Index: ossp-pkg/str/TODO
RCS File: /v/ossp/cvs/ossp-pkg/str/TODO,v
co -q -kk -p'1.20' '/v/ossp/cvs/ossp-pkg/str/TODO,v' | diff -u /dev/null - -L'ossp-pkg/str/TODO' 2>/dev/null
--- ossp-pkg/str/TODO
+++ -	2024-05-20 03:07:38.417892667 +0200
@@ -0,0 +1,65 @@
+   ____  _        
+  / ___|| |_ _ __ 
+  \___ \| __| '__|
+   ___) | |_| |   
+  |____/ \__|_|   
+                
+  Str - String Library                 
+
+  TODO
+
+  o look at printf ideas from ../stdio/printf!!
+  o look at ../strnatcmp
+   
+  The following *have* to be done before 1.0.0:
+
+  o perhaps a str_shave() or str_strip() function which
+    allows one to strip off characters out of a set from
+    begin and end of a string. Usually for zapping whitespaces
+    from begin and end of string.
+    => str_shave(" \t foo bar ", " \t\n") => "foo bar"
+
+  o use stdio-style "t" strings insead of STR_TRIGRAPH bitmasks
+  o finish str_test.c: more tests
+  o Dean wrote: btw -- str_compare taking a mode argument doesn't let you trim a few
+    cycles by going into strcmp/strcasecmp which are usually hand-tuned
+    assembly beasts on most boxes.  i'd go for str_compare and
+    str_casecompare.
+
+  The following *could* be done before 1.0.0:
+
+  o Unicode/UTF-8 support!!
+  o str_encode/str_decode:
+    - plain base64
+    - strict base64 (with newlines and padding, etc.)
+  o str to/from symbol conversion:
+    - long str_symbol(char *str);
+  o str_format merged with strptime
+  o str_locate could become a "mode" argument which
+    indicates whether "s" is small or large and if
+    it is small a (faster) brute force search could be done.
+  o str_parse could support s/.../g or even m/.../g 
+  o one could add "vstr" support with own malloc
+    Dean wrote: have you seen djb's stralloc stuff? it's part of qmail.
+    it looks like a nice way to avoid thinking about lengths.
+
+> I really like your Str library, but I have a question:
+>
+> What is its equivalent og the Perl 'g' modifier, such as in this example
+> (Perl):
+>
+> $teststr = 'hello/world/nice/to/meet/you';
+> $teststr =~ s/\//:/g;
+> # $teststr == 'hello:world:nice:to:meet:you';
+>
+> I couldn't find it in the documentation. Has this been implemented yet?
+> If not, is there some other way of doing it?
+
+Hmmm... yes, I initially left our the /g variant because I thought just
+about _matching_ like in m/.../g. But that was silly. You're right,
+for s/.../ the global flag is very important. It is currently still
+not implemented, but I've now added it to my TODO list for Str. OTOH
+your s/\//:/g usually can be performed more easier without str_parse,
+of course. Just use str_span() in a loop and replace the underlaying
+character in each step until str_span reached the end of the string.
+