RE: Where could I read more about ipd/ipda/bpd/bpda values calculation?

2006-10-02 Thread Andrejus Chaliapinas
Hi Manuel,

> unfortunately for the purpose of understanding the area tree, the
> terminology, and the geometry the XSL-FO specification is your best
> friend

Yes, I've heard that before. Hope that support/extensive documentation will
grow up later. It's usually the case with dynamically progressing projects.

> I don't quite understand hwy you are so focused on the test cases. A
> test case is exactly that - it tests a specific piece of functionality.

I'm not so focused on them as it may seems, but good (and clear) test case
always helps me in my development life to kind of summarize what is achieved
so far. Due to my only initail involvement phase in FOP development - clear
test run (when you don't get warnings regarding feature is not implemented,
while you are running applied patch for that exactly feature) is a must for
me to see if project development (and test cases are part of that) just
follows some usual best practices and doesn't contain a lot of
incompleteness/half-of-way code.

> So we are getting back to the start. You need to understand what the
> area tree is suppose to look like for the feature you want to test.
> Then you can write the checks for the particular testcase.

Yes, will exactly do that now via everything I could find on that topic on
the Web.

Will ask more reasonable questions later.

Andrejus



Re: Where could I read more about ipd/ipda/bpd/bpda values calculation?

2006-10-02 Thread Manuel Mall
On Monday 02 October 2006 17:12, Andrejus Chaliapinas wrote:
> Hello,
>
> While some of you already told me that I should look through XSL-FO
> for those calculations (Jeremias mentioned how I would get value of
> 14400) - I still would like to ask you if you could point me to any
> document (probably with several samples), which discusses that in
> more details.
>
> I'm working on table auto layout feature right now (well, try to see
> if I could help and if that help could be delivered in timely matter)
> and try to understand first what is the goal of testcases and how are
> they prepared. Very often values to be checked in them are blocks
> ipda/ipd. But in table_table-layout_auto_1.xml.at.xml Area Tree file
> I could see that all 3 blocks under further check to be run over them
> have the same ipd="149181" and ipda="151181", though different
> bpd/bpda/left-offset values. Should we place for those different
> values instead in the testcase file? I'm a little bit confused.
>
> Any help would be appreciated.
>
> Andrejus

Andrejus,

unfortunately for the purpose of understanding the area tree, the 
terminology, and the geometry the XSL-FO specification is your best 
friend (although it usually takes one a long time to come to real 
friendly terms with this piece of work). In many aspects does the FOP 
implementation of XSL-FO follow the nomenclature and abstract models in 
the XSL-FO specification. Most importantly areas and the area tree are 
the major concepts used in the XSL-FO spec to describe the XSL-FO 
semantics and FOP models these fairly faithfully.

I don't quite understand hwy you are so focused on the test cases. A 
test case is exactly that - it tests a specific piece of functionality. 
The layout engine testcases are based around the notion of checking 
that the FOP generated area tree matches the expectations. Which aspect 
of the area tree is tested depends on the feature the particular 
testcase is suppose to exercise. Only checks relevant to the feature 
tested are usually included.

So we are getting back to the start. You need to understand what the 
area tree is suppose to look like for the feature you want to test. 
Then you can write the checks for the particular testcase.

Test cases are typically not something you do iteratively and change or 
expand a lot. They should be small testing a very particular feature 
only.

Manuel