1519231 - Stop building nsssysinit for Firefox

Status: RESOLVED FIXED

(NSS :: Build, enhancement)

Description

[Martin Thomson [:mt:]]

In Bug 818686 we noticed that nsssysinit.so was being built in gyp builds [1]. This it not built when MOZILLA_CLIENT was set in make builds [2].

Given that we don't use the default database location in Firefox AND the code works fine without nsssysinit [3], I think that we can stop building and linking this. It's not big, but every bit counts.

[1] gyp: https://searchfox.org/nss/rev/7fcf00883fb4d25a1bd085c824489c405b98666a/nss.gyp#32
[2] make: https://searchfox.org/nss/rev/7fcf00883fb4d25a1bd085c824489c405b98666a/lib/Makefile#39-43
[3] https://searchfox.org/nss/rev/7fcf00883fb4d25a1bd085c824489c405b98666a/lib/pk11wrap/pk11load.c#467-472

Comment 1

[Martin Thomson [:mt:]]

Ted, you did the original port. Would you agree with this assessment?

Comment 2

[Martin Thomson [:mt:]]

Attachment: Bug 1519231 - Disable nsssysinit for Firefox builds, r?ted

This disables building of nsssysinit in gecko builds.  This code isn't used by
Firefox and we could use less code.

Comment 3

[Martin Thomson [:mt:]]

nss-try looks good: https://treeherder.mozilla.org/#/jobs?repo=nss-try&resultStatus=success%2Ctestfailed%2Cbusted%2Cexception%2Cusercancel%2Crunning%2Cpending%2Crunnable&classifiedState=unclassified&selectedJob=221192508&revision=f1c3ba2f05d785928efc2aa79c9067a778f57d9d

Comment 4

[(not currently active) Ted Mielczarek]

Nice catch! It does indeed sound sensible. The original gyp files were somewhat auto-generated by evaluating the Makefiles. While I did compare results it's likely that some nuance was lost.

Comment 5

[Martin Thomson [:mt:]]

https://hg.mozilla.org/projects/nss/rev/f0f598a244605e3739acf75e956339dff4ce03d0