| Ticket UUID: | 536838 | ||
| Title: | [subst] doc clarification | ||
| Type: | Bug | Created on: | 2002-03-29 18:39:25 |
| Submitter: | dgp | Assigned to: | dgp |
| Subsystem: | 18. Commands M-Z | Severity: | |
| Priority: | 5 Medium | Last modified: | 2002-04-18 23:31:54 |
| Status: | Closed | Closed by: | dgp |
| Resolution: | Fixed | Closed on: | 2002-04-18 16:31:54 |
| Version: | obsolete: 8.4a5 | ||
| Description: | ||||
It should be noted that the -no* options to
[subst] only influence the "top level" of
substitution. Even with -novar, variable
substitutions do take place with command
substitutions, and even with -nocommands,
command substitutions do take place within
variable substitutions (array indices).
% set a 1
1
% subst -novar {$a [string equal $a 1]}
$a 1
% proc b {} {return c}
% array set foo {c c b b [b] tricky}
% subst -nocommands {[b]$foo([b])}
[b]c
| ||||
| User Comments: | ||||
dgp added on 2002-03-30 01:57:46:
Logged In: YES
user_id=80530
It should also be noted that [subst] can only have a
return code of TCL_OK or TCL_ERROR. If a substitution
yields some other return code, notably TCL_RETURN, but
other extended return codes as well, they are effectively
converted to TCL_OK.
I seem to recall Brent Welch wishing for a means to
return extended return codes from [subst] for use in
web page template processing so that something like
subst {
# lots of template stuff with a
return -code 404 "Page not found"
# blah
}
could be used.
dgp added on 2002-03-30 02:28:10:
Logged In: YES user_id=80530 Actually, [return] acts much like [continue]. Said another way, [continue] is equivalent to [return ""]. Extended return codes too. dkf added on 2002-04-05 15:46:41:
Logged In: YES user_id=79902 Care to submit a detailed description of what you're after? dgp added on 2002-04-06 04:19:20:
File Added - 20637: sdoc.patch Logged In: YES user_id=80530 Here's a patch. dkf added on 2002-04-18 20:28:16:
Logged In: YES user_id=79902 Documentation change looks good to me; assigning back to you for application. (I was actually interested in what was meant by the handling of TCL_RETURN, though I'd like to note that I've always found it easier to pass that sort of thing using the ::errorCode variable.) dgp added on 2002-04-18 23:31:54:
Logged In: YES user_id=80530 committed to HEAD (8.4a5) | ||||
Attachments:
- sdoc.patch [download] added by dgp on 2002-04-06 04:19:20. [details]