mirror of
https://github.com/mozilla/pdf.js.git
synced 2025-04-19 22:58:07 +02:00
dbb82f05fc
During initial parsing of every PDF document we're currently creating a few `1 kB` strings, in order to find certain commands needed for initialization. This seems inefficient, not to mention completely unnecessary, since we can just as well search through the raw bytes directly instead (similar to other parts of the code-base). One small complication here is the need to support backwards search, which does add some amount of "duplication" to this function. The main benefits here are: - No longer necessary to allocate *temporary* `1 kB` strings during initial parsing, thus saving some memory. - In practice, for well-formed PDF documents, the number of iterations required to find the commands are usually very low. (For the `tracemonkey.pdf` file, there's a *total* of only 30 loop iterations.) |
||
|---|---|---|
| .. | ||
| core | ||
| display | ||
| images | ||
| shared | ||
| doc_helper.js | ||
| interfaces.js | ||
| license_header.js | ||
| license_header_libre.js | ||
| pdf.image_decoders.js | ||
| pdf.js | ||
| pdf.worker.entry.js | ||
| pdf.worker.js | ||
| worker_loader.js | ||