"Greg Smith" <[EMAIL PROTECTED]> wrote: > I seem to have missed the start of this discussion. I have a couple of > observations based on reading the message on the list:
Some of it was a P.S. in one of my replies in the "smart tabulations" thread, other parts were discussed on or after July 24 in a thread named "Indicators", started by me. > 1. Are you planning on implementing a compile time choice of styling buffer > size, or a run time choice. I would much prefer a run-time choice (or was it > a compile time choice of including this feature)? Compile-time choice for including the feature was an option. I personally prefer #3, whose enumeration Neal stripped from his reply, which was essentially; add some API to allow for 1,2 or 4-byte style setting during runtime, all of the old code and lexers use the old API and don't know the difference (even if their internals are different). > 2. If you do this, a 0 size as well as 1, 2 and 4 would be very much > appreciated. As Neil says, there are folks with huge log files or huge data > tables (really... I have users who generate 100 MB and more of text). Sounds reasonable. I see no reason why it wouldn't make sense to include a 0 styling bytes option. > 3. I also assume that you will make sure that the API is unchanged for folks > using 1 byte (and possibly also 0 byte styling) as logically there is no > need to change it. However, the API obviously has to change for >1 byte (and > maybe for 0 byte). The API doesn't need to change. If we *add* API functions for the new functionality, then all of the old code can use the old API without change. This preserves the backwards compatability of all code, as long as styling bytes > 0. - Josiah _______________________________________________ Scintilla-interest mailing list [email protected] http://mailman.lyra.org/mailman/listinfo/scintilla-interest
