Custom Query (301 matches)
Results (112 - 114 of 301)
| Ticket | Resolution | Summary | Owner | Reporter |
|---|---|---|---|---|
| #150 | wontfix | MOC output on "No relevant classes found" | ||
| Description |
This is a rather weird one, but I'm asking for your opinion on that... When MOC processes a file without any "mocable" classes, it creates a zero-length output file. Unfortunately, if the target drive is on a NetWare? server, the timestamp of such a zero-length file is not updated (which is not MOC's fault). Anyway, as a result of that MAKE will re-invoke MOC for those classes - even though nothing has changed - and then cause the empty MOC result to be compiled. Due to this the target will be re-linked, which can be quiet time consuming. A simple workaround for that problem is to always write something into the MOC result files. A carriage return would be sufficient, however I'm wondering why they don't put the standard "this is a MOC generated file blah blah" comment header in... |
|||
| #151 | fixed | QFile::copy fails | ||
| Description |
Not sure if it makes sense to add a ticket for a problem that is only present in 4.5.1 and that is fixed in 4.6.2 but I just want to mention it. The QTemporaryFile class in 4.5.1-GA lacks the OS/2 specific changes. That leads to QFile::copy() not working correctly. |
|||
| #152 | fixed | Apply performance options | ||
| Description |
Since we need to use any piece of performance we can get and we don't have to support i386 CPUs, I suggest to let qmake generate the compiler options (for GCC4.x): -march=i486 -mtune=generic Don't expect wonders from that, but a small speed gain can be measured. The down side is, that the created code grows a bit in size. |
|||
