m_Modified
bool Parser::CacheNeedsUpdate()
{
if (m_UsingCache)
{
wxCriticalSectionLocker locker(s_TokensTreeCritical);
return m_TokensTree->m_Modified;
}
return true;
}
But, In-fact we don't use cache at all, so should be removed too.After a short discussion about if we need the cache at all anymore I'd say if this turns out to be of no use anymore we should rather remove the whole cache facility, not only this variable. Otherwise caching is broken.
The cache was too big(eg. the codeblocks.cbp should have 100M memory size) to save to hard disk.The case wasn't saved to the project file abut another (binary based) file. So the size would not be an issue. The question is if we need this to improve speed of scanning through large projects...
The cache was too big(eg. the codeblocks.cbp should have 100M memory size) to save to hard disk.The case wasn't saved to the project file abut another (binary based) file. So the size would not be an issue. The question is if we need this to improve speed of scanning through large projects...
But we need more IO(read/write) time, when save to hard disk, we need loop the whole tree, their need more time.The cache was too big(eg. the codeblocks.cbp should have 100M memory size) to save to hard disk.The question is if we need this to improve speed of scanning through large projects...
So, I personally think this is not needed.I personally believe it would only make sense for system headers, like STL, BOOST or alike. Or simply the compiler's include path would be something where a cache might e useful.