{"id":319228,"date":"2022-02-03T00:02:49","date_gmt":"2022-02-03T08:02:49","guid":{"rendered":"https:\/\/dreness.com\/blog\/?p=319228"},"modified":"2022-02-03T00:02:51","modified_gmt":"2022-02-03T08:02:51","slug":"splashguard-of-departure","status":"publish","type":"post","link":"https:\/\/dreness.com\/blog\/archives\/319228","title":{"rendered":"Splashguard of Departure"},"content":{"rendered":"\n

Trivia question: what kind of protection should you wear when calling diff<\/code> from a Shortcuts.app “Run Shell Script” action?<\/p>\n\n\n\n

You should affix the Splashguard of Departure, ideally with some +charisma to offset the posix misalignment. This is because “Run Shell Script” actions are executed with an implicit set -e<\/code>. I’ll elaborate.<\/p>\n\n\n\n

set -e<\/code> means the script (and by extension, the whole shortcut) is halted whenever any of the commands executed in the script exit non-zero. It is customary for tools of the posix persuasion to set their exit status to zero upon success and non-zero upon failure. The trick with common but not mandatory behaviors is that the exceptions tend to sting. diff<\/code> and a few of its siblings are exceptional in this regard, although you wouldn’t learn about that directly from the diff<\/code> man page. Instead, you have to accept the challenge found at the bottom of the diff<\/code> man page:<\/p>\n\n\n\n

The full documentation for diff is maintained as a Texinfo manual.  If the info and diff programs are properly installed at your site, the command\n              info diff\nshould give you access to the complete manual.<\/pre>\n\n\n\n
Respect to elders notwithstanding, Texinfo is a total train wreck every time I try to use it, so please accept this shortcut:<\/p>\n\n\n\n
andre@boom ~ % info diff 2> \/dev\/null | grep -A 1 'exit status of'\n   An exit status of 0 means no differences were found, 1 means some\ndifferences were found, and 2 means trouble.\n--\n   An exit status of 0 means no differences were found, 1 means some\ndifferences were found, and 2 means trouble.\n--\n   An exit status of 0 means `diff3' was successful, 1 means some\nconflicts were found, and 2 means trouble.\n--\n   An exit status of 0 means no differences were found, 1 means some\ndifferences were found, and 2 means trouble.<\/pre>\n\n\n\n
For example, in this shortcut<\/a>, the first action completes because diff<\/code> exits zero, but the second action causes the shortcut to fail with Exit code: 1<\/code>. If it’s a big script, it might even take you a while to pinpoint the problem.<\/p>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
Normally I would run newproc.d<\/code> to monitor command executions in real time, but that requires disabling SIP, which I usually have enabled<\/em> these days due to this unfortunate constraint<\/a>. I don’t know if newproc.d<\/code> would have been the last place I had to look, but I suspect it would have.<\/p>\n\n\n\n
Instead I turned to supraudit<\/a><\/code> which does show all the commands executed from the script – EXCEPT the one that fails the whole shortcut, which is probably a clue if you’re looking closely enough. Console logs are mostly useless in this case without ‘private data<\/a>‘, which is inconvenient to get – for your protection, of course. I eventually found the culprit by testing progressively smaller scripts until diff<\/code> alone was causing failures.<\/p>\n\n\n\n
Right then, the remedy: Splashguard of Departure is easily crafted from two ingots and one eternal truth. Its magical effects include the formation of a specialized tribunal that serves to contain the representations of diff<\/code>\u2019s quality of departure that are not tolerated in this locale. In the following example<\/a>, eternal truth (i.e. \/bin\/true<\/code>) is replaced with a small placard which works just as well.<\/p>\n\n\n\n
\"\"<\/a><\/figure>\n","protected":false},"excerpt":{"rendered":"
Trivia question: what kind of protection should you wear when calling diff from a Shortcuts.app “Run Shell Script” action? You should affix the Splashguard of Departure, ideally with some +charisma to offset the posix misalignment. This is because “Run Shell … Continue reading →<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5,4,14,6],"tags":[],"class_list":["post-319228","post","type-post","status-publish","format-standard","hentry","category-development","category-os-x","category-pro-tip","category-scripts"],"_links":{"self":[{"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/posts\/319228","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/comments?post=319228"}],"version-history":[{"count":6,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/posts\/319228\/revisions"}],"predecessor-version":[{"id":320186,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/posts\/319228\/revisions\/320186"}],"wp:attachment":[{"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/media?parent=319228"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/categories?post=319228"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dreness.com\/blog\/wp-json\/wp\/v2\/tags?post=319228"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}