Fixes #623 #624
Fixes #623 #624
Conversation
| return Arrays.stream(COMMA_SPLITTER.split(unparsedHeaderValue)) | ||
| // this is special case, since it's a single char and not a regex reserved one, | ||
| // it defaults to a simple fast lookup @see java.lang.String#split(java.lang.String) | ||
| return Arrays.stream(unparsedHeaderValue.split(",")) |
There was a problem hiding this comment.
should use instead a precompiled regex, String#split is expensive
There was a problem hiding this comment.
this is a special case, if you look at the code of how String handles this you'll notice that it will not use regular expressions at all.
… to minimize restarts and object allocations
There was a problem hiding this comment.
I wrote a few examples as test before I saw your new version and all pass with the new version except this one:
assertEquals("[\"a\"]", HeaderParser.convertToParsedHeaderValues("\\\"a\\\"", s -> s).toString());
I assume that the header does not make sense, it seems that the old version removed \" from anywhere in the string and the new one does it only for the values after =
If we can ignore this case, everything else is ok from my perspective
|
I was reading the RFC and quoted strings are only on values so I only check after the equals sign. I will read the RFC again to confirm this. |
|
https://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html Quoted strings are only after the equals sign. |
|
ok, then the previous implementation was a bit more lenient, we can probably ignore that |
Fixes #623