[dpdk-web] SSI or similar

[dpdk-web] SSI or similar

[Andriy Berestovskyy]

Hi Thomas,
Is there a chance to enable SSI (server-side includes) or any other scripting option, so we could reduce code duplication inside the html files?

Here is the SSI documentation for the Nginx:
http://nginx.org/en/docs/http/ngx_http_ssi_module.html

Andriy

Re: [dpdk-web] SSI or similar

[Thomas Monjalon]

Hi,

2015-12-14 21:38, Andriy Berestovskyy:
> Hi Thomas,
> Is there a chance to enable SSI (server-side includes) or any other scripting option, so we could reduce code duplication inside the html files?
> 
> Here is the SSI documentation for the Nginx:
> http://nginx.org/en/docs/http/ngx_http_ssi_module.html

Thanks for the suggestion.
The current flat HTML files can be tested locally with just a browser.
How a change would be tested with SSI?

Re: [dpdk-web] SSI or similar

[Andriy Berestovskyy]

> On 14 Dec 2015, at 21:53, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> Thanks for the suggestion.
> The current flat HTML files can be tested locally with just a browser.
> How a change would be tested with SSI?

At the moment all the links inside the HTML files are not relative (i.e. start with a slash), so browser as is does not render the pages correctly anyway.

Having SSI enabled we could separate the content from the rest of the page, say:
/dev.html: // the whole page to test with an HTTP server
  includes inc/header.html
  includes inc/menu.html
  includes content/dev.html // flat HTML file to test with just a browser
  includes inc/footer.html

So the „content" directory could be used to store flat HTML files, but without any headers/menus/footers.


Alternatively, we could build (i.e. concatenate) HTML pages using a makefile or a script. So we could have a structure as above without any code duplication. Then we could build the pages with headers and menus to test them locally or to upload to a web-server.


Andriy

Re: [dpdk-web] SSI or similar

[Thomas Monjalon]

2015-12-14 22:51, Andriy Berestovskyy:
> > On 14 Dec 2015, at 21:53, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> > Thanks for the suggestion.
> > The current flat HTML files can be tested locally with just a browser.
> > How a change would be tested with SSI?
> 
> At the moment all the links inside the HTML files are not relative (i.e. start with a slash), so browser as is does not render the pages correctly anyway.

Yes you are right.
And the links do not use the .html suffix, it is found automatically
by the server.

> Having SSI enabled we could separate the content from the rest of the page, say:
> /dev.html: // the whole page to test with an HTTP server
>   includes inc/header.html
>   includes inc/menu.html
>   includes content/dev.html // flat HTML file to test with just a browser
>   includes inc/footer.html
> 
> So the „content" directory could be used to store flat HTML files, but without any headers/menus/footers.

OK for the inc directory but not for content. I would prefer having
the content in the original file.

> Alternatively, we could build (i.e. concatenate) HTML pages using a makefile or a script. So we could have a structure as above without any code duplication. Then we could build the pages with headers and menus to test them locally or to upload to a web-server.

Yes a "make local" to allow testing would be nice.

I would suggest to do it in 2 steps:
1/ "make local" for testing
2/ split with server-side includes

Please do not submit such change before end of January.
We need to make the workflow living smoothly first.
Thanks

Re: [dpdk-web] SSI or similar

[Andriy Berestovskyy]

Ok, agreed.

Andriy

> On 15 Dec 2015, at 06:43, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> 
> 2015-12-14 22:51, Andriy Berestovskyy:
>>> On 14 Dec 2015, at 21:53, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
>>> Thanks for the suggestion.
>>> The current flat HTML files can be tested locally with just a browser.
>>> How a change would be tested with SSI?
>> 
>> At the moment all the links inside the HTML files are not relative (i.e. start with a slash), so browser as is does not render the pages correctly anyway.
> 
> Yes you are right.
> And the links do not use the .html suffix, it is found automatically
> by the server.
> 
>> Having SSI enabled we could separate the content from the rest of the page, say:
>> /dev.html: // the whole page to test with an HTTP server
>>  includes inc/header.html
>>  includes inc/menu.html
>>  includes content/dev.html // flat HTML file to test with just a browser
>>  includes inc/footer.html
>> 
>> So the „content" directory could be used to store flat HTML files, but without any headers/menus/footers.
> 
> OK for the inc directory but not for content. I would prefer having
> the content in the original file.
> 
>> Alternatively, we could build (i.e. concatenate) HTML pages using a makefile or a script. So we could have a structure as above without any code duplication. Then we could build the pages with headers and menus to test them locally or to upload to a web-server.
> 
> Yes a "make local" to allow testing would be nice.
> 
> I would suggest to do it in 2 steps:
> 1/ "make local" for testing
> 2/ split with server-side includes
> 
> Please do not submit such change before end of January.
> We need to make the workflow living smoothly first.
> Thanks

Re: [dpdk-web] SSI or similar

[Thomas Monjalon]

Hi Andriy,

2015-12-15 06:43, Thomas Monjalon:
> 2015-12-14 22:51, Andriy Berestovskyy:
> > Alternatively, we could build (i.e. concatenate) HTML pages using a makefile or a script. So we could have a structure as above without any code duplication. Then we could build the pages with headers and menus to test them locally or to upload to a web-server.
> 
> Yes a "make local" to allow testing would be nice.
> 
> I would suggest to do it in 2 steps:
> 1/ "make local" for testing
> 2/ split with server-side includes
> 
> Please do not submit such change before end of January.
> We need to make the workflow living smoothly first.
> Thanks

Are you volunteer to write such script? Or anyone else?
I'd love to have a "make local" command to test website updates.

Re: [dpdk-web] SSI or similar

[Andriy Berestovskyy]

Hi Thomas,

> On 27 Jan 2016, at 15:01, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> Are you volunteer to write such script? Or anyone else?
> I'd love to have a "make local" command to test website updates.

Sure, no problem. I’ll do this on the weekend.

Andriy

Re: [dpdk-web] SSI or similar

[Thomas Monjalon]

2016-01-27 21:50, Andriy Berestovskyy:
> Hi Thomas,
> 
> > On 27 Jan 2016, at 15:01, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> > Are you volunteer to write such script? Or anyone else?
> > I'd love to have a "make local" command to test website updates.
> 
> Sure, no problem. I’ll do this on the weekend.

Great!
Which language would you like to use? (I do not like perl)

Re: [dpdk-web] SSI or similar

[Andriy Berestovskyy]

Hi  Thomas,

> On 28 Jan 2016, at 08:50, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> 
> 2016-01-27 21:50, Andriy Berestovskyy:
>> Hi Thomas,
>> 
>>> On 27 Jan 2016, at 15:01, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> Which language would you like to use? (I do not like perl)

I don't think we should invest time in making a generic HTML tree parser. So I would do this with a makefile + shell + sed changing hrefs, srcs etc.

Another option is to adopt wget or similar tool. I'll have a look if it would be possible.

What do you think?

Andriy

Re: [dpdk-web] SSI or similar

[Thomas Monjalon]

2016-01-28 10:48, Andriy Berestovskyy:
> Hi  Thomas,
> 
> > On 28 Jan 2016, at 08:50, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> > 
> > 2016-01-27 21:50, Andriy Berestovskyy:
> >> Hi Thomas,
> >> 
> >>> On 27 Jan 2016, at 15:01, Thomas Monjalon <thomas.monjalon@6wind.com> wrote:
> > Which language would you like to use? (I do not like perl)
> 
> I don't think we should invest time in making a generic HTML tree parser. So I would do this with a makefile + shell + sed changing hrefs, srcs etc.
> 
> Another option is to adopt wget or similar tool. I'll have a look if it would be possible.
> 
> What do you think?

I totally agree with simple shell + sed solution :)