As Alex noted in Bug 164527, Comment 40, the string specifications in format strings are not standard and don't build with non-Cocoa SDKs or even with Cocoa SDKs that are not recent enough.
Created attachment 296645 [details]
I thought about using "PRIs" instead of "STRING_SPECIFICATION" (as an analog to "PRId8", et al), but wasn't sure everyone would get the reference. Also, I've been encouraged by multiple parties to spell out my symbol names, so I went with "STRING_SPECIFICATION". But if there's support for "PRIs" (or perhaps "PRIsPUBLIC"), I could use that instead.
Comment on attachment 296645 [details]
View in context: https://bugs.webkit.org/attachment.cgi?id=296645&action=review
We do use a lot of PRI* in places like NetworkCacheStatistics.cpp, but PRIs isn't standardized and I think this is indeed better.
> +#define DEBUG_LOG(format, ...) RELEASE_LOG(Network, "#PLT: %p - " STRING_SPECIFICATION "::" STRING_SPECIFICATION ": " format, this, DEBUG_LOG_EXPAND_AND_QUOTE(DEBUG_CLASS), __FUNCTION__, ##__VA_ARGS__)
http://www.cplusplus.com/reference/cstdio/printf/ calls this a "specifier" not a "specification". Let's call them specifiers. If they're called specifications somewhere I'm not familiar with, that's ok, too.
Created attachment 296664 [details]
Comment on attachment 296664 [details]
Clearing flags on attachment: 296664
Committed r209625: <http://trac.webkit.org/changeset/209625>
All reviewed patches have been landed. Closing bug.