Compile of abs(long long) fails on ARM processor

[parrisimo@…]

Error: transcode.cpp:541: error: call of overloaded 'abs(long long int)' is ambiguous

Fix: Changing abs to llabs allows the compile to be successful.

Info: Another user reported a similar problem on ​Ticket 1885, but ​fixed problem by adding special function MyAbs?. I don't know which solution is preferred by the developers.

[m Parris <parrisimo@…>]

patch changing abs to llabs

[Gary Buhrmaster <gary.buhrmaster@…>]

This issue was discussed in ​http://www.gossamer-threads.com/lists/mythtv/dev/522106

I agree with Lawrence Rust regarding compiler/platform idiosyncratic behaviors, and would instead suggest that using the Qt function here is likely to be the most portable solution. Attached is an alternative patch using qAbs (all uses of abs changed).

[beirdo]

At this time, we do not specifically support ARM processors. What platform are you compiling for, and with what compiler tools versions, etc?

With no way to verify this, it's probably not going to be immediately viewed as terribly high priority as we have too many other platforms to support as it is. However, it would be nice to get it working.

[beirdo]

Kinda a feature request... Anyways, please do not set the milestone.

[danielk]

Gavin, I think qAbs() makes sense for long long, due to current state of flux with long long in C99 vs C++03 vs C++11. i.e. it is required in C99 and C++11, but not in C++03 which is our current minimum C++.

[Gavin Hurlbut <ghurlbut@…>]

In 13163358dd54687628233ce4285f09435c0066c2/mythtv:

Remove the use of abs(long long) as it freaks out ARM compilers

Rather than using qAbs or llabs, it's simple enough in the specific situations
to just avoid using abs() altogether.

Fixes #10915