Enable PHC on early beta 72 builds
Categories
(Core :: Memory Allocator, enhancement)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox72 | --- | fixed |
People
(Reporter: n.nethercote, Assigned: n.nethercote)
References
Details
Attachments
(1 file)
We want to enable PHC on early beta builds in 72, in early December.
| Assignee | ||
Comment 1•6 years ago
|
||
Some edited comments from Slack:
- we need to figure out how to propagate
EARLY_BETA_OR_EARLIERto the build script - it's an
AC_SUBST, it's available inmoz.build. Not in configure, though, because it's all in autoconf - we can "parse" the
defines.shfile in python configure, and make it available inbuild/moz.configure/memory.configure.
glandium: Any suggestions on breaking this down further, and who should do it?
Updated•6 years ago
|
Comment 2•6 years ago
|
||
Nick, it looks like we're at the point where we should just unconditionally enable and let it ride the train. 2 weeks into the beta we can uplift a backout.
| Assignee | ||
Updated•6 years ago
|
| Assignee | ||
Comment 3•6 years ago
|
||
Updated•6 years ago
|
Updated•6 years ago
|
Comment 5•6 years ago
|
||
| bugherder | ||
| Assignee | ||
Comment 6•6 years ago
|
||
I just discovered that this commit doesn't work. The not default_replace_malloc condition here succeeds, which disables PHC. So PHC hasn't been enabled on Beta 72, which explains why no new PHC crashes have come in.
Mike, what would be the right way to enable this on Beta? Remove the not default_replace_malloc condition? Something else?
Comment 7•6 years ago
|
||
I guess you can remove/comment the entire condition. Another solution would be to restore the original code and put an explicit --enable-phc in the relevant mozconfigs.
Description
•