Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 336751

Summary:

HTMLResourceEncodingGuesser reading stream unnecessarily

Product:

[WebTools] WTP Source Editing

Reporter:

Nick Sandonato <nsand.dev>

Component:

wst.html

Assignee:

Nick Sandonato <nsand.dev>

Status:

RESOLVED FIXED

QA Contact:

Nick Sandonato <nsand.dev>

Severity:

normal

Priority:

CC:

thatnitind

Version:

3.2.3

Keywords:

performance

Target Milestone:

3.2.4

Flags:

thatnitind: review+

Hardware:

OS:

Windows XP

Whiteboard:

WI60406

Attachments:

Description	Flags
patch	none

Description Nick Sandonato

2011-02-09 14:54:56 EST

After work was done on Bug 333925, I came to realize another inefficiency. Right now, the HTMLResourceEncodingDetector will try to read in bytes for the EncodingGuesser to work on; however, the EncodingGuesser only ever will work for Japanese encodings.

So it seems like we're doing a lot of extra work.

Comment 1 Nick Sandonato

2011-02-09 14:58:08 EST

Created attachment 188617 [details]
patch

Patch makes it so that only if the EncodingGuesser is capable of making a guess will it bother reading from the stream. In a large enough workspace (~1020 HTML files) I saw IFile#getContentDescription() times go from ~10ms per file down to about .66ms per file.

Comment 2 Nick Sandonato

2011-03-02 10:24:32 EST

Code checked in. Thanks, Nitin.