ColorCode-Universal icon indicating copy to clipboard operation
ColorCode-Universal copied to clipboard

JSON Parsing - deadlock/stuck on parsing Json to HTML

Open JochnGst opened this issue 2 years ago • 5 comments
trafficstars

Im trying to parse this JSON sipped to my Blazor page. But because of some weird RegEx parsing issue the process get stuck without any Exception. Can somebody tell me where there could be a Problem? here you can find my test project: ColorCodeTest

This is my Test Code

string _jsonString3 = "{\r\n \"raw_causes\": [\r\n      \"Winterglatter Fahrbahn\",\r\n      \"Nicht angepasste Geschwindigkeit\",\r\n      \"test3\"\r\n    ]\r\n      }";
try
{
    var _formatter = new HtmlFormatter();
    var language = ColorCode.Languages.FindById("json");
    _jsonHtml = _formatter.GetHtmlString(_jsonString3, language);

}
catch (Exception)
{

    throw;
}

It stuck after the array Element \"Winterglatter Fahrbahn\", when it call regexMatch = regexMatch.NextMatch(); and I have no idear why this happends

        private void Parse(string sourceCode,
                           CompiledLanguage compiledLanguage,
                           Action<string, IList<Scope>> parseHandler)
        {
            Match regexMatch = compiledLanguage.Regex.Match(sourceCode);

            if (!regexMatch.Success)
                parseHandler(sourceCode, new List<Scope>());
            else
            {
                int currentIndex = 0;

                try
                {
                    while (regexMatch.Success)
                    {
                        string sourceCodeBeforeMatch = sourceCode.Substring(currentIndex, regexMatch.Index - currentIndex);
                        if (!string.IsNullOrEmpty(sourceCodeBeforeMatch))
                            parseHandler(sourceCodeBeforeMatch, new List<Scope>());

                        string matchedSourceCode = sourceCode.Substring(regexMatch.Index, regexMatch.Length);
                        if (!string.IsNullOrEmpty(matchedSourceCode))
                        {
                            List<Scope> capturedStylesForMatchedFragment = GetCapturedStyles(regexMatch, regexMatch.Index, compiledLanguage);
                            List<Scope> capturedStyleTree = CreateCapturedStyleTree(capturedStylesForMatchedFragment);
                            parseHandler(matchedSourceCode, capturedStyleTree);
                        }

                        currentIndex = regexMatch.Index + regexMatch.Length;
                        regexMatch = regexMatch.NextMatch();
                    }
                }
                catch (Exception ex)
                {

                    throw;
                }

                string sourceCodeAfterAllMatches = sourceCode.Substring(currentIndex);
                if (!string.IsNullOrEmpty(sourceCodeAfterAllMatches))
                    parseHandler(sourceCodeAfterAllMatches, new List<Scope>());
            }
        }

JochnGst avatar Aug 04 '23 07:08 JochnGst

I found out that there is a conflict with the Key-LanguageRule

new LanguageRule(
     $@"[,\{{]\s*({Regex_String})\s*:",
     new Dictionary<int, string>
         {
             {1, ScopeName.JsonKey}
         }),

for my case it works when I use this RegEx $@"[,\{{]\s*(""\w*"")\s*:" But I know that this will not catch all edge cases for a JSON-Key

JochnGst avatar Aug 04 '23 14:08 JochnGst

I found out that there is a conflict with the Key-LanguageRule

new LanguageRule(
     $@"[,\{{]\s*({Regex_String})\s*:",
     new Dictionary<int, string>
         {
             {1, ScopeName.JsonKey}
         }),

for my case it works when I use this RegEx $@"[,\{{]\s*(""\w*"")\s*:" But I know that this will not catch all edge cases for a JSON-Key

I made this to clean up the RegEx pattern:

       public static List<string> ExtractKeys(string jsonString)
       {
           var keys = new List<string>();
           var matches = Regex.Matches(jsonString, "[,\\{]\"(.*?)\"\\s*:");
           foreach (Match match in matches) { keys.Add(match.Groups[1].Value); }
           return keys;
       }

GuildOfCalamity avatar Oct 06 '23 19:10 GuildOfCalamity

@GuildOfCalamity @JochnGst Encountering the same problem, what is the reasonable solution?

niltor avatar Mar 25 '24 09:03 niltor

This has become a serious issue for us; showing the colored JSON of certain Intune policies locks up our whole Blazor app making it unusable and taking down the environment with it.

image

Yomodo avatar Jul 26 '24 08:07 Yomodo

I believe the issue relates to excessive regex backtracking when parsing json keys.

Atomic groups can be used by tweaking the original LanguageRule from

new LanguageRule(
     $@"[,\{{]\s*({Regex_String})\s*:",
     new Dictionary<int, string>
         {
             {1, ScopeName.JsonKey}
         })

to

new LanguageRule(
     $@"[,\{{]\s*(?>{Regex_String})\s*:",
     new Dictionary<int, string>
         {
             {1, ScopeName.JsonKey}
         })

Kompiler avatar Sep 09 '24 01:09 Kompiler

I am seeing this issue too. The rendering hangs when processing JSON of any size and type.

justinasmussen avatar Nov 10 '25 21:11 justinasmussen