About the order of reading source files.

[Shigio YAMAGUCHI]

I have found a user visible change which is not discussed.

The order of reading file is different between GTAGS and GRTAGS(and GSYMS).
This behavior is different from former version.

% gtags -v
[Sat Aug 27 19:01:13 JST 2005] Gtags started.
 Using default configuration.
[Sat Aug 27 19:01:13 JST 2005] Creating 'GTAGS'.
 [1] extracting tags of gtags.c                 <- read gtags.c first
 [2] extracting tags of const.h
[Sat Aug 27 19:01:13 JST 2005] Creating 'GRTAGS'.
 [1/2] extracting tags of const.h               <- read const.h first.
 [2/2] extracting tags of gtags.c
[Sat Aug 27 19:01:13 JST 2005] Creating 'GSYMS'.
 [1/2] extracting tags of const.h
 [2/2] extracting tags of gtags.c
[Sat Aug 27 19:01:13 JST 2005] Done.

It seems that this is because createtags() uses both of find_xxxx() and
gfind_xxx() to read file names. Though using gfind_xxx() improves efficiency
by about 3%, this behavior seems not to be consistent.

I would like to return to former way because:
o Consistency is more important than 3% of efficiency.
o I cannot explain why the order of reading file is different because of
  tag file.

What do you think?
--
Shigio YAMAGUCHI <address@hidden> - Tama Communications Corporation
PGP fingerprint: D1CB 0B89 B346 4AB6 5663  C4B6 3CA5 BBB3 57BE DDA3