RE: Question about mustEmitSingleEntryDirectory tests

Hi Khaled,

Thank you for the status update and really happy to see that you made a great progress, congrats!
Yes, test one appears to be the only one sensible option here, I was tempted to simply modify the description of the test but decided to postpone it until the WG call next week.

As far as the number of tables is concerned, I guess the motivation behind using just one shared table that is known in advance was to simplify the testing process itself but it is certainly possible to make the test more rigorous by having more than one table shared and simply do a full check of the table directory to make sure that none of the tables are duplicated. Let's discuss it during our call next week (meanwhile, we could also continue this discussion here on the email list).

Thank you,
Vlad


-----Original Message-----
From: Khaled Hosny [mailto:khaledhosny@eglug.org] 
Sent: Wednesday, July 22, 2015 7:56 PM
To: Levantovsky, Vladimir
Cc: WebFonts WG
Subject: Re: Question about mustEmitSingleEntryDirectory tests

Hi,

OK, that makes sense. So only Test one is needed here, I assume the input font can have multiple shared tables or does it really need to be only one shared table?

Regards,
Khaled

On Wed, Jul 22, 2015 at 01:36:05PM +0000, Levantovsky, Vladimir wrote:
> Hi Khaled,
> 
> I've been reading the test description over and over again and can't really make sense of it as written. However, based on what the spec says - there are two different test scenarios where we need to check that:
> - the TableDirectory entry for a shared table is not duplicated, and
> - the table itself appears only once in the output bytestream.
> 
> Both cases are covered by their respective individual tests ("mustEmitSingleEntryDirectory" and " mustNotDuplicateTables") so I guess the "Test two" description of "mustEmitSingleEntryDirectory" could be a lingering mistake that went unnoticed until now.
> 
> Thank you,
> Vlad
> 
> 
> 
> -----Original Message-----
> From: Khaled Hosny [mailto:khaledhosny@eglug.org]
> Sent: Tuesday, July 21, 2015 7:53 PM
> To: WebFonts WG
> Subject: Question about mustEmitSingleEntryDirectory tests
> 
> The description of the second test in [1], confuses me a bit: It says it is not a font collection but it has a shared table, but what kind of sharing do non-collection fonts have?
> 
> Regards,
> Khaled
> 
> 1. 
> https://www.w3.org/Fonts/WG/wiki/TestPlan20-AuthoringTool#mustEmitSing

> leEntryDirectory
> 

Received on Thursday, 23 July 2015 14:23:56 UTC