silly little script<\/a> to do that at a file level only. Example output:<\/p>\n\r\n% patchdeps\r\nNote: This is based on filename collisions only, so may overreport conflicts\r\nif patches touch different parts of the same file. (TODO)\r\n \r\nA bug-663281-deque X * * * * * * \r\nA bug-663281-deque-test | : : : : * : \r\nA bug-642054-func-setline X | * : : : : : \r\nA bug-642054-js_MapPCToLineNumber--' | * : : : : : \r\nA bug-642054-rwreentrant | : X : : : : : \r\nA algorithm--------------------------' X | * * * * * \r\nA system-libunwind X | | : * : * : * * : \r\nA try-libunwind------------------------' | | : X : * : * * : \r\nA backtrace------------------------------' | X * * * | * : * * * : * * * *\r\nU shell-backtrace | | : * : | : : : : : : : : : :\r\nU M-reentr---------------------------------' | : : : | : : : : : : : : : :\r\nU M-backtrace--------------------------------' X : : | : : : : : : : * : :\r\nU activities-----------------------------------' X : | : : : : * * : X * *\r\nU profiler---------------------------------------' X | * : * * X * * | * *\r\nU bug-675096-valgrind-jit--------------------------' | * : * : | : : | : :\r\nU bug-599499-opagent-config--------------------------' X * : * | * : | : :\r\nU bug-599499-opagent-----------------------------------' X X * | : * | : :\r\nU bug-642320-gdb-jit-config------------------------------' | * | * : | : :\r\nU bug-642320-gdb-jit---------------------------------------' X | : * | : :\r\nU import-libunwind | | : : | : :\r\nU libunwind-config-------------------------------------------' | X X | : :\r\nU warnings-fixes-----------------------------------------------' | | | : *\r\nU bug-696965-cfi-autocheck---------------------------------------' | | X :\r\nU mystery-librt-stuff----------------------------------------------' | | :\r\nU bug-637393-eval-lifetime | | :\r\nU register-dwarf-----------------------------------------------------' | :\r\nU bug-652535-JM__JIT_code_performance_counters-------------------------' X\r\nU JSOP_RUNMODE-----------------------------------------------------------'\r\n<\/pre>\nHow to read it: patches that have no conflicts earlier in the stack are shown without a line next to them. They’re free spirits; you can “sink” them anywhere earlier in your queue without getting conflicts. (The script removes their lines to make the grid take up less horizontal space.)<\/p>\n
Any other patch gets a horizontal line that then bends up to show the interference pattern with earlier patches. All in all, you have a complete interference matrix showing whether the set of files touched by any patch intersects the set of files for any other patch.<\/p>\n
‘X’ marks the first conflict. After that, the marker turns to ‘*’ and the vertical lines get broken. (That’s just because it’s mostly the first one that matters when you’re munging your queue.)<\/p>\n
So the patch named “backtrace” conflicts with the earlier “algorithm” patch, as well as the even earlier “bug-642054-js_MapPCToLineNumber” and others. The “M-reentr” patch only touches the same stuff as “bug-642054-rwreentrant” (not surprising, since “M-…” is my notation for a patch that needs to be folded into an earlier patch.) “system-libunwind” doesn’t conflict with anything earlier in the queue, and so can be freely reordered in the series file to anywhere earlier than where it is now — but note that several later patches touch the same stuff as it does. (It happens to be a patch to js\/src\/configure.in.)<\/p>\n
Useful? Not very. But it was kinda fun to write and I find myself running it occasionally just to see what it shows, so I feel the entertainment value was worth the small investment of time. Though now I’m tempted to enhance it by checking for collisions in line ranges, not just in the files…<\/p>\n
I suppose I could make a mercurial extension out of it, but that’d require porting it from Perl to Python, which is more trouble than it’s worth. (Yes, I still use Perl as my preferred language for whipping things together. Even though I dislike the syntax for nested data structures, I very much like the feature set, and it’s still the best language I’ve found for these sorts of things. So phbbbttt!)<\/p>\n","protected":false},"excerpt":{"rendered":"
A little while back, I was again contemplating a tangled patch queue, considering how to rework it for landing. I thought it’d be nice to see at a very basic level which patches in the queue were going to be problematic, and which I could freely reorder at whim. So I whipped together a silly […]<\/p>\n","protected":false},"author":206,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[477,289,137],"_links":{"self":[{"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/posts\/300"}],"collection":[{"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/users\/206"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/comments?post=300"}],"version-history":[{"count":0,"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/posts\/300\/revisions"}],"wp:attachment":[{"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/media?parent=300"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/categories?post=300"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.mozilla.org\/sfink\/wp-json\/wp\/v2\/tags?post=300"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}