Summary: | kbibtex failed to build on ARM since qreal is float instead of double | ||
---|---|---|---|
Product: | [Applications] KBibTeX | Reporter: | Robin Lee <robinlee.sysu> |
Component: | Compiling/packaging | Assignee: | Thomas Fischer <fischer> |
Status: | RESOLVED FIXED | ||
Severity: | normal | ||
Priority: | NOR | ||
Version: | 0.6 | ||
Target Milestone: | --- | ||
Platform: | Other | ||
OS: | Linux | ||
Latest Commit: | http://commits.kde.org/kbibtex/51e64e0c22b6ff39d07843e34fe667ebf128c1a9 | Version Fixed In: | |
Attachments: |
Patch to fix the ARM build
Migrating StarRating from double to float |
Created attachment 95016 [details]
Migrating StarRating from double to float
Wouldn't a complete migration from double to float solve the problem as well? On x86/amd64, it works without a noticeable difference. How about ARM?
This patch does not only change all double/qreal cases to float, but also refactors percentForPosition into StarRating::Private.
Please test if this patch works on ARM, too.
The new patch looks great and also fixes the ARM build. Git commit 80ff29053cf40a739ae5b3447eee16d7819d9f66 by Thomas Fischer. Committed on 30/10/2015 at 17:50. Pushed by thomasfischer into branch 'kbibtex/0.6'. Migrating StarRating to float, solving ARM issues As reported by Robin Lee in KDE bug 353898, qreal is defined differently in Qt on different platforms. Code that compiles on x86 platforms fails on ARM. By consequently using float instead of qreal, this issue can be avoided. Beside fixing this bug, some code has been refactored by moving a function from the class's public interface (header) to a private class. M +15 -15 src/gui/widgets/starrating.cpp M +3 -4 src/gui/widgets/starrating.h http://commits.kde.org/kbibtex/80ff29053cf40a739ae5b3447eee16d7819d9f66 Git commit 273dbbd019aa035b80d6bbfccb948dc9fb76b2d5 by Thomas Fischer. Committed on 30/10/2015 at 17:55. Pushed by thomasfischer into branch 'kbibtex/0.7'. Migrating StarRating to float, solving ARM issues As reported by Robin Lee in KDE bug 353898, qreal is defined differently in Qt on different platforms. Code that compiles on x86 platforms fails on ARM. By consequently using float instead of qreal, this issue can be avoided. Beside fixing this bug, some code has been refactored by moving a function from the class's public interface (header) to a private class. M +15 -15 src/gui/widgets/starrating.cpp M +3 -4 src/gui/widgets/starrating.h http://commits.kde.org/kbibtex/273dbbd019aa035b80d6bbfccb948dc9fb76b2d5 Git commit 51e64e0c22b6ff39d07843e34fe667ebf128c1a9 by Thomas Fischer. Committed on 30/10/2015 at 17:56. Pushed by thomasfischer into branch 'master'. Migrating StarRating to float, solving ARM issues As reported by Robin Lee in KDE bug 353898, qreal is defined differently in Qt on different platforms. Code that compiles on x86 platforms fails on ARM. By consequently using float instead of qreal, this issue can be avoided. Beside fixing this bug, some code has been refactored by moving a function from the class's public interface (header) to a private class. M +15 -15 src/gui/widgets/starrating.cpp M +3 -4 src/gui/widgets/starrating.h http://commits.kde.org/kbibtex/51e64e0c22b6ff39d07843e34fe667ebf128c1a9 |
Created attachment 94991 [details] Patch to fix the ARM build Please refer to the patch. Use qreal instead of double Since qreal is float on ARM, using double here will fail to compile.