Comment base64

1 files changed, 7 insertions, 4 deletions

diff --git a/base64.cpp b/base64.cpp
index 7e0524a..a6a85e9 100644
--- a/base64.cpp
+++ b/base64.cpp
@@ -6,7 +6,7 @@ using namespace std;
 namespace Base64{
 
 	char alphabet[65]="ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
-	uint8_t revbet[128]={
+	uint8_t revbet[128]={ //reverse look-up table of `alphabet` (for decoding)
 #define XX (127)
 		XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX, //0-15
 		XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX,XX, //16-31
@@ -19,6 +19,8 @@ namespace Base64{
 #undef XX
 	};
 
+	//This is the standard base64 algorithm; most of the code deals with correct padding, actually.
+	//There are a lot of magic numbers here, but they should be obvious regarding the algorithm.
 	string encode(const string &data){
 		int sz=data.size();
 		if(sz==0)return {};
@@ -50,24 +52,25 @@ namespace Base64{
 		return res;
 	}
 
+	//The inverse of `encode`.
 	string decode(const string &dataS){
 		int szS=dataS.size();
 		if(szS==0)return {};
 		uint8_t data[szS];
 		int sz=0;
-		for(char c : dataS){
+		for(char c : dataS){ //First filter away all non-base64 characters (probably mostly newlines, if any)
 			if(revbet[c&0x7f]!=127)data[sz++]=revbet[c&0x7f];
 		}
 
 		int blocks=sz/4;
 		int endlen;
-		if(sz%4==0){
+		if(sz%4==0){ //Detect padding; tries to be nice and forgive bad padding
 			if(data[sz-1]=='='){
 				blocks--;
 				if(data[sz-2]=='=')endlen=1;
 				else endlen=2;
 			} else endlen=0;
-		} else endlen=sz%4-1;
+		} else endlen=sz%4-1; //padding not present... assume the data is OK sort-of?
 		string res(3*blocks+endlen,'\0');
 		int x;
 		for(int i=0;i<blocks;i++){