Since we use gnu_printf format on mingw, the windows "I" modifier
is no longer recognized by the format checker. Use "ll" instead.
Meanwhile, MSVC understands that as well.
This removes the language specification and BNF descriptions from the
PR release notes (they're too long for that) to a separate document
in the libcom/src/as area. That should be combined into the chapter
on Access Security from the AppDevGuide someday.
This imports the function-specifics that were documented in the
Application Developer's Guide, section "IOC Shell".
Other changes include:
- Removing unneeded "@typedef" and "@enum"
- Adding "()" and "@ref" for interlinking
- Removing duplicated documentation between
iocsh & iocshLoad, and iocshCmd & iocshRun
Supports the PR:
https://github.com/epics-docs/epics-docs/pull/141
Document that epicsSnprintf() adds a nil on success,
capacity permitting.
Note OS specific ambiguity on error. Linux and MSVC
inject nil for most failures (bad format).
Not clear if vxworks/fioFormatV() does.
Not clear if long description of non-compliant
return value is still accurate.
Corrects two issues. 012139638d
added a dec-ref in the status==EPERM path, but neglected to
add a matching inc-ref. So if a joinable pthread_create()
actually errored with EPERM, then no reference is left for the
joiner.
Secondly, and far more suble, is a data race if pthread_create()
succeeds in starting a new thread, which ends and free()'s
pthreadInfo before pthread_create() writes pthreadInfo->tid .
Most errors found with:
codespell -L cach,thst,odly,aslo,parm,parms,inpu,ges,prset,pevent,ptd,pring,valu,noo,noe,ned,inout,ro,siz,froms,nd,fo,singl,sart,multy,tthe,allong,ment,inate,nodel,tring,alse,ture,thi,wille,numer
Some more manually found (its -> it's)
c++20: Do not use apostrophe (e.g. can't) in unquoted #error message
