July 04, 2010 [phobos] cacheIsBack and general policy suggestions | ||||
---|---|---|---|---|
| ||||
The introduction of cacheIsBack in std.algorithm.Map makes the unittests fail in release mode. I'm investigating the reason. Please ALWAYS build unittest in both debug and release mode prior to committing. This can be easily done on Unixen by running make -f linux.mak unittest Also, all data members of Map have a trailing underscore but cacheIsBack doesn't. I'm not hung on any particular rule but mixing two conventions in the same module, let alone the same type, is just bad. That goes for a bunch of "if(" added in a module that used "if (" everywhere. Please ALWAYS follow the conventions of at least the module you're changing. Andrei |
July 04, 2010 [phobos] cacheIsBack and general policy suggestions | ||||
---|---|---|---|---|
| ||||
Posted in reply to Andrei Alexandrescu | Are you sure this isn't some really obscure Linux-specific DMD bug? The std.algorithm unittests pass for me on Windows in both release and debug mode.
On 7/4/2010 4:23 PM, Andrei Alexandrescu wrote:
> The introduction of cacheIsBack in std.algorithm.Map makes the unittests fail in release mode. I'm investigating the reason.
>
> Please ALWAYS build unittest in both debug and release mode prior to committing. This can be easily done on Unixen by running
>
> make -f linux.mak unittest
>
> Also, all data members of Map have a trailing underscore but cacheIsBack doesn't. I'm not hung on any particular rule but mixing two conventions in the same module, let alone the same type, is just bad. That goes for a bunch of "if(" added in a module that used "if (" everywhere.
>
> Please ALWAYS follow the conventions of at least the module you're changing.
>
>
> Andrei
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
>
|
Copyright © 1999-2021 by the D Language Foundation