buffer optimization: long comments do not trigger errorBufferTooSmall

[dzzie]

this should do the trick give it a try.

[dzzie]

actually one note on this, the prototype should be moved to the private: section of the .h its not for public consumption as it expects that a read of len has already occured

[stevemarple]

It fails the regression test. Values which are expected to be read from test.ini are not found. Below is the diff that make regressiontest outputs:

Using file test.ini
   File open? true
     Looking for key "mac"
-      Value of mac is "01:23:45:67:89:AB"
+      Error: key not found (6)
     Looking for key "mac" in section "network"
-      Value of mac is "01:23:45:67:89:AB"
+      Error: section not found (5)
     Looking for key "mac" in section "network2"
-      Value of mac is "ee:ee:ee:ee:ee:ee"
+      Error: section not found (5)
     Looking for key "mac" in section "fake"
       Error: section not found (5)
     Looking for key "ip"
-      Value of ip is "192.168.1.2"
+      Error: key not found (6)
     Looking for key "gateway"
       Value of gateway is "192.168.1.1"
     Looking for key "hosts allow" in section "network"
-      Value of hosts allow is "example.com"
+      Error: section not found (5)
     Looking for key "hosts allow" in section "network"
-      Value of hosts allow is "example.com"
+      Error: section not found (5)
     Looking for key "hosts allow" in section "network2"
-      Value of hosts allow is "sloppy.example.com"
+      Error: section not found (5)
     Looking for key "hosts allow" in section "network2"
-      Value of hosts allow is "sloppy.example.com"
+      Error: section not found (5)
     Looking for key "string" in section "misc"
-      Value of string is "123456789012345678901234567890123456789001234567890"
+      Error: section not found (5)
     Looking for key "string2" in section "misc"
-      Value of string2 is "a string with spaces in it"
+      Error: section not found (5)
     Looking for key "pi" in section "misc"
-      Value of pi is "3.141592653589793"
-    Pi: 3.14159
+      Error: section not found (5)
+    Could not read "pi" from section "misc" in test.ini
+      Error: section not found (5)
 ----
 Done
make: *** [regressiontest] Error 1

(Lines beginning with "-" are the reference output, lines which begin with "+" are from the current output.)

[dzzie]

sorry I just realized i had a bug in this on the way home from the grocery store, one sec

[dzzie]

ok I amended the commit, I realized I needed to scan the existing buffer for \n before scanning ahead, I was testing on a truncated ini, this time I ran it through the full set sorry

[stevemarple]

That looks good, it passes the regression tests now.

I noticed that I haven't made the library available for automatic installation via the IDE. I plan to update my master copy and submit for inclusion on the lists of third-party libraries. Then I'll merge your changes and push an updated version. That way users can revert if necessary. Thanks for your contribution, Steve.

[dzzie]

one other thought I just had...I wonder if I need to pass the i counter into the readUntilNewLine function and start from there in the for loop instead of i=0,

since the buffer has already been partially processed and pos++ that many times already.

nothing showed up in the tests, but I will see if I can construct one. Simple as the logic seems Its kind of tricky to nail down just right :-\

edit: I cant seem to construct a scenario that bugs it out, the change didnt hurt anything either. The data i thought that might trigger it would have been a comment line with space preceeding the comment char, followed immediately by setting.