| 29 Oct 2024 |
 Scrumplex | both -devedition and -beta fail to apply /nix/store/9lx1746yy56pn3dpc35l1xr6l3lykbin-env_var_for_system_dir-ff111.patch in the PR above | 20:10:38 |
| Scrumplex | Looking at the trunk (tip?) it just seems like it was moved around in the file. I will try to update the patch for 133 onwards | 20:11:20 |
|  syd installs gentoo (they/them) joined the room. | 20:13:37 |
 hexa | ok | 20:17:36 |
| Scrumplex | Current patch versus new patch. Seems like the function signature has changed slightly:
--- pkgs/applications/networking/browsers/firefox/env_var_for_system_dir-ff111.patch 2024-08-10 21:13:22.142083987 +0200
+++ pkgs/applications/networking/browsers/firefox/env_var_for_system_dir-ff133.patch 2024-10-29 21:25:17.412737463 +0100
@@ -10,13 +10,13 @@
#include "prprf.h"
#include "nsIAppStartup.h"
-@@ -309,7 +310,8 @@ static nsresult GetSystemParentDirectory(nsIFile** aFile) {
+@@ -297,7 +297,8 @@ static nsresult GetSystemParentDirectory(nsIFile** aFile) {
"/usr/lib/mozilla"_ns
# endif
;
-- rv = NS_NewNativeLocalFile(dirname, false, getter_AddRefs(localDir));
+- rv = NS_NewNativeLocalFile(dirname, getter_AddRefs(localDir));
+ const char* pathVar = PR_GetEnv("MOZ_SYSTEM_DIR");
-+ rv = NS_NewNativeLocalFile((pathVar && *pathVar) ? nsDependentCString(pathVar) : reinterpret_cast<const nsCString&>(dirname), false, getter_AddRefs(localDir));
++ rv = NS_NewNativeLocalFile((pathVar && *pathVar) ? nsDependentCString(pathVar) : reinterpret_cast<const nsCString&>(dirname), getter_AddRefs(localDir));
# endif
if (NS_SUCCEEDED(rv)) {
| 20:26:49 |
| Scrumplex | This new patch applied successfully. Let's see if it will build | 20:27:55 |
| Scrumplex | Nope :(
> 1:02.52(B checking for nss >= 3.106... no(B(B
> 1:02.52(B ERROR: Requested 'nss >= 3.106' but version of NSS is 3.105.0(B(B
| 20:29:39 |
| hexa | on staging | 20:29:44 |
| Scrumplex | I see. Sooo rebase the PR on staging? | 20:30:00 |
| hexa | because ... reasons | 20:30:02 |
| hexa | let me check | 20:30:14 |
| Scrumplex | Or should I move the beta ones out of this PR and handle them in a separate staging PR? | 20:30:23 |
| hexa |
- Bug 1925975 - NSS 3.106 should be distributed with NSPR 4.36.
| 20:30:30 |
| Scrumplex | * Or should I move the beta/133+ ones out of this PR and handle them in a separate staging PR? | 20:30:31 |
| hexa | because nspr | 20:30:31 |
| hexa | ideally, yeah | 20:30:40 |
| hexa | and nspr causes a rebuild of nss, which is mass-rebuildy | 20:30:53 |
| hexa | can you pick 795da483b5cf1c00dea6a14e381a12572135a7c3 into your pr? | 20:32:09 |
| hexa | "should" should allow us to to do it anyway | 20:32:25 |
| Scrumplex | Into the one targeting master? | 20:32:43 |
| hexa | yes | 20:34:06 |
| Scrumplex | But I assume you don't want me to push that right? It's just for testing the build right? | 20:35:04 |
| hexa | just push it | 20:35:10 |
| hexa | it will be fine | 20:35:12 |
| Scrumplex | Oh I see that it causes <100 rebuilds. I assumed it would be way more | 20:36:05 |
| Scrumplex | I have another rebase in the oven to make sure all packages build when bisecting (moving nss and
@Scrumplex
buildMozillaMach before all other commits). | 20:37:53 |
| Scrumplex | * I have another rebase in the oven to make sure all packages build when bisecting (moving nss and buildMozillaMach before all other commits). | 20:37:57 |
| hexa | nss_latest is cheap, nss isn't | 20:37:58 |
| Scrumplex | Looks good so far. Both devedition and beta are quite a few minutes into the build. I also was building thunderbird while trying to get the Firefoxes working. Will report back once all builds are done | 20:46:36 |
| Scrumplex | My Thunderbird build got OOM killed during linking 🥲 | 21:37:42 |
