Oh yes - log file/dump analyis ...
That was one of the things I was jealous of my Codewright colleagues.
I think 'static' lang. support could be better done taking the
prg.e example which also provides even more lang. features.
But it's clearly not that user friendly and resides in a dark edge ...
For (huge) logs/dumps with frequently changing 'patterns' to search for it's another thing (and a bit off-topic).
And sometimes it's not possible/useful to use a real debugger especially for driver/embedded development (as I curr. do) where
printf() is your friend.
So I thought about it and this idea came into my mind:
@Slickteam and friends:
If the process buffer ('Build tab' / .process buffer) is displayed in an MDI window and the error markers are set (implicitely by a 'Build' tool or by a user written
load_file_into_process_buffer_and_set_error_markers() macro) the error markers could be used to display s.th. useful in the 'Defs' tab which would implicitely provide some basic tagging support (list_tags()) too.
Thanks to the new error parsing support it's quite convenient to create / set / select the proper regexp to use.
The 'load...' macro could be added to the Build-tab context menu ('Send Compile output to Editor window' is already there).
And if e.g. a debug log file provides some filename/linenumber information we'd get a 'goto source from logfile' feature for free
(I already make use of __FILE__, __LINE__ (supported by most C compilers) for debug output and I think I'll try to write this 'load_...) macro mentioned above. Even w/o 'Defs'-tab support this could help me a lot.)
Similar things apply to the '.searchN' buffers (SearchN-tabs) when displayed as MDI win.
The given matches could be used accordingly as described above providing a better outline / handling especially for larger search results.
In additon the 'Symbol' toolbar could display the matched location.
This could be a cool (and non-intrusive) feature making a difference.
And maybe it's not that hard to implement. I think a lot of the required functionality is already there.
What do you think about it ?
HS2