summaryrefslogtreecommitdiff
path: root/doc/UsingLibical.md
Commit message (Collapse)AuthorAgeFilesLines
* lots of spelling fixes found by codespellAllen Winter2021-11-281-12/+12
|
* doc/UsingLibical.md – typosДилян Палаузов2021-11-201-5/+5
|
* various - use https instead of http where we canAllen Winter2019-12-141-4/+4
| | | | | Update Eric's email Remove softwarestudio.com which no longer exists as a domain
* doc/UsingLibical.md - fixes code examplePatrick Elsen2017-07-111-7/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | While running the code example locally, I ran into a bug where icalcomponent_vanew() would crash, but only when I was using more than 4 variable arguments. Code to reproduce is: icalcomponent_vanew( ICAL_VEVENT_COMPONENT, icalproperty_new_dtstamp(atime), icalproperty_new_dtstamp(atime), icalproperty_new_dtstamp(atime), icalproperty_new_dtstamp(atime), icalproperty_new_dtstamp(atime), 0); My machine is a 64-bit OS X machine. After a bit of fiddling and experimentation I think I found the cause, which is that on my machine, `sizeof(int) == 4` whereas `sizeof(void*) == 8`. In icalcomponent_add_children(), varargs are interpreted as `void *`, and checked for equality with 0 (therefore, the is automagically cast to a `void*`. As to why this only happened when there are more than 4 varargs, my best guess is that since the first N arguments are passed via registers on x86-64, it would be passed as a zeroed-register, whereas any arguments after that would be passed on the stack, where obviously anything that is on the stack after the 4-byte 0 would make this code crash. This should probably be marked in the documentation for the vanew functions (if it hasn't already).
* doc/UsingLibical.md - cleans up the example code, somewhatPatrick Elsen2017-07-111-326/+303
|
* doc/UsingLibical.md - converts doc/UsingLibical.txt to markdown formatPatrick Elsen2017-07-111-0/+1462