Hello,
I thought some people may have spare cycles and are willing to help with the documentation. Here is some updates I would like to get for modules:
1) adding ids for parameters, functions, and other sections of the docbook xml files. Now we generate the alphabetic indexes pointing to the online html readmes, would be better to point directly to the exact section. Practically, for example, sections such as next from async module:
<section> <title><varname>workers</varname> (int)</title>
should become:
<section id="async.p.workers"> <title><varname>workers</varname> (int)</title>
To get unique ids over all files, I suggest using following pattern:
[module name] [dot] [type] [dot] [title]
The [type] should be: - p - parameters - f - functions - m - mi commands - r - rpc commands - s - statistics
2) Add RPC commands to readmes -- some of modules already do, but many don't. There is a perl script that should get the list of RPC commands from source code, but it requires some patch to a perl lib, not working on all systems. Moreover, the output is quite minimalistic, not easy to improve the content or add examples. Last generated index is:
- http://www.kamailio.org/docs/docbooks/3.4.x/rpc_list/rpc_list.html
3) Document selects -- although their name are quite suggestive, a bit of text around won't harm.
- http://www.kamailio.org/wiki/cookbooks/devel/selects
Because 1) and 2) requires git access, the easiest way to contribute is to provide patches to the xml files. If anyone commits to do an extensive work on these tasks in short term, we may eventually arrange for git access. The 3) is in the wiki, everyone can register an account and then update the pages.
I would recommend these guidelines to be followed from now on by developers when adding new elements to the module documentation.
Cheers, Daniel
Hi, I've added a new event route to the tm module and would like to document it in the README. (I've already added it to the wiki) With regards to providing automatic indexing, can we standardise a 'type' for the event_route section of documentation. For example 'e': <section id="tm.e.branch-failure"> ... </section>
Also, the tm module docs are split into separate xml files, with a top-level section <section id="tm.functions" ...>. Should this format be kept and does there need to be a standard for the top-level section ids?
Finally, I'm adding the id's to the websocket module where the MI commands include a dot e.g. "ws.enable". Does this affect the id generation, should it be replaced with underscore <section id="websocket.m.ws_enable">?
Regards, Hugh
On 25/03/2013 16:50, Daniel-Constantin Mierla wrote:
Hello,
I thought some people may have spare cycles and are willing to help with the documentation. Here is some updates I would like to get for modules:
- adding ids for parameters, functions, and other sections of the
docbook xml files. Now we generate the alphabetic indexes pointing to the online html readmes, would be better to point directly to the exact section. Practically, for example, sections such as next from async module:
<section> <title><varname>workers</varname> (int)</title>
should become:
<section id="async.p.workers"> <title><varname>workers</varname> (int)</title>
To get unique ids over all files, I suggest using following pattern:
[module name] [dot] [type] [dot] [title]
The [type] should be:
- p - parameters
- f - functions
- m - mi commands
- r - rpc commands
- s - statistics
- Add RPC commands to readmes -- some of modules already do, but many
don't. There is a perl script that should get the list of RPC commands from source code, but it requires some patch to a perl lib, not working on all systems. Moreover, the output is quite minimalistic, not easy to improve the content or add examples. Last generated index is:
- Document selects -- although their name are quite suggestive, a bit
of text around won't harm.
Because 1) and 2) requires git access, the easiest way to contribute is to provide patches to the xml files. If anyone commits to do an extensive work on these tasks in short term, we may eventually arrange for git access. The 3) is in the wiki, everyone can register an account and then update the pages.
I would recommend these guidelines to be followed from now on by developers when adding new elements to the module documentation.
Cheers, Daniel
Hello,
On 4/5/13 12:51 PM, Hugh Waite wrote:
Hi, I've added a new event route to the tm module and would like to document it in the README. (I've already added it to the wiki) With regards to providing automatic indexing, can we standardise a 'type' for the event_route section of documentation. For example 'e':
<section id="tm.e.branch-failure"> ... </section>
it is ok, we can reserve e for event routes. I created a page on wiki to collect guidelines about wrinting docbook files for README:
- https://www.kamailio.org/wiki/devel/module-docbook-readme
Anyone feel free to add there.
Also, the tm module docs are split into separate xml files, with a top-level section <section id="tm.functions" ...>. Should this format be kept and does there need to be a standard for the top-level section ids?
My interest was initially for the elements that are referred in alphabetic indexes, the rest is more a decision of the developer, but if done it should try to keep a pattern.
Finally, I'm adding the id's to the websocket module where the MI commands include a dot e.g. "ws.enable". Does this affect the id generation, should it be replaced with underscore <section id="websocket.m.ws_enable">?
I think is ok to keep it as the name, not replacing with '_', the prefix is important to have unique names across modules and inside modules (in eventuality there is a function named xyz and a rpc command xyz).
Cheers, Daniel
Regards, Hugh
On 25/03/2013 16:50, Daniel-Constantin Mierla wrote:
Hello,
I thought some people may have spare cycles and are willing to help with the documentation. Here is some updates I would like to get for modules:
- adding ids for parameters, functions, and other sections of the
docbook xml files. Now we generate the alphabetic indexes pointing to the online html readmes, would be better to point directly to the exact section. Practically, for example, sections such as next from async module:
<section> <title><varname>workers</varname> (int)</title>
should become:
<section id="async.p.workers"> <title><varname>workers</varname> (int)</title>
To get unique ids over all files, I suggest using following pattern:
[module name] [dot] [type] [dot] [title]
The [type] should be:
- p - parameters
- f - functions
- m - mi commands
- r - rpc commands
- s - statistics
- Add RPC commands to readmes -- some of modules already do, but
many don't. There is a perl script that should get the list of RPC commands from source code, but it requires some patch to a perl lib, not working on all systems. Moreover, the output is quite minimalistic, not easy to improve the content or add examples. Last generated index is:
- Document selects -- although their name are quite suggestive, a
bit of text around won't harm.
Because 1) and 2) requires git access, the easiest way to contribute is to provide patches to the xml files. If anyone commits to do an extensive work on these tasks in short term, we may eventually arrange for git access. The 3) is in the wiki, everyone can register an account and then update the pages.
I would recommend these guidelines to be followed from now on by developers when adding new elements to the module documentation.
Cheers, Daniel