survey - snippet engines - should snipmate be merged with ultisnips or xptemplate?

[Marc Weber]

Do we need at least three? Should they be merged so that more effort
will be put into less engines?

http://mawercer.de/~marc/vam/snipmate-survey.php

There are some small new feature requests such as
- making snipmate snippets be context aware
- adding priorities to snippets

And I'm wondering whether I should implement them or make the other two
widely used engines (xptemplate or ultisnips) read snipmate snippets
instead in some way - and put my effort into one of those projects.

Thanks for any feedback. You can read the results yourself following the
link on that page.

Thanks for participating forming your future :)

Marc Weber

[herm...@free.fr]

Hello,

Sorry for the delay, I'll try to be constructive regarding your question.
As the survey text box is too restrictive I'll elaborate here.

Marc Weber wrote:
> [...]

> Do we need at least three? Should they be merged so that more effort
> will be put into less engines?

I wouldn't put effort into simplistic engines.
snipmate strength was its placeholders engine, and the syntax for simple snippets. Alas it wasn't able to do much.

As far as I'm concerned, I need many advanced things from template engines:
- the possibility for one template-file to include another template-file, and with arguments ; and to have control-statements, at least the "if else endif"

- the possibility to have variation points => it shall be possible to override included template-files on user- or project-basis (typically, with mu-template I provide a default header for C/C++ source files ; any user can prefer another format by default like a doxygen "/**@file" header ; but when working in particular projects we need to be able to specialize the header to include the correct licensing text)

- the possibility to fetch information and interact with the user ("I see your file is named foobar.h, the name of your new class is foobar, isn't it ?")

- it should be able to fetch project information before the template expansion kicks in -- it's required in order to correctly support template-file overriding on (/per?) project-basis.

- surrounding templates are a plus.

- plus many small things: i18n friendly, folding friendly, etc.

> There are some small new feature requests such as
> - making snipmate snippets be context aware
> - adding priorities to snippets
>
> And I'm wondering whether I should implement them or make the other
> two widely used engines (xptemplate or ultisnips) read snipmate
> snippets instead in some way - and put my effort into one of those
> projects.

I'd do the later.

HTH,

--
Luc Hermitte
http://lh-vim.googlecode.com/