Don't go into an infinite loop when trying to convert positions with w idth of 0. Hopefully fixes issue 6005

Beta 8 does not have crashes in OffsetOfWrappedLine. Increase testing fraction to 10% since I haven't seen any crashes yet in the nightly build

Don't search more than 500,000 cells in a very long line to avoid hangs. This is tricky to split up into multiple searches since lines can change.

- Enable it for nightly build users
- 1% of lookups are checked for correctness, leaving a crashlog on failure
- Clear the DWC cache when dropping a line. I beleive this was the cause of the assertions in beta 6 and 7.
- Make the check for nightly build faster

Disable number-of-wrapped-lines cache. There's a very common crash in OffsetOfWrappedLine starting in beta 6. This is the likely culprit. Issue 5980

Fix a bug in regex search where a regex like xa* would return multiple search results against a document like xaaaa (xa, xaa, etc.). Now that search results are properly sorted, this causes strange behavior. For example, only part of a match might be selected. Now, we ignore results that are a subset of a previous result. This works because the search progressively shortens and re-searches the document (e.g., xaaa, xaa, xa, x). Issue 5937

Make a copy of the raw line block data in -[LineBlock dictionary]. It looks like OS 10.13 needs the data in the coder asynchronously. This was always a risky optimization. It used to work because nothing could be mutated between willEncode and the actual encoding, but it seems that is no longer the case.

Add a cache to LineBlock that remembers where double-width characters that get moved to the next line exist. The cache is off by default except for nightly build users. For others it can be enabled with an advanced pref. Issue 5855.

Limit the number of characters that are searched for a word to 100k (was unlimited) and also don't search for the bounds of a wrapped line farther than they could possibly extend when a limit is present. This helps but does not completely fix some hangs when there is a very long line in a narrow window. Issue 5855.

Add smart case sensitivity to find on page as the default option. Automatically upgrade users from case-insensitive substring search to smart case sensitivity. General style cleanup of find view controller code. Issue 4690

Revert "Improve performance of enumerating characters by adding enumeration APIs to LineBlock, LineBuffer, and VT100Screen. Converting ascii screen_char_t to NSString was also optimized."

This reverts commit de3f4276.

Improve performance of enumerating characters by adding enumeration APIs to LineBlock, LineBuffer, and VT100Screen. Converting ascii screen_char_t to NSString was also optimized.

Fix a bug where automatic profile switching didn't work in fullscreen. This happened because when you ssh'ed the first RemoteHost interval would be on the last line in the terminal which was blank. APS would switch profiles, which would cause the screen to resize if the fonts were different sizes. When converting the ranges of these intervals from their source width to their destination width, they would be thrown out because they were on blank lines and when trimmed their length was 0. The fix is to tolerate these ranges being empty (their actual width is irrelevant). There was another issue in -[LineBuffer coordinateForPosition:width:ok:] where, when given a position equal to the end of the line buffer, it would return a coordinate that was at or after all the empty lines at the end of the line buffer. This commit fixes that problem as well, since it would have incorrectly moved the RemoteHost we're trying to fix down to the bottom of the screen.

I only modified those that were in the Headers group in Xcode.

Allow the cursor to enter the right margin when restoring from scrollback buffer. Fixes a bug where bash prompts get screwed up when resizing the window to smaller than the screen's width. Issue 3788. Also fix a few test failures and improve the output of compactLineDumpWithHistoryAndContinuationMarks.

Conflicts:
	sources/CaptureTrigger.m
	sources/PTYSession.m
	sources/Trigger.m

Add the ability to save history with window state restoration. Also up the deployment target to 10.7.

Move source into sources/ directory. Delete a few unused images in the process and fix a reference to an Aqua image in FindView

Remember the character attributes at the end of each line to fill unspecified cells when restoring from line buffer. Whie in the grid, it's stored in the last column where the continuation mark lives. When stored in the line buffer it goes in a new metadata per-line structure alongside timestamp. It is just a screen_char_t.

Improve performance when looking for a URLAction on a very long line. Limit the number of characters considered and optimize OffsetOfWrappedLine for the no-DWC-present case.

Increase warning levels and fix warnings. The only inexcusably disabled warning (possibly) is for comparisons of different signs, simply because there are so many of them, in large part because methods like -count and -length return unsigned integers. IMO, the proliferation of unsigned ints is more dangerous than turning off the warning.

Improve performance on very long lines by assuming there are no DWC's until one is seen, which allows a faster algorithm for counting the number of lines since we don't care about DWCs that get wrapped to the next line.

Introduce annotations on selections (notes) and improve marks. See intervalTrees branch for the individual commits.

Fix bug where DVR frames were encoded incorrectly. Work around bug where sometimes strings report a crazy (negative) length, causing search to go into an infinite loop in line buffer. Fix bug where divisionView was in the wrong place when a bottom bar is present.