YamaKen
yamak****@bp*****
2005年 9月 16日 (金) 18:07:30 JST
ヤマケンです。
最近uimのコーディングスタイルについてちょっと議論がありましたが、
整形や細かい論理の記述スタイル( (!p) と (p == NULL) 等)は個々人
の好みが大きく異なっているので、そういった問題のない慣用表現を先
にまとめてみるのもアリかもしれないと思いました。
以下のように、頻出のコード表現について定型とそうすべき理由を書き
連ねておき、特に理由がない限りはそれに従うという形はどうでしょう
か?
SigSchemeでの井上さんとの議論が発端ですが、uimまでカバーしても問
題ないんじゃないかと思います。
doc/IDIOMS:
* Traversing a list
for (rest = lst; !NULLP(rest); rest = CDR(rest)) {
elm = CAR(rest);
<code body>
}
- The element extraction "elm = CAR(rest)" is not performed in the expression
part of 'for', but in the block. It is intended that a null list can be
applied to the idiom without adding special handlings for it.
- 'lst' is the target list. It varies to 'args', 'procs', 'names' and so on
in accordance with the circumstances.
- 'rest' indicates that it points to the tail part of the list through the
traverser.
If you want to indicate which rest of, append the original list name such
as 'rest_args'. Don't omit the 'rest' prefix since it is required to
indicate the variable's role in the idiom.
Direct overwriting of the list such as "args = CDR(args)" should be avoided
if no significant reason exist, because it may be misunderstood as pointing
to the original list always.
- 'elm' indicates that the object is extracted from a list.
It should not arbitrarily be named such as 'obj' or 'val' since the names
indicate neither its role in the idiom nor its content specification.
If any confusion or misunderstanding will not be involved around the code,
it may be named reflecting its content, such as "arg = CAR(rest_args)" or
"sym = CAR(rest)". Select better name case by case.
-------------------------------
ヤマケン yamak****@bp*****