Add semicolon and hex notation support to the MOS VDU command#118
Add semicolon and hex notation support to the MOS VDU command#118HeathenUK wants to merge 2 commits intobreakintoprogram:mainfrom
Conversation
Conscious decision to not support comma (rather than whitespace) delimited values, as this would increase code complexity and depth exponentially given the potential for whitespace, comma and semicolon to all act as a delimiter.
src/mos.c
Outdated
| value_str += 2; | ||
| len -= 2; |
There was a problem hiding this comment.
I think these 2 lines should be removed and let strtol() call below to deal with this prefix. Otherwise, the parsing code will allow input like 0x0xFE to be parsed as a valid number.
Also, the independent checks for prefix/suffix should probably be converted to sequence of if / elseif. The idea is to accept as valid only numbers with at most one prefix or suffix, not any combination of all possible prefixes/suffixes.
There was a problem hiding this comment.
Agreed on those two lines, since strtol can handle 0x prefixes there's no need.
We discussed the second point and it costs us nothing to deal with situations where a user may double up on notation.
| // Check for '0x' or '0X' prefix | ||
| if (len > 2 && (value_str[0] == '0' && tolower(value_str[1] == 'x'))) { | ||
| base = 16; | ||
| } | ||
|
|
||
| // Check for '&' prefix | ||
| if (value_str[0] == '&') { | ||
| base = 16; | ||
| value_str++; | ||
| len--; | ||
| } | ||
| // Check for 'h' suffix | ||
| if (len > 0 && tolower(value_str[len - 1]) == 'h') { | ||
| value_str[len - 1] = '\0'; | ||
| base = 16; | ||
| } |
There was a problem hiding this comment.
this is cool - belt & braces style
I had a wild thought on hex handling... it might be nice for hex values to automatically be variable/arbitrary length. so 0x1 or 0x12 would send a single byte, 0x123 or 0x1234 two bytes, 0x12345 or 0x123456 three bytes, and so on
you could just interpret the string one byte at a time. the only tricky part really is dealing with odd length strings, where the first character is a stand-alone nibble rather than a full byte character pair.
3 participants
Conscious decision to not support comma (rather than whitespace) delimited values, as this would increase code complexity and depth exponentially given the potential for whitespace, comma and semicolon to all act as a delimiter.