Oh yeah, I forgot to look there. Although now that I do, that looks really familiar. I think I might have tried to use it very early on but got confused by that array so just went off and wrote my own. Unfortunately, that bundle uses field_edits, which I've recently found to (currently) be incapable of handling some of my values properly. I've filed an RFE that describes my problem.
https://cfengine.com/bugtracker/view.php?id=174 Cfengine is missing one very critical component for complete CSV (or in this case, Equal Sign Separated Values) handling via field_edits: a text delimiter for values. An inserted text value cannot contain a character that matches the field delimiter (or presumably the value delimiter). -- /* Wes Hardin */ Maxim Integrated Products: Innovation Delivered On 07/01/2010 10:20 PM, Daniel V Klein wrote: > Heh - just use "set_variable_values" from cfengine_stdlib.cf! Or see how it > is done there - it's rather clever (I wish I thought of it :-) > > -Dan > >> Forum: Cfengine Help >> Subject: Looking for suggestions on ordering >> Author: Authority >> Link to topic: https://cfengine.com/forum/read.php?3,17628,17628#msg-17628 >> >> Normal ordering for edit_line bundles puts insert_lines before replace_patter >> ns. For my current problem, I'm trying to reverse them. >> >> Given a simple syntax like this: >> >> option = value >> >> I would like to check that the option exists and simply replace the current v >> alue with the new value and if the option doesn't exist, add the line. This >> is problematic though because the syntax of my files are flexible regarding w >> hite space but insert_lines is not. If my insert_lines promiser does not hav >> e any whitespace around the equals sign but the file does, or if the option i >> s indented, even though the option and value are as they should be, Cfengine >> will insert a new line. Replace_patterns is powerful enough to handle and pr >> eserve whitespace but I'm having trouble stopping the insert_line from happen >> ing after replace_patterns has already done its work. I only want the insert >> _line to happen if the replace_pattern couldn't find and replace the value. >> >> I've found that a replace_patterns promise that doesn't match anything (and t >> hus doesn't replace anything) is still marked as kept, which I can kind of un >> derstand (nothing to do is just as good as already being done), but it makes >> it very hard to determine if replace_patterns took any action. >> >> I've tried setting a class at the start of the bundle that blocked the insert >> _lines promises and was cancelled by replace_patterns, but that class was rea >> ctivated when the bundle was re-evaluated so my inserts never happened. >> >> I'm not going to argue that the normal ordering be changed as I think for mos >> t cases, it is optimal already, but I'm wondering if a new promise result cla >> ss attribute (for use in class bodies) might be worthwhile, something along t >> he lines of match_found, that would activate classes if the regex in the curr >> ent promise succeeded. >> >> Has anyone else dealt with such ordering issues? Any suggestions/tricks on h >> ow to do this? >> >> _______________________________________________ >> Help-cfengine mailing list >> Help-cfengine@cfengine.org >> https://cfengine.org/mailman/listinfo/help-cfengine > _______________________________________________ > Help-cfengine mailing list > Help-cfengine@cfengine.org > https://cfengine.org/mailman/listinfo/help-cfengine > _______________________________________________ Help-cfengine mailing list Help-cfengine@cfengine.org https://cfengine.org/mailman/listinfo/help-cfengine
