Re: Run pgindent now?
От | Bruce Momjian |
---|---|
Тема | Re: Run pgindent now? |
Дата | |
Msg-id | 20150524023339.GE28512@momjian.us обсуждение исходный текст |
Ответ на | Re: Run pgindent now? (Andres Freund <andres@anarazel.de>) |
Ответы |
Re: Run pgindent now?
|
Список | pgsql-hackers |
On Sun, May 24, 2015 at 04:16:07AM +0200, Andres Freund wrote: > On 2015-05-23 21:36:50 -0400, Bruce Momjian wrote: > > pgindent run on HEAD and committed. > > - if (IsA(node, Aggref) || IsA(node, GroupingFunc)) > + if (IsA(node, Aggref) ||IsA(node, GroupingFunc)) > > There's a bunch of changes like this. Looks rather odd to me? I don't > recall seing much code looking like that? Wow, that is quite odd. I saw another case where it might have done this because the line was >80 characters without it, but not in this case. > Also, does somebody have an idea to keep pgindent from butchering inline > asm like: > /* > * Perform cmpxchg and use the zero flag which it implicitly sets when > * equal to measure the success. > */ > - __asm__ __volatile__( > - " lock \n" > - " cmpxchgl %4,%5 \n" > - " setz %2 \n" > -: "=a" (*expected), "=m"(ptr->value), "=q" (ret) > -: "a" (*expected), "r" (newval), "m"(ptr->value) > -: "memory", "cc"); > + __asm__ __volatile__( > + " lock \n" > + " cmpxchgl %4,%5 \n" > + " setz %2 \n" > + : "=a"(*expected), "=m"(ptr->value), "=q"(ret) > + : "a"(*expected), "r"(newval), "m"(ptr->value) > + : "memory", "cc"); > + > return (bool) ret; Ah, we have a file /pgtop/src/tools/pgindent/exclude_file_patterns which has excluded files, and s_lock.h is one of them. I think /include/port/atomics/arch-x86.h needs to be added, and the pgindent commit on the file reverted. Are there any other files with __asm__ lines like that? -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + Everyone has their own god. +
В списке pgsql-hackers по дате отправления: