::/home/robh/unity/unityfiles/text/UnityWikiBugsGraveyard:1082968909:: Moved by SimonBooth on Mon 26 Apr 2004
<<< I've been thinking of adding a list of "pseudo-whitespace" characters to sqaure-brackets links that get removed when generating page names, just like actual whitespace. Hence, ["Huck" Marbury] would link to HuckMarbury, [Huck's Books] would link to HucksBooks, and Edwin Arnold-Mallory would link to EdwinArnoldMallory. Can anyone see any problems this would cause? -- RobHague >>>
Moved by SimonBooth on Mon 26 Apr 2004
<<<Same thing is I think making it confused about AllTheThngsIRead and All The Things I Read -- StevenKitson >>>
<<< This would be fixed by changing the regexp in spilt_title (in UnityWiki.py) to ([a-zA-Z])([A-Z]) - could someone with CVS access do the honours? (At least, that would fix the display of the pagename for square bracket links; you could also probably fix WikiWords with a little regular expression jiggery-pokery.) -- RobHague >>>
<<< would ([a-zA-Z])([0-9A-Z]) be a good idea - it would fix TheUnityTimesChronicleArticlesJanuary1923 -- SimonBooth >>>
<<< Looks good to me -- RobHague >>>
<<< Yarr. It be done. -- SimonBooth >>>
<<< Shiver-me-timbers, it be on the server. However, it doesn't work for single letters in some circumstances (eg, [All The Things IRead]); after a little experimentation, I realised that this is because the RE consumes that character after the space as well as the one before it. (([a-zA-Z])(?=[0-9A-Z])|([0-9](?=[a-zA-Z]))) seems to work OK. -- RobHague >>>
<<< Shiver me shurikens. Though he did sell out. 'tis changed (again). -- SimonBooth >>>
<<< I think I'm due for a keel-hauling; I forgot to mention that the above regular expression only matches one group, so the replacement string should be "\1 ", not "\1 \2". I've updated it on the SRCF version. Sorry. -- RobHague >>>
<<< OK; now the SRCF and CVS versions are in sync, and I think everything is working. Does this mean we actually have a fixed bug? Could it be? -- RobHague >>>