| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Add Intel copyright to files that Intel has had non-trivial contribution | Thiago Macieira | 2016-01-21 | 1 | -0/+1 |
| | | | | | | | | | | I wrote a script to help find the files, but I reviewed the contributions manually to be sure I wasn't claiming copyright for search & replace, adding Q_DECL_NOTHROW or adding "We mean it" headers. Change-Id: I7a9e11d7b64a4cc78e24ffff142b506368fc8842 Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com> | ||||
| * | Updated license headers | Jani Heikkinen | 2016-01-15 | 1 | -14/+20 |
| | | | | | | | | | | | | From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/ Updated license headers to use new LGPL header instead of LGPL21 one (in those files which will be under LGPL v3) Change-Id: I046ec3e47b1876cd7b4b0353a576b352e3a946d9 Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com> | ||||
| * | QVersionNumber: compile with -Wzero-as-null-pointer-constant | Marc Mutz | 2015-09-13 | 1 | -1/+1 |
| | | | | | | | Change-Id: I1514864f1c7ae0d260aad368e2dc4de84061732c Reviewed-by: Keith Gardner <kreios4004@gmail.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> | ||||
| * | Add a quick optimization for QVersionNumber's constructors | Thiago Macieira | 2015-07-30 | 1 | -1/+11 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If the data fits inline, let's store it using the dummy member. GCC, Clang and ICC optimize the code on all architectures I tested to one single store. Previously, the function for "return QVersionNumber(5,4,0);" was: x86-64: movb $7, (%rdi) movb $5, 1(%rdi) movb $4, 2(%rdi) movb $0, 3(%rdi) x86: movb $7, (%eax) movb $5, 1(%eax) movb $4, 2(%eax) movb $0, 3(%eax) ia64: addl r17 = 7, r0 adds r16 = 1, in0 adds r15 = 2, in0 adds r14 = 3, in0 st1 [in0] = r17 addl r17 = 5, r0 ;; st1 [r16] = r17 addl r16 = 4, r0 ;; st1 [r15] = r16 st1 [r14] = r0 armv7a: mov r1, #7 mov r2, #5 strb r1, [r0] mov r1, #4 strb r2, [r0, #1] mov r2, #0 strb r1, [r0, #2] strb r2, [r0, #3] mips32: li $3,7 # 0x7 sb $3,3($4) li $3,5 # 0x5 sb $3,0($4) li $3,4 # 0x4 sb $3,1($4) sb $0,2($4) mips64: li $3,7 # 0x7 sb $3,7($4) li $3,5 # 0x5 sb $3,0($4) li $3,4 # 0x4 sb $3,1($4) sb $0,2($4) ppc32: li 10,7 stb 10,3(3) li 10,5 stb 10,0(3) li 10,4 stb 10,1(3) li 10,0 stb 10,2(3) ppc64: li 10,7 stb 10,7(3) li 10,5 stb 10,0(3) li 10,4 stb 10,1(3) li 10,0 stb 10,2(3) Now it is: x86-64: movq $263431, (%rdi) x86: movl $263431, (%eax) ia64: addl r14 = 263431, r0 ;; st8 [in0] = r14 armv7a: movw r3, #1287 movt r3, 4 str r3, [r0] mips32: li $3,84148224 # 0x5040000 addiu $3,$3,7 sw $3,0($4) mips64: li $3,321 # 0x141 dsll $3,$3,50 daddiu $3,$3,7 sd $3,0($4) ppc64: lis 9,0x504 sldi 9,9,32 ori 9,9,7 std 9,0(3) ppc32: lis 9,0x504 ori 9,9,7 stw 9,0(3) All assembly listings from GCC 4.8.1, but the Clang and ICC outputs are identical or at least very similar (I tested Clang for ARM, MIPS and PowerPC). Both MIPS and PowerPC were compiled in big-endian mode and this listing shows that the 64-bit implementation is correct. Additionally, the output is also the same for GCC when using brace initialization (that is, return QVersionNumber{5,4,0}). Clang and ICC couldn't optimize that. Change-Id: I9a4a4c9fc83f1182401f63fd2da829c935a8c9da Reviewed-by: Keith Gardner <kreios4004@gmail.com> Reviewed-by: Marc Mutz <marc.mutz@kdab.com> | ||||
| * | Refactor QVersionNumber so it stores values in-class | Thiago Macieira | 2015-07-29 | 1 | -10/+157 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | The common case of QVersionNumber is that there are few segments and each segment is a small integers. So instead of allocating a QVector<int>, just store those numbers in the class itself if possible. Think of this as a "Small String Optimization" for QVersionNumber. QVector<int> costs 16 + 4*N bytes, plus malloc overhead. After this change, QVersionNumber(5,4,0) will have an overhead of zero. The memory layout is explained in the header. I've coded it so big endian also works, but I have not tested it at all. Aside from the special functions for QVersionNumber and operator>>, all the rest of the algorithm could have been left unchanged. I only updated segments(), normalized(), compare(), commonPrefix() and fromString() to take advantage of the smaller implementation in a more efficient way. Note: QVersionNumber's constructors often leave half of the object or more uninitialized. That's not a problem. Change-Id: I4a2a0ce09fce2580f02d678e2f80b1dba74bac9d Reviewed-by: Marc Mutz <marc.mutz@kdab.com> | ||||
| * | Remove the ref-qualified versions of segments and normalized | Thiago Macieira | 2015-07-16 | 1 | -40/+3 |
| | | | | | | | | | | | They can't be ref-qualified if the QVersionNumber object doesn't actually hold a QVector<int>, as the next commit will make it: for segments(), there might not be a QVector to be moved; for normalized(), the common case will be that there's no gain in ref-qualifying. Change-Id: I4bfb8b8765a502c0de6aed693752217106e575a2 Reviewed-by: Marc Mutz <marc.mutz@kdab.com> | ||||
| * | Refactor QVersionNumber to use the public functions when possible | Thiago Macieira | 2015-07-16 | 1 | -2/+2 |
| | | | | | | | | | | Use segmentCount() and segmentAt() instead of going to m_segments. This is done in preparation for a major refactor of QVersionNumber that will store the version numbers in the object itself, without QVector. Change-Id: I03dbdee59a3c74c21a0a4e70c1bb9182250f6223 Reviewed-by: Marc Mutz <marc.mutz@kdab.com> | ||||
| * | Added QVersionNumber to QtCore's public API | Keith Gardner | 2015-07-10 | 1 | -0/+193 |
| | | | | | | | | | | [ChangeLog][QtCore] Added QVersionNumber. Change-Id: I11acc1fae3dc9368a72593afcfa2e462c53a620e Reviewed-by: Marc Mutz <marc.mutz@kdab.com> Reviewed-by: Keith Gardner <kreios4004@gmail.com> Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com> | ||||
| * | Make QVersionNumber private | Thiago Macieira | 2014-11-13 | 1 | -201/+0 |
| | | | | | | | | | | | We're not ready. [ChangeLog][EDITORIAL] Remove all mentions of QVersionNumber. Change-Id: I03ad95992982eb3177f982c1eeddb6a6bc29336c Reviewed-by: Keith Gardner <kreios4004@gmail.com> Reviewed-by: Lars Knoll <lars.knoll@digia.com> | ||||
| * | Long live QVersionNumber! | Keith Gardner | 2014-08-09 | 1 | -0/+201 |
| The class provides compare operators, stream operators, and hashing functions. This class aims to be compatible with (but not restricted to) the Semantic Versioning 2.0 standard (semver.org). [ChangeLog][QtCore] Added QVersionNumber class Done-with: Marc Mutz <marc.mutz@kdab.com> Change-Id: I244c8ccc002909af03987a2df052734d1a8621a9 Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com> | |||||
 |
index : qt/qtbase.git | |
| Qt Base (Core, Gui, Widgets, Network, ...) |
| summaryrefslogtreecommitdiffstats |