#3178 closed task (fixed)
Upgrade to nss-3.12.8
| Reported by: | Lars Bamberger | Owned by: | |
|---|---|---|---|
| Priority: | normal | Milestone: | x-future |
| Component: | BOOK | Version: | SVN |
| Severity: | normal | Keywords: | |
| Cc: |
Description
New version: nsss-3.12.8
Compiles fine with the current instructions. nss-3.12.5-standalone-1.patch also applies to this version.
This is a dependency of ticket:3176
This depends on ticket:3177
Change History (5)
comment:1 by , 14 years ago
comment:2 by , 14 years ago
| Owner: | changed from to |
|---|
Weird, mozilla usually barfs during the build if nss or nspr are too old. This time, it just ignored my too-old version of nss.
comment:3 by , 14 years ago
And now I remember why I *hate* nss - the oldest part of the mozilla codebase, requiring wierd and obscure imprecations to do anything. At least I've managed to run the testsuite, after digging around to understand the DOMSUF explanation.
On openssl-1.0.0a (which works fine in use with firefox and epiphany), 4 tests (all ssl) fail and 3698 pass. The output on stderr or stdout showed
selfserv: PR_Bind returned error -5982: Local Network address is in use
I also note that the first completed run of the test took 41 minutes. After throwing that build away, I managed to (twice) run the tests in less than 12 minutes with the same result. The failures were labelled in the html as 'CRL SSL CLient Tests' with the message "Fatal - selfserv process not detectable". The test numbers were 455, 1549, 2529, 2597 (but see below - the number of any particular test can apparently vary).
Eventually found a system with gcc-4.5.1 and openssl-0.9.8. All 5732 tests (sic) passed, and the test numbers in the html do not relate to those on the other box. This is a slightly faster box, with the cpu always running at full speed, and a faster disk, but the test still took 22 minutes.
I conclude that running these tests is something that should be left to consenting adults.
This ticket obsoletes ticket:3109