78787 - Instrument key areas of Mail/News code for measuring performance automatically.

Status: RESOLVED INVALID

(SeaMonkey :: MailNews: Message Display, defect)

Description

[[e-mail] (gone - use [e-mail] instead)]

Preface:  The scope of this bug is intended for Seth and I to share our ideas
and implementations of instrumenting and reporting Mail/News performance results.  

Note:  Some of these require significant planning, as the timing I do is "what
the user sees", not limited to the APIs in mail/news code alone (we call layout,
etc.)

Overview: The following things need to be instrumented, in order of priority.

Add timers (or uncomment them, if they exist already)

1.  Folder Loading Speed
2.  Message Display Speed
3.  Batch Delete 5 Messages

Other ideas to implement:

1.  After the timers are written, dump using PR_LOG with an Environment
Variable, and output to a logfile.
2.  We will need to be able to parse this logfile (preferably automagically) and
take the results.
3.  (Slightly complex).  Get this data, and using a Perl script, plot this data
using GNU Plot.

Comment 1

[[e-mail] (gone - use [e-mail] instead)]

Attachment: Allows the entering of a column of data, calculates averages and then displays it.

Comment 2

[[e-mail] (gone - use [e-mail] instead)]

This is something for later on, when we have an HTML page to write 
the values of these variables/arrays to.  This doesn't get us very far, but I 
wrote it anyway, because a) it helps me to learn Javascript b) it eliminated use 
of the calculator c) this can be used as a starting point for someday having an 
HTML page with static data for the following columns:

Target (Outlook Express 5.5)
Netscape 4.76
Netscape 6.01 (and eventually 6.5) RTM

In the future, all that might have to be done weekly is write data to 3 columns 
(or if we really get sophisticated, only the current week's data, if we can 
store the 2 previous weeks' data) inside of the 
http://www.mozilla.org/mailnews/win_perfomance_results.html page. 

Entering data into columns, and shifting whole cells over (cut and pasting) is 
actually what takes up quite a bit of my time, and it very repetitive.)

Comment 3

[lchiang]

stephen - thanks for taking initiative on this, along with Seth!

Comment 4

[kmurray1115]

This is great stuff.

Comment 5

[[e-mail] (gone - use [e-mail] instead)]

Attachment: Comments of what needs to be done to output automatically

Comment 6

[[e-mail] (gone - use [e-mail] instead)]

*** Bug 23053 has been marked as a duplicate of this bug. ***

Comment 7

[[e-mail] (gone - use [e-mail] instead)]

Comment on attachment 33146 [details]
Allows the entering of a column of data, calculates averages and then displays it.

obsolete (using new python scripts)

Comment 8

[[e-mail] (gone - use [e-mail] instead)]

Comment on attachment 33568 [details]
Comments of what needs to be done to output automatically

also obsolete

Comment 9

[[e-mail] (gone - use [e-mail] instead)]

Cavin, can you hook up logging (using NSPR_LOG_MODULES) for Folder Loading
Speed, similar to what Jean-Francois did for the msgcompose:5 metric?  Folder
Loading Speed has gotten so fast lately that even on a slow machine, it's
becoming hard to get precise tests when performance enhancements land for that
area.  Shirley also has mentioned this to me in her internationalization
performance testing (which follows my page's methodology).  Stopwatch just isn't
accurate enough for something as important as this.

Comment 10

[ji]

Yes, on a 266 MHz machine, loading of a 100M folder that contains about 1,010
msgs is too fast to measure accurately using a stopwatch.

Comment 11

[Arun Prasannan (polar)]

This bug is being marked EXPIRED as it has seen no activity in a very long time.

If you think that the issue reported might still be relevant, please test with a recent release of SeaMonkey and if the problem persists feel free to re-open the report. Thank you.

http://www.seamonkey-project.org/

Comment 12

[Mark Banner (:standard8)]

Bulk reopening incorrectly expired bugs - no activity does not constitute no bug - these need proper checking.

Comment 13

[Wayne Mery (:wsmwk)]

I'll step out on a limb and say this won't happen, given that key areas are (moved to) mailnews core. And mailnews core continues to improve its test