Incorrect regex patterns

[kemonats]

Examples that are considered correct

@STRUCT;1 HEADER="STRUCT;1"

@STRUCT=1 FIELD_NAME="@STRUCT" VALUE="1"

INT1===1 FIELD_NAME="INT1==" VALUE="1"

@2STRUCT1 Whether the first character can be a digit?

In addition, the line type check can be simplified. Since the input file has a simple format , you can do parsing without using regex, it will probably be faster.

By the way, for which processors it is provided.

[PNieck]

Most of these example are not consistent with our file syntax

First character can be a digit

[PNieck]

As for the processors in this project we use armv7a9-zynq7000(Xilinx Zynq 7000)

[kemonats]

That is why I showed them that they do not conform to your file syntax and are considered appropiate.

First example: line = @STRUCT;1\n

#define WORD "[A-Za-z0-9_.,-\\+]"
#define HEADER_PATTERN "^[[:space:]]*@(" WORD "+)[[[:space:]]*(#.*)?[[[:space:]]*$"

substituting WORD we have

HEADER_PATTERN "^[[:space:]]*@([A-Za-z0-9_.,-\\+]+)[[:space:]]*(#.*)?{[:space:]]*$"

Short description:

• char @ - header,
• then uppercase letters STRUCT - ok,
• char; (0x3b) is considered correct because it is in range ,-\ (0x2c - 0x5c),
• char 1 - ok
• char \n - ok

That is, the parser will interpret this line as a header with the name STRUCT;1. The header name can contain uppercase and lowercase letters, digits, an underscore character, a period, characters between 0x2c-0x5c, a \ , plus sign.

If this pattern is to follow your file syntax it should look like this:

HEADER_PATTERN "^[[:space:]]*@([A-Za-z0-9_]+)[[:space:]]*(#.*)?$"

i.e.

• at the begining can be [[:space:]]
• char @
• the first group one or more upper and lower letters, digits, char _ ([A-Za-z0-9_]+) or instead of + give {1,16}
• next can be [[:space:]]
• the second group can be comment (#.*)?

[PNieck]

Ok, yes I agree. Now I see where is the problem.