jar protocol seems to read in the entire file and then copy parts of it into necko's buffer. Rather it would be better if it uncompresses into necko's buffer directly. This would save - the extra copy - the malloc/free cost - the fragmentation associated with allocations
Per dveditz: "care should be taken so we dont dont seek too much for every time we fill necko's buffer"
Keywords: footprint, nsbeta1, perf
Priority: -- → P2
Target Milestone: --- → mozilla1.0
This is a perf win. Not footprint. Will prevent 200+ allocation of big buffers and will save on extra memcpy.
Keywords: footprint, nsbeta1 → nsbeta1+
This is fixed in bug 189528, so can be closed now. *** This bug has been marked as a duplicate of 189528 ***
Status: ASSIGNED → RESOLVED
Last Resolved: 15 years ago
Resolution: --- → DUPLICATE
Status: RESOLVED → VERIFIED
Component: General → Networking: JAR
QA Contact: general → networking.jar
You need to log in before you can comment on or make changes to this bug.