Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
It often happens when creating a new entry that the first paragraph dissapears from the view of the page, although it is in the source code. The usual repair is to write a line of two empty before the first paragraph and then everything OK. It is unpleasant to have such unexpected feature, which may be unnoticed by the new contributors and requires more chekcing from the creators.
This is maybe part of a more general bug:
for instance also sometimes (I don’t understand under which conditions exactly) displayed math where the double dollar signs are not separated by a blank like from the rest of the text lead to an “Application error” when trying to save.
And then we recently saw that anchors in headlines lead to faulty rendering when not separated by a line break.
So generally line breaks and blank lines seem to help.
So generally line breaks and blank lines seem to help.
It is logical that in incorrect string manipulation in programming boundaries are critical. So I would expect that it is not a single bug for all the cases you noted. Thank you for putting the things together, however. BTW, I heard you gave (an impressive and) lively lecture minutes ago :)
I think that unusual punctuation (such as semicolons and colons) can cause the first paragraph to disappear. I assume that something is triggering the code for when these are used as markup (which should only apply when the special character is the first non-space character, however). But I was never able to figure out a consistent rule for this bug.
Could someone post a link to a page where this has happened? Thanks.
There’s one that I put up somewhere in the history the various Sandbox pages, but I no longer know where.
Right, here’s one: Sandbox revision #101. (Rollback to look at the first line of the source.)
I’ve confirmed this using a commandline installation of maruku. I can reproduce it with the text:
aa: b
c
which produces:
<p>c</p>
(actually, neither the b nor the c are necessary. The space after the colon is necessary, as are the two characters before it.)
I’ve reported this to Jacques.
It’s not a bug, it’s a feature. From Jacques:
That syntax, where you have any number of lines at the beginning of the document, which are of the form:
alphanumeric key + ':' + any number of spaces + valueis used by Maruku for a number of things (e.g., for the metatdata of S5 slideshows). Of course, ’aa’ is a gibberish key, but aside from that, the principle is the same.
It’s not a bug, it’s a feature.
I am not sure if you are talking the same things anymore, but statistically good percentage of normal pragraphs starting directly from the starting line eats up the paragraph and the writer may not notice this. This can not be a feature.
and the writer may not notice this.
An unrelated but similar problem is the following: if one forgets, accidentally deletes or misprints the closing-tag of a definition/theorem/proof-environment, then the entire entry will terminate at that point. This, too, is often easy to miss, especially for longer entries. After editing one really has to scroll through the entire entry to see if everything is there as intended.
You can see how it is used for slideshows on the instiki instiki. It’s also documented on the Maruku features page that “Email style” headers allow you to specify meta-data for the document. So it is documented. We can argue that we don’t like it, but once one knows about it then it’s pretty easy to get round.
As for Urs’ bug, what would be quite nice would be a way to see the errors that Maruku generates when rendering a page, at least if the page was viewed straight after a “save” command. That’s because all this sort of thing is reported by Maruku, but at the moment the errors go in the apache error logs on the server. I’d say that that was a reasonable feature request.
The work-around is just a blank line at the beginning of the page, I think. I’ve been starting every page with one blank line on principle since noticing this.
It looks like
title: blah, blah, blah
could be what people have been asking for elsewhere, a way to specify a name of the page which is not the same as the name used in [[links]]. However, this is only supported in slideshows, so we’d have to ask for it to be supported in ordinary pages.
We can argue that we don’t like it, but once one knows about it then it’s pretty easy to get round.
Andrew, any wiki is having some text and some special symbols and special words. In LaTeX every command starts with backslash so commands are distinguished. I can understand that any user of some software should use the syntax from it and understand the syntax. But wiki environment has those things like brackets of various kinds tgo be interpreted as links and not as text. LaTeX commands are within dollar signs and alike to be distinguished as well. If one just cuts and pastes ordinary text for the introduction one should expect all be OK. Single colon is an ordinary symbol from 1950-s mechanical typewriters. It is a normal symbol in texts. I spent many times lots of time on slow connection correcting randomly my normal text by inserting various things until it displayed correctly. Was I supposed to know which page of maruku manual to read when this was bogging me down ?
I can understand that statements like category:people is not an ordinary text, and it looks to me that there are simultaneously more features to such a situation: first of all there is only one word before colon, which is spelled with small letter and which is a key word. It is unpleasant to be against such special usages.
Zoran, I am in no way saying that you are unjustified in your frustration with this. Even that when one knows about it then it’s a bit daft and could be better implemented to spot random colons. All I’m saying is that once we’ve grumbled then we should think about the best way to (a) avoid it ourselves, and (b) ensure that others know about it.
I agree that the Wiki syntax is a bit of a mess. All wiki-syntaxes are a mess. But that’s because they aren’t designed by one person who sat down for a long time and worked out a whole system. Rather they were developed by people who took a bit from this system, and a bit from that, and a bit from the other. So it’s not surprising that there is the occasional inconsistency. Also, this sort of system hasn’t been around for all that long so not all of the wrinkles have been ironed out.
So, no, you weren’t supposed to know which page in the manual to read. You did exactly what you were supposed to do - ask about it here (though if you had done it when you first spotted it, it might not have caused you so much grief).
By the way, not every command in LaTeX starts with a backslash. There are traps there for the unwieldy. If I wanted to say that something was approximately three dollars, I might naively type ~$3 and be a little bit surprised at the outcome.
(Edited for grammar)
LaTeX commands are within dollar signs and alike to be distinguished as well. If one just cuts and pastes ordinary text for the introduction one should expect all be OK. Single colon is an ordinary symbol from 1950-s mechanical typewriters. It is a normal symbol in texts.
The mechanical typewriter in my parents’ garage has a dollar sign as well as a colon. I realise that dollar signs don’t belong in ordinary text in every country, but from where I stand, this “email header” notation actually makes a lot more sense than LaTeX.
I just wish that we had something to do with it.
Toby, in my understanding, which is here much less informed than yours (you know all the encodings besides and more!), the ordinary text in most languages (including those in Greek, Latin, cyrillic and devanagari alphabets, at least) traditionally consists only of
A) Letters, say a,b,c,d…
B) Interpunctions: comma, dot, colon, round brackets, question-mark, semicolon and exclamation mark, dash or hyphen (there is no distinction in the writing of most languages, and I do not understand the one in English), quotation, apostrophe (maybe the list is too ling to be consider normal, at least quotations and appostrophes are usually of special meaning in computers!)
C) Digits
In my grammar school impression, everything else are special (financial, computer science, mathematical, chemical, meteorogical) and abbreviation symbols (including ampersand in English, not meaningful/allowed in standard Croatian for example), usually not in grammar books, and not universally valid – euro sign, copyright sign, paragraph sign, dollar sign, tilde for estimate, sigh signs, Bourbaki bend mark, masterspace (now called monkey), special (square etc.) brackets (square brackets maybe good for some languages like Japanese), plus, equality, less than, more than, computer symbols for emoticons…
Of course, accentuation has special status which is out of this discussion, I think.
Programming languages do not have the intention to have special text as a default so there unless in special syntax or sort of string environment, one does not expect a usual text to be a usual text. So they like to take say a semicolon or colon as a special symbol.
I mostly agree with your list of specials, but not the dollar sign. Possibly this is just an American thing, but it really does feel like an ordinary symbol. (The cent symbol does not, however, possibly because it doesn’t appear on the typewriter.)
this is just an American thing,
Yes, that’s because in other parts of the world it’s a habit to wreck the currency every 50 years or so, hence there is no use to set any currency symbols in stone – or in typewriter keys, for that matter. Dollars are outdated here as of late 16th century or so. For the next currency to be abandoned they invented the symbol “€”. No need to remember it… ;-)
Urs is usually optimistic about unification in mathematics, but seems not so in unifications in Europe :)
That’s why they invented the generic currency symbol “¤”.
Is that meant to represent a WW II floating mine, ready to explode on contact!
1 to 23 of 23