1 |
> |
2 |
> The problem is that writing things in XML will be even harder than the |
3 |
> pretty clean syntax you proposed. |
4 |
> |
5 |
|
6 |
Maybe a little more laborious, but not really that hard. Take this for |
7 |
example: |
8 |
|
9 |
use_var={"name":"esd", |
10 |
"description":"Enable enlightenment sound daemon support.", |
11 |
"priority":"OPTIONAL", |
12 |
"depends":"media-sound/esound"} |
13 |
|
14 |
To write this in XML it would look like this: |
15 |
|
16 |
<usedatabase> |
17 |
|
18 |
<useflag name=esd> |
19 |
<description>Enable enlightenment sound daemon support</description> |
20 |
<priority>OPTIONAL</priority> |
21 |
<depends>media-sound/esound</depends> |
22 |
</useflag> |
23 |
|
24 |
</usedatabase> |
25 |
|
26 |
(OK, not quite as _pretty_, but I've seen worse. The syntax, though, is |
27 |
clean.) |
28 |
|
29 |
Also, I don't think it would be too painful because it's not like |
30 |
someone would have to spend all day editing/adding entries. It would be |
31 |
pretty easy to write a script to prompt for the name/desc/etc. and |
32 |
generate the XML. |
33 |
|
34 |
Benefits of using XML also include: |
35 |
- Rules, such as declaring "priority" flags, can be created in and |
36 |
enforced by a schema. |
37 |
- Validation in general. |
38 |
- Consistent formatting would also be easy with XML. |
39 |
- Flexibility. |
40 |
|
41 |
(Not to sound like a commercial.) |
42 |
|
43 |
|
44 |
|
45 |
Mikael Hallendal wrote: |
46 |
|
47 |
> tor 2001-12-06 klockan 19.12 skrev Zach Forrest: |
48 |
> |
49 |
>>Before using python as an example, I had actually played with the idea |
50 |
>>of using XML. Any thoughts on this? I know that compared to a plain text |
51 |
>>file both of these notions seem somewhat cumbersome, but there are a |
52 |
>>couple of arguments for using a more structured approach. One, |
53 |
>>programatically handling the data is easier. Using either python or XML |
54 |
>>(and a parser) allows named access rather positional access to the |
55 |
>>fields. It also makes the definition of "records" more clear and easier |
56 |
>>to understand for new developers (i.e. me). It wouldn't be too difficult |
57 |
>>to parse XML (even into a dictionary). |
58 |
>> |
59 |
> |
60 |
> Yes, I thought about using XML too, as you say, for parsing the file |
61 |
> it'll be better since we don't narrow it down to python. Any language |
62 |
> that can parse XML (most) can be used for tools and such. |
63 |
> |
64 |
> The problem is that writing things in XML will be even harder than the |
65 |
> pretty clean syntax you proposed. |
66 |
|
67 |
> |
68 |
>>As far as GConf is concerned, I think that the base system should have |
69 |
>>as few dependencies as possible. Python is already required, and, in my |
70 |
>>opinion, with its XML capabilities (or just using dictionaries) I |
71 |
>>believe a solution along these lines would be preferable. |
72 |
>> |
73 |
> |
74 |
> Yes, GConf is IMHO not an option today, perhaps in the future when it: |
75 |
> 1) is more stable/tested for these kind of things. |
76 |
> 2) has nice tools for editing the contents. |
77 |
> |
78 |
> Regards, |
79 |
> Mikael Hallendal |
80 |
> |
81 |
> |