Closed Bug 66467 Opened 24 years ago Closed 10 years ago

use atoms everywhere (and not strings) for "us-ascii", "UTF-8", etc.

Categories

(MailNews Core :: Internationalization, defect)

x86
All
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED WONTFIX
Future

People

(Reporter: sspitzer, Unassigned)

Details

(Whiteboard: intl)

looking around at the i18n code and in mime, we do a lot of string allocations and strcmps charsets. (grep for "UTF-8" and "us-ascii" for example) we should fix the code to use atoms, instead of strings. it would save us memory and time. That's just a hunch. I haven't quantified to prove this Don't bother on starting on this until we get good quantify data to back it up.
Moving to future for now. One thing I found is converting to charset atom needs to call do_GetService, and that is expensive (more expensive than strcmp). If we change to atom, we need to use atom in everywhere so we don't need to convert back and forth between atom and string.
Status: NEW → ASSIGNED
Target Milestone: --- → Future
updating summary. yes, we need to use them everywhere instead of converting back and forth. I doubt this will be a big performance win, but we should still do it, since it is the right thing.
Summary: use atoms, not strings, for "us-ascii", "UTF-8", etc. → use atoms everywhere (and not strings) for "us-ascii", "UTF-8", etc.
Product: MailNews → Core
Product: Core → MailNews Core
QA Contact: ji → i18n
Assignee: nhottanscp → nobody
Status: ASSIGNED → NEW
Whiteboard: intl
The gain we'd get from atomization is probably not worth the effort required to do so--relatively few checks per individual (and very short) strings.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.