switch clog to cat in status_tinc.php #691
Conversation
"tail -n 50" I think there is no reason to "sed" all log file
there is can be too many connections in logs
|
|
||
| exec("{$clog_path} /var/log/tinc.log | /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); | ||
| if ($firmware['config_version'] >= 19.7) { | ||
| exec(system_log_get_cat() . ' ' . $logfile . "| /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); |
There was a problem hiding this comment.
Rather than $logfile here, it should use sort_related_log_files() like the other example.
| if ($firmware['config_version'] >= 19.7) { | ||
| exec(system_log_get_cat() . ' ' . $logfile . "| /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); | ||
| } else { | ||
| exec("{$clog_path} /var/log/tinc.log | /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); |
There was a problem hiding this comment.
Why use the full hardcoded path to the log here, rather than {$logfile} which is populated above?
|
|
||
| exec("{$clog_path} /var/log/tinc.log | /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); | ||
| if ($firmware['config_version'] >= 19.7) { |
There was a problem hiding this comment.
This isn't a proper version test. The config revision is not a reliable indicator of the pfSense version. It should be based on $g['product_version'] or similar to test the actual version directly.
There was a problem hiding this comment.
in case of $g['product_version'] we need to enumerate all supported versions
better to use kernel version number
There was a problem hiding this comment.
No, it is not. That does not necessarily determine the pfSense version either. It must use product_version. If you need to compare, there are convenience functions like pfs_version_compare() that facilitate a proper comparison.
| if ($firmware['config_version'] >= 19.7) { | ||
| exec(system_log_get_cat() . ' ' . $logfile . "| /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); | ||
| } else { | ||
| exec("{$clog_path} /var/log/tinc.log | /usr/bin/sed -e 's/.*tinc\[.*\]: //'", $result); | ||
| } | ||
|
|
There was a problem hiding this comment.
See comments above which also apply to this similar block.
(And since they are practically identical, I have to wonder if this could be simplified significantly, though that's likely out of scope for this particular PR)
| if (preg_match("/Connections:/", $line)) { | ||
| $begin = $i; | ||
| if ($usr == 'USR1') { | ||
| foreach ($result as $line) { |
There was a problem hiding this comment.
The USR1 and else block are still pretty similar. Looks like you could use the if ($usr == 'USR1') ... else test to set the beginning and ending match strings only, and then just have one foreach.
There was a problem hiding this comment.
fixed pfs_version_compare() + foreach
Changelog: * Fixed: Version `Qt_5.15' not found (required by /usr/bin/ksnip). (#712) * Fixed: CI packages show continuous suffix for tagged build. (#710) * Fixed: kImageAnnotator not translated with deb package. (#359) * Fixed: Windows packages increased in size. (#713) * Fixed: The string 'Actions' is not available for translation. (#729) * Fixed: HiDPI issue with multiple screen on Windows. (#668) * Fixed: Snipping Area not closing when pressing ESC. (#735) * Fixed: Sometimes "Snipping Area Rulers" not shown after starting rectangular selection. (#684) * Fixed: Cursor not positioned correctly when snipping area opens. (#736) * Fixed: Mouse cursor not captured when triggered via global shortcut. (#737) * Fixed: Dual 4K screens get scrambled on X11. (#734) * Fixed: VCRUNTIME140_1.dll was not found. (#743) * Fixed: Screenshot area issue when monitor count changes on Windows. (#722) * Fixed: Wayland does not support QWindow::requestActivate(). (#656) * Fixed: Wrong area is captured on a Wayland screen scaling. (#691) * Changed: Enforce xdg-desktop-portal screenshots for Gnome >= 41. (#727) * Fixed kImageAnnotator: Crash while typing text on wayland. (#256) * Changed kImageAnnotator: Show scrollbar when not all tools visible. (#258)
Changes since 1.4.0: v1.5.1 - fix: Properly close created files by @beatbrot in #702 - fix: Skip auth type prompt if already set by @ankitpokhrel in #701 - @beatbrot made their first contribution in #702 Full Changelog: ankitpokhrel/jira-cli@v1.5.0...v1.5.1 v1.5.0 This release brings the support for mTLS authentication along with some other features like setting affects version, updating the estimate, etc. - feat: Enable issue edit to read body from stdin by @erpel in #619 - feat: Affects version by @damianoneill in #642 - feat: Add mtls authentication for client certificate auth by @markhatc- in #615 - feat: Add support for updating the estimate by @chapmanc in #669 - fix: Issue with no-input on create by @ankitpokhrel in #655 - fix: Jira init broken due to authtype value by @ankitpokhrel in #694 - fix: Bring bearer back by @ankitpokhrel in #696 - fix: Respect jira timezone by @ankitpokhrel in #697 - chore: Bump go & alpine versions by @pbnj in #691 - ci: Upgrade workflow + linter by @ankitpokhrel in #695 - dep: Upgrade all by @ankitpokhrel in #643 - @erpel made their first contribution in #619 - @damianoneill made their first contribution in #642 - @pbnj made their first contribution in #691 - @chapmanc made their first contribution in #669 Full Changelog: ankitpokhrel/jira-cli@v1.4.0...v1.5.0
Redmine Issue: https://redmine.pfsense.org/issues/9740
Ready for review
There is no /usr/local/sbin/clog in pfSense 2.5
using "cat" instead