mozers: > 1. Fixed SciTe freezing when command.input is used. Before the fix, > the StdIn function launched from a script would wait for the data and > freeze Scite, if no text selection was made.
This patch pretends that there is one byte of input available. I can't see all the ramifications of this but I expect a single NUL is sent to the process which doesn't appear to me to be a good outcome. Probably its just trying to affect the logic so that the termination code is reached. I'd prefer something more explicit with less side effects here. > 2. Fixed error with text output (in UTF-8 only!) initiated by > command.input and $(CurrentSelection). It was not possible to use an > external script to process the text containing Cyrillic symbols. See > the example above. This change has many repercussions which is why it hasn't been done. One (but not the only) problem that is not addressed here is that the output pane may have a different encoding to the edit pane. > 3. Now each programming language can have its own abbreviations file!!! > For example: > abbrev.$(file.patterns.web)=$(SciteDefaultHome)\abbrev\html.abbrev > abbrev.$(file.patterns.php)=$(SciteDefaultHome)\abbrev\php.abbrev While having a per-language abbreviation file may be better, it should be completely optional, with current behaviour preserved for existing users. > (Why wasn't it in the first version SciTE? I do not understand :-/ Some abbreviations, such as names are independent of language and its easier to find and edit a single file. > 4. The autocomplete function now gets the templates both from the > edited text and the corresponding .api file! This decreases selectivity and may be a win for some people but its not my preference. > 5. The SQL folding is improved. > A new parameter fold.only.begin (for SQL) has been added. It declares > that folding should begin only after "begin". > E.g. in SyBase SQL a construction of type "if then -> end if" requires > fold.only.begin=0 > while in Firebird SQL a construction of type "if thet begin -> end" requires > fold.only.begin=1 > > (the best way is to use "fold.sql.only.begin" - amendment Neil Hodgson) Changed to default to 0 to preserve current behaviour and changed variable name to foldOnlyBegin to be more obvious. Not yet committed. > It is a pity that you ignore these sweets :( > Waiting for official version which has them. A large patch set is less likely to be examined by anyone as it is too much work and its less obvious what is there. Proposing a simple change like per-language abbreviations is more likely to interest people and thus lead to a conversation with other users. If no one else responds then I get to be responder-of-last-resort and my normal reaction is that if nobody else has expressed an interest then its unlikely I'll want them. Neil _______________________________________________ Scite-interest mailing list [email protected] http://mailman.lyra.org/mailman/listinfo/scite-interest
