Anthony Ettinger wrote: > It works everywhere else (FF, Opera, Safari), but IE7 seems to not > like the "?" in a url, and wraps it within a table cell.
Stay tuned to seeing other browsers behave that way, too. According to Unicode line breaking rules, a break is permitted between a question mark "?" and an alphabetic character, even when no space intervenes. Although HTML or CSS specifications do not require Unicode conformance, there seems to be a trend of changing browsers to apply some of the line breaking rules. This is a long story; see http://www.cs.tut.fi/~jkorpela/html/nobr.html In practice, this means that there will be more and more situations where you want to use either the nonstandard HTML markup <nobr>...</nobr> or its CSS counterpart, white-space: nowrap, when your text contains special characters (or, as here, punctuation marks in contexts other those of a natural language). Despite the name, the white-space property affects line breaking in general, not just whitespace handling. (Even the CSS 2.1 draft says that "This property declares how whitespace inside the element is handled", painting a wrong picture, but the description of the nowrap value says "suppresses line breaks within text".) > Markup is a cell in a table like this: > > <td class="pan"><a > href="foo.com?long_url=here">foo.com?long_url=here</a></td> First of all, in practical terms, it might be best to avoid the problem. URLs should be avoided in document content, as opposite to normal descriptive link texts. But sometimes you wish to write _about_ URLs, and then you need to mention them. > CSS is: > > td.pan { overflow: hidden; width: 12em; } > td.pan > *:first-child { display: block; overflow: hidden; width: > 12em; } Actually this doesn't quite the demonstrate the problem, since your sample URL is shorter than 12em, in any normal font. Anyone who wants to test your example needs to modify it to see the effect. > So I tried entity encoding the question mark as "<" That can't help, since "<" means exactly the same as "?". It is converted during the lexical processing of HTML markup, so the notational difference cannot have an effect on rendering. > I also tried { white-space: nowrap; } on the td CSS rule, but that > didn't do it either...again this is only IE7. This is where things get really interesting. In a later note you added: "Adding it to both rules worked" And in fact, it has the desired effect when placed in the second rule, td.pan > *:first-child { ... }. You probably first tried it in the first one, then added to the second, and didn't check what happens when you put it in the second one _only_. It seems that you have encountered a serious bug in IE 7. By CSS specifications, the white-space property is inherited, so the <a> element should inherit it from its parent, <td>, since no style sheet assigns a white-space value for the <a> element. This seems to be a general bug in IE 7. If, for example, we have <span>foo <i>zip zap</i> bar</span> and we assign white-space: nowrap to the <span>, then the effect should extend to the inner element's content "zip zap" as well. However, IE 7 breaks between "zip" and "zap" when it sees that fit. Even if we think that "suppresses line breaks within text" does not mean _all_ text in the content of the element but just directly contained text, inheritance rules say that inner elements inherit the value. However, we could still characterize the IE 7 behavior as "feature, not bug", if we (rather articially) describe it so that the browser style sheet contains the rule * { white-space: normal; } This is different from the fact that normal is the initial value. Such a hypothetical browser style sheet prevents any inheritance for the white-space property. This bug has serious implications for any use of white-space: you need to set it for inner elements as well. I guess it generally suffices to replace foo { white-space: nowrap; } by foo, foo * { white-space: nowrap; } for any selector foo. It appears that IE 7 implements <td nowrap> in a manner that does not quite correspond to a white-space: nowrap declaration for the cell element, as we might expect. It affects whitespace handling but not e.g. line breaks after "?". The nonstandard <nobr>...</nobr> markup is more effective. Thus, using the CSS white-space property is the simplest way _provided_ that you take a precaution against the lack of inheritance on IE 7. Jukka K. Korpela ("Yucca") http://www.cs.tut.fi/~jkorpela/ ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
