| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
See Bug #5306.
(cherry picked from commit 117776e0c01e68a8b65584d86d7b8b56fe75c9d0)
|
|
|
|
|
|
|
|
|
| |
Make status column sort order work correctly with the icons.
Also refactor get_service_status_icon() output string construction to be cleaner and clearer.
(2.4, 2.3)
(cherry picked from commit 446505a9f9be7f43e4515658f1a5444bc3732a3f)
|
|
|
|
|
|
| |
CSR request
(cherry picked from commit 7fd7fbcff3304285f4407bec2ae62bab7195bcc4)
|
|
|
|
| |
(cherry picked from commit 9c91c7bd747074b8cdaa90e8810f0c2df081f72d)
|
| |
|
|
|
|
| |
and for all IPv6 items
|
|
|
|
|
|
|
|
|
| |
As far as I can see, filter_generate_user_rule() is always supposed to be called with 'ipprotocol' set to 'inet' or 'inet6'. The cases of rules for both ('inet46') are handled by calling filter_generate_user_rule() twice, passing 'inet' then 'inet6'.
So at this point, if 'ipprotocol' is blank, then it is from an old rule, and it [can|should|must] default to 'inet'.
This would provide a generic fix for old rules that do not have 'ipprotocol' specified.
The other thing that could be done is make some upgrade code that fills in 'ipprotocol' on old rules at upgrade.
|
|
|
|
| |
(cherry picked from commit 3d8b01e8c6392b4177572d540c8160c7e6e071ca)
|
|
|
|
| |
(cherry picked from commit 4fb2b17772928f39add5fc0529e94ed07a09de31)
|
| |
|
|
|
|
|
|
| |
Fix #7051
This reverts commit 04665e78537906f7375668ca665cba17f95a4864.
|
|
|
|
| |
This reverts commit 855826896509a1a0bec77a51535a8f004b4ca570.
|
|
|
|
| |
(cherry picked from commit 7abc3f992e5dd5bff53495844ce944163d6d1d9b)
|
|
|
|
|
|
|
|
|
|
|
| |
In some places ldap_get_groups has:
```
return memberof;
```
It should have the "$" in front, so it will return the $memberof array (that is empty when this happens).
This causes issues for callers that expect to have a return value that is either false, an empty array, or an array of the groups.
(cherry picked from commit 0241b34f1a33c3ae83fdf817c8c374b10775335a)
|
|
|
|
|
|
|
|
|
|
|
|
| |
When there is an upgrade, the echo here was outputting a stale value of the version. For example, on first upgrade from 2.3.3-DEVELOPMENT to 2.3.3-RC the console had:
pfSense (pfSense) 2.3.3-DEVELOPMENT amd6 Sat Feb 11 14:24:27 CST 2017
Bootup complete
FreeBSD/amd64 (myhost.localdomain) (ttyv0)
*** Welcome to pfSense 2.3.3-RC (amd64 full-install) on myhost ***
That is a bit confusing for users to be sure which version it is at this point.
|
| |
|
|
|
|
|
|
| |
- change $do_ping default value to 'true' (which emulates the previous default behavior) to avoid any unexpected results
(cherry picked from commit 20cf8d8e20fa28c16e86ce0d91e57e4d78427d26)
|
|
|
|
|
|
| |
- Reworked based on recent comments from @rbgarga
(cherry picked from commit c516cb287a78f7b05459e7fcba410f443d8eb8af)
|
|
|
|
|
|
| |
- incorporate suggestions from @rbgarga with slight modification
(cherry picked from commit 6c2f093000b05285546e81dd1a578fc9b573b72b)
|
|
|
|
|
|
|
|
|
|
|
| |
/etc/inc/util.inc:
- arp_get_mac_by_ip() updated to support IPv6
- attempt at code streamline
/usr/local/www/services_dhcp_edit.php:
- streamline code, now just a simple call to arp_get_mac_by_ip()
(cherry picked from commit dd83f869b79a858bd74c7a8bb4adcd49217445b0)
|
| |
|
|
|
|
| |
status, for use by a shell script and status.php. Ticket #7046
|
| |
|
|
|
|
| |
(cherry picked from commit 40dcb4b61a2c1213a0b3e213c78fddac845a0117)
|
|
|
|
| |
(cherry picked from commit 5af0922d75724e1eac89017173457f57842387f8)
|
|
|
|
| |
(cherry picked from commit ac5ee07ee1daef2f43e728895290ca6d11efe0f3)
|
|
|
|
| |
(cherry picked from commit cb5961d1fa64a45cbec5ef5d677b57f8d62f50b5)
|
| |
|
|
|
|
| |
(cherry picked from commit 821110e8ff76564c23783c554fc89cd9458683ac)
|
|
|
|
| |
(cherry picked from commit 6924a2bf34a70cd33284a28ca3575f33f9834375)
|
|
|
|
|
|
|
|
|
| |
If "description_state" is requested here, there are too many "%s" substitutions in the string for sprintf().
Also, to help translators, number all the "%s" substitutions. Then translators can rearrange the text and variable order if they need to.
I removed the commented-out get_service_control_GET_links() while I am here, because it also contained sprintf() strings that had multiple "%s" substitutions. So I had found it when searching for lines containing multiple "%s". Might as well remove it "in the next version" as the comment says.
(cherry picked from commit b3119a61c8f5841f92e2c1fe945f90d9e98a843c)
|
|
|
|
|
| |
Not sure what was the idea here, but these are not used anywhere, do nothing as they immediately call ```return false;``` plus the second one is also misnamed.
(cherry picked from commit edba33b5a567ab8c9d4827fa26a25bd9649e3fac)
|
|
|
|
|
|
|
|
| |
* rename function args to be clearer what they do ($local_only was quite ambiguous, at first sight it could mean any of: "don't update local catalog copy", "only check local catalog copy", or "only report local installed pkgs")
* merge nested if () conditions
* refactor minor code stuff
* rewrite explanatory comment
(cherry picked from commit 04daf8b1f016e17bede4ad00db46f2855d3e7c1f)
|
|
|
|
| |
(cherry picked from commit 2f633b526075b2ed5e0e160ef6f0d025b509bd70)
|
|
|
|
| |
(cherry picked from commit e5f96a2cb3c0cad0c828148bd7b8d45c130a9b17)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(resubmit)
Resubmit of PR #3157 with fix.
The issue in #3157 was that `pkg info` and `pkg search`, undocumented in man pages, seem to handle things differently if no packages match the pattern string. `pkg search` gives an error "No packages match [$pkgs]", whereas `pkg search` doesn't give an error.
I didn't come across this because it's not stated in the man pages for pkg info, and because I tested it with a bare install but at least one pkg, while @NYOB used it on a bare install.
The fix which I've tested and seems to work nicely, is to test using `pkg info -e` whether the pkg info call will hit a "no packages exist" issue or not. If `pkg info -e` returns 0 then it's safe as there's at least one matching pkg (any other errors are allowed to occur and be detected as usual). If it returns anything else, then there's no matching pkgs, so the main `pkg info` is skipped leaving $out and other variables at the correct values they need anyway, showing a null match/none found.
For details of the underlying issue being fixed, and the issue that this resubmit fixes, see the original PR.
(cherry picked from commit 587988f6df7e1d3b1a369b93a5634620129e3d9d)
|
| |
|
|
|
|
| |
(cherry picked from commit f0b1358dfe520ad3b771127127daed970ba2c0a0)
|
|
|
|
| |
(cherry picked from commit cde28bfa0e11f268485ec1f6ccb73a3a2f66448f)
|
|
|
|
| |
(cherry picked from commit 587995fb57f91894d1f8eb6b296a9fe2fa111fac)
|
| |
|
|
|
|
|
|
| |
so users can grant access to Services->UPNP
Note: Status->UPNP already has a priv and it works.
(cherry picked from commit a5a899e4388f2737a6d1cdc82c7325c20fb72ee4)
|
|
|
|
|
| |
from a user that does not have admin or all page privilege.
(cherry picked from commit 166540830275318c8dec9199d8a9ee0e605f606a)
|
|
|
|
| |
(cherry picked from commit ddf99718d5f1f4545483c39d3759fdfbb788b0fb)
|
| |
|
| |
|
|
|
| |
When 'any' is selected as the ICMP type, do not write 'icmp-type any' in the rule, just leave it out.
|
|
|
|
| |
only when ipprotocol is inet6
|
| |
|
|
|
|
| |
(cherry picked from commit a2d23e88596deab6bbed2818385a0b72c913843a)
|