Returning codelet results that are over a thousand lines long is somewhat senseless, because, if a user requires all relevant matches for a symbol, they'll probably ultimately resort to downloading the file from its source and operating on it manually (and more efficiently). Capping the codelets the client receives at a length of 1,000 lines would also decrease load/syntax-highlighting times.
The excerpt of the codelet that we return needs to be intelligently selected -- optimally, so that the number of matches it contains is maximized, but we can settle for a simpler algorithm because it doesn't much of a difference.
Returning codelet results that are over a thousand lines long is somewhat senseless, because, if a user requires all relevant matches for a symbol, they'll probably ultimately resort to downloading the file from its source and operating on it manually (and more efficiently). Capping the codelets the client receives at a length of 1,000 lines would also decrease load/syntax-highlighting times.
The excerpt of the codelet that we return needs to be intelligently selected -- optimally, so that the number of matches it contains is maximized, but we can settle for a simpler algorithm because it doesn't much of a difference.