Debian Patches
Status for sogo/5.12.10-1
| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| disable_is_localhost_test.patch | =================================================================== | no | ||||
| libxml2.patch | =================================================================== | no | ||||
| 0002-Change-sogo-backup-location-and-update-cronjob.patch | Change sogo-backup location and update cronjob Change sogo-backup location to /var/backups/sogo and update sogo-backup.sh location in cronjob to /usr/sbin/sogo-backup. |
Jeroen Dekkers <jeroen@dekkers.ch> | no | 2013-04-13 | ||
| 0003-Fix-may-be-used-uninitialized-in-this-function-warni.patch | Fix "may be used uninitialized in this function" warnings. =================================================================== |
Jeroen Dekkers <jeroen@dekkers.ch> | yes | 2014-04-09 | ||
| 0005-Remove-build-date.patch | Remove build date | Jeroen Dekkers <jeroen@dekkers.ch> | no | 2014-10-05 | ||
| fix_test_rendering_named_value_order.patch | Make test_rendering's named-value ordering check order-independent CardVersitRenderer renders a CardElement's named values by iterating NSDictionary's -allKeys, whose enumeration order is not guaranteed by the Foundation/GNUstep contract. test_rendering's case 5 hard-coded only two of the six possible orderings of its three named values (NAMED1/NAMED2/NAMED3), which happened to be the only two produced on little endian architectures. On big endian architectures a third, equally valid, ordering is produced instead, causing the test to fail even though the rendered vCard is functionally correct. . Rewrite the assertion to accept any of the six possible orderings of the three named fragments, which makes the test pass on every architecture and removes the need to disable it on big endian. =================================================================== |
Jordi Mallach <jordi@debian.org> | not-needed | 2026-08-18 | ||
| 0007-Do-not-use-OpenSSL-when-we-are-configured-to-use-Gnu.patch | Do not use OpenSSL when we are configured to use GnuTLS | Jeroen Dekkers <jeroen@dekkers.ch> | no | 2019-01-13 | ||
| 0008-Unset-MAKEFLAGS-and-MFLAGS-in-configure.patch | Unset MAKEFLAGS and MFLAGS in configure This fixes a build failure when building the package with parallel make. |
Jeroen Dekkers <jeroen@dekkers.ch> | no | 2019-01-13 | ||
| 0009-Omit-signedViewer-altogether-when-not-using-openssl.patch | Omit signedViewer altogether when not using openssl | Hanno Stock <hanno.stock@indurad.com> | no | 2019-07-30 | ||
| python3.patch | =================================================================== | no | ||||
| disable_isIpv4_test.patch | Disable `isIPv4` test This test fails on still unknown scenarios (machines with default IPv6 routes?) and generates lots of unwanted test failures in Debian. =================================================================== |
Jordi Mallach <jordi@debian.org> | not-needed | |||
| cross.patch | Fix FTCBFS. | Yavor Doganov <yavor@gnu.org> | no | 2025-01-26 | ||
| fix_mime_address_generation_test.patch | Fix brittle MIME address generation test The NGMimeAddressHeaderFieldGenerator test fails in isolated build environments (like Debian sbuild) because the SOGo framework resources are not fully assembled/installed. This causes NGMime to fall back to a default encoding that results in "doubled" address strings for simple inputs. This patch allows the test to accept both the standard and the fallback formats. |
Jordi Mallach <jordi@debian.org> | not-needed | 2026-05-12 | ||
| fix_html_injection_import_masking.patch | Fix broken masking of CSS @import in stringWithoutHTMLInjection The @import stripping branch of -[NSString(SOGoURLExtension) stringWithoutHTMLInjection:stripAngular:] used the wrong replacement template for its regex substitution: withTemplate:@"onrep***=", copy-pasted from the unrelated onrepeat= sanitizer defined just above it. Instead of masking "@import" occurrences inside <style> blocks as "@im****" (as the verification loop right after it expects), it replaced them with the literal string "onrep***=", corrupting the sanitized output and leaving the CSS injection payload effectively unmasked. . This caused the TestNSString+Utilities test_stringWithoutHTMLInjection unit test to fail. =================================================================== |
Jordi Mallach <jordi@debian.org> | no | 2026-08-17 |
All known versions for source package 'sogo'
- 5.12.10-1 (sid)
- 5.12.1-3+deb13u2 (trixie, trixie-security)
- 5.8.0-2+deb12u3 (bookworm, bookworm-security)
