingw doesn't support linker scripts, which qmake used
for passing larger numbers of input file names to the linker.
Instead of using a fullblown linker script for this, just use a plain
response file, which both lld and binutils ld support.
Change-Id: I3aace7902fa6ca861a0a9fe67feaa236e7ea417b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Remove the DefaultCharIsUnsigned member and XML property because the
MSVC project file no longer has a specific XML tag for that setting;
instead, the /J option should go directly into the AdditionalOptions
tag.
Task-number: QTBUG-69611
Change-Id: I8f386427b7384ae09553f66193a2c9ea2b98dbc8
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The fix was for MSVC2010 which is now no longer supported
and causes a warning:
msvc_vcproj.cpp(217,12): warning: unknown pragma ignored [-Wunknown-pragmas]
#pragma optimize( "g", off )
by clang-cl.
This reverts commit 9f54846d95.
Task-number: QTBUG-63512
Change-Id: I395b964d09ac04a7dc852e43d7d95669948eea52
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
instead of trying to reverse-engineer it from the final target including
extension and possible bundle path, construct the basename explicitly.
this avoids that we mangle the filename if the actual target contains a
period for some reason.
Task-number: QTBUG-70097
Change-Id: I0bae9f010ab82e258680830250f8e28656f09d67
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
turns out that flushing the ids together with the ProFile cache was an
abysmal idea, as the latter expires after a few seconds after loading
the project, while references to the ProFiles (and thus the underlying
file ids) are kept for as long as the project stays loaded. the early
flush would cause re-use of the file ids, which would wreak all kinds of
havoc when re-loading projects.
but just ref-counting the vfs class is insufficient as well, as then the
ProFile cache (which expires after a timeout) could outlive all VFS
instances and thus refer to ids which were discarded and later re-used.
so we ref-count, and additionally let the cache instance hold a
reference to the vfs class.
this is sync-up with qt creator; no actual effect on qmake itself.
amends 190aa94be.
Change-Id: Idd4478ffc1c2405b3b83df32fba45b1f687f6a18
Reviewed-by: Robert Loehning <robert.loehning@qt.io>
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>
(cherry picked from qtcreator/d03fb350672d311dccc06f0bcb4da744a3c99745)
(cherry picked from qtcreator/1ddfb443b686ef04cc0e28363308ce70d01f0d73)
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
For single config mode (no debug_and_release) extra compilers wouldn't
get added to the vcxproj file.
Single config mode creates a temporary project, and that was incomplete.
Multi config mode, on the other hand, directly operates on the "real
data" and wasn't affected by this problem.
Task-number: QTBUG-69769
Change-Id: I9cd942e43d80adbeac9a3c8fbe1a5766bc9645a0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
A switch statement using enumeration type midlErrorCheckOption as
condition had an enumeration value from type midlStructMemberAlignOption
as a case label. This had only coincidentally the intended effect, since
the intended value (midlErrorCheckOption::midlEnableCustom) and the
actually used one
(midlStructMemberAlignOption::midlStructMemberAlignOption) have both
the value 0.
Change-Id: I73b337f23e733a1a6fb80517e29365e01838238a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Generated C++ source files with extensions other than .cpp would not get
compiled, because the code ignored all other C++ source extensions like
.cc and .cxx.
Fix this by respecting the value of QMAKE_EXT_CPP and QMAKE_EXT_C.
Task-number: QTBUG-69770
Change-Id: I097dfef6920e353a351c97891cdbfdc9a859815f
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
qt creator's clang code model is a bit more picky than the old one, so
we need a project that approximately works.
while we're at it, inline qmake.pri, add some missing files, and
beautify the source lists.
Change-Id: I87ca1db2ee3e55ea08e4c23f7913e882ab44fd21
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Some distributions do not define MINGW_HAS_SECURE_API globally,
resulting in methods like wgetenv_s not being declared in the
headers.
This is probably to keep compatibility with Windows XP. Anyhow,
we don't support Windows XP anymore, so we can safely add the
define.
Note that this is not necessary for the mingw-builds distro,
which is the only one we test and support. Anyhow, I don't
see any risk in adding these for other distributions.
Diff was provided by Philippe Dunski in the bug report.
Task-number: QTBUG-67443
Change-Id: I3a64b11541fe95e527ed44bbf8ad94469d457d3d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
We now treat -o foo/bar/baz as a request to generate the output in the
foo/bar directory with baz as the output name, or if foo/bar/baz is already
a directory, in the foo/bar/baz directory with the default output name.
We take care to handle generator specific directory structures, so
that the project directory does not get merged into OUT_PWD. This is
done in runQmake(), before parsing the project file, so that OUT_PWD
will be correct during project parsing. The individual generators are
then passed the filename relative to the final output directory.
Each generator now also makes sure to add the right project suffix
to the output file, so -o foo will result in foo.pro or foo.vcproj,
instead of just foo.
Task-number: QTBUG-44408
Change-Id: I26990cec0c0458bee2b88dbb86322617a85f54b5
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Makes it a bit more clear why all the Xcode settings were lost.
Task-number: QTBUG-45113
Change-Id: I3b19edb02a24673f56e77d3a1fb7cc76584c73fd
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
CONFIG+=lrelease enables that all .ts files in
TRANSLATIONS or EXTRA_TRANSLATIONS are compiled by
lrelease.
EXTRA_TRANSLATIONS is a new variable that is only
processed by lrelease, but not lupdate - this
is useful for translation files that are supposed to
be empty, because they match the language of the
original translation sources.
If embed_translations is also set, the generated .qm
files will be made available through the Qt resource
system under :/i18n/. Alternatively, the user can
specify an installation target by setting
QM_FILES_INSTALL_PATH.
Note that relative paths in TRANSLATIONS are not taken
into account. That is,
TRANSLATIONS = component1/de.ts component2/de.ts
will cause a conflict.
[ChangeLog][qmake] New CONFIG options lrelease and
embed_translations were added. CONFIG+=lrelease does
run lrelease on translation files listed in TRANSLATIONS
and EXTRA_TRANSLATIONS. CONFIG+=embed_translations does
include the generated .qm files as resources under
:/i18n/.
Change-Id: I94db5b8431d07b24f59b2c332ede91450f9c0c58
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
this allows for dynamic generation of the dependencies.
Task-number: QTBUG-61267
Change-Id: If5b8aed6b9e4bde189cc3ba6a5f13dcf8def3a1e
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
It's 160ish lines and adequately isolated. Still, it *might* be
contributing to compilers being slow (unlikely though that seems) so
seemed worth tidying up anyway.
Task-number: QTQAINFRA-2117
Change-Id: I9e55e677552c273fdf3480ccefc229fd6fd2b66a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The Xcode project name may be affected by e.g. the -o argument to qmake,
so we can't assume it's based on the target.
Change-Id: Ibb9f4265017ffcfe26bd8734758dcb30237c704f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This reverts the following commits:
d12d2949d126c3bec09b49b08f96e8
We can't easily predict all code paths for QDesigner
with such a microoptimization. We also don't want
to generate three different string constructions
depending on some sophisticated heuristics.
[ChangeLog][uic] The -no-stringliteral option is now deprecated and
UIC will not generate QStringLiteral anymore.
Task-number: QTBUG-65251
Task-number: QTBUG-51602
Change-Id: I34a5a1934a8df19c5c84ac2ba8e5168ce5665037
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
the variables are quoted correctly for commands, which is incompatible
with quoting for dependencies under mingw. so insert the paths as
literals, where we can control quoting.
this fixes building in directories with spaces, which i broke in
7c34e0a7b by using different quoting styles for deps and commands in the
first place.
this breaks the hypothetical use case where somebody wants to override
TARGET or DESTDIR (or DESTDIR_TARGET under windows) on the *make*
command line. not sure why anyone would do that - just do it at the
*qmake* level.
we did not get rid of OBJECTS, because that would cause significant
duplication in the makefile (not that it would matter too much, given
the dependency lists ...). this isn't a problem, because these are
short relative paths which are not expected to contain "funny"
characters.
an alternative would have been to change the variables' quoting and
eliminate them from the commands instead, but that would be
backwards-incompatible, because commands are "user-servicable".
for the same reason, we cannot get rid of the variables entirely.
Change-Id: Ic7592c7fc67d8b7d2b64de80808365cd1c3f79d0
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
under windows, libraries can have a numeric suffix derived from VERSION,
and (under MinGW) a unix-like "lib" prefix - neither of which .prl files
have. therefore, we had to make the back-mapping from the library to the
.prl file reverse-engineer the original TARGET's name. we verify whether
we actually got the right file by comparing the target specified inside
the .prl file with what we started from.
this fixes linking of transitive deps of static deps.
the alternative of changing the .prl naming pattern to avoid the
back-mapping was discarded, as a) it would be backwards incompatible and
b) it would break project-internal -lfoo references to versioned libs.
Change-Id: Ia9b899fe6a5700fee528bd1dacf130caf083cdd6
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
