Large modifiers (#32)

* npf__int_t and npf__uint_t typedef'd for large or normal. add macros to extract and writeback through specified types

* fix up readme, make all float forms just do %f

* bigger coversion buffer, more conformance tests

* writeback large modifier tests

* ignore .ii

* negative float conformance

* initialize val to calm gcc

.gitignore

@@ -7,6 +7,7 @@ cmake-build*
 
 # preprocessed / asm files
 *.i
+*.ii
 *.s
 
 # Object files

.ycm_extra_conf.py

@@ -5,7 +5,7 @@
 flags = [
     '-DNANOPRINTF_IMPLEMENTATION',
     '-DNANOPRINTF_USE_FLOAT_FORMAT_SPECIFIERS=1',
-    '-DNANOPRINTF_USE_C99_FORMAT_SPECIFIERS=1',
+    '-DNANOPRINTF_USE_LARGE_FORMAT_SPECIFIERS=1',
     '-DNANOPRINTF_USE_WRITEBACK_FORMAT_SPECIFIERS=1',
     '-Wall',
     '-Wextra',

README.md

@@ -2,9 +2,9 @@
 
 [![CircleCI](https://circleci.com/gh/charlesnicholson/nanoprintf.svg?style=shield)](https://circleci.com/gh/charlesnicholson/nanoprintf) [![](https://img.shields.io/badge/license-public_domain-brightgreen.svg)](https://github.com/charlesnicholson/nanoprintf/blob/master/LICENSE)
 
-nanoprintf is an implementation of snprintf and vsnprintf for embedded systems that aims for C11 standard compliance. 
+nanoprintf is an implementation of snprintf and vsnprintf for embedded systems that aims for C11 standard compliance.
 
-nanoprintf makes no memory allocations, uses less than 100 bytes of stack, and is around 5KB of ARM Cortex-M object code when optimized with all the bells and whistles turned on (slightly larger on x64).
+nanoprintf makes no memory allocations, uses less than 100 bytes of stack, and is smaller than 5KB of ARM Cortex-M object code when optimized with all the bells and whistles turned on (slightly larger on x64).
 
 nanoprintf is a [single header file](https://github.com/charlesnicholson/nanoprintf/blob/readme/nanoprintf.h) in the style of the [stb libraries](https://github.com/nothings/stb). The rest of the repository is tests and scaffolding and not required for use.
 
@@ -92,6 +92,18 @@ Like `printf`, `nanoprintf` expects a conversion specification string of the fol
 	* `%n`: Write the number of bytes written to the pointer vararg
 	* `%f`/`%F`: Floating-point values
 
+## Limitations
+
+No wide-character support exists: the `%lc` and `%ls` fields behave like `%c` and `%s`, respectively.
+
+No decimal exponent notation support exists: `%e` and `%E` behave like `%f` and `%F`, respectively.
+
+No "dynamic" decimal / exponent notation support exists: `%g` and `%G` behave like `%f` and `%F`, respectively.
+
+No hexadecimal exponent notation support exists: `%a` and `%A` behave like `%f` and `%F`, respectively.
+
+Pull requests welcome!
+
 ## Acknowledgments
 
 Float-to-int conversion is done using [Wojciech Muła](mailto:zdjęcia@garnek.pl)'s [algorithm](http://0x80.pl/notesen/2015-12-29-float-to-string.html).