\fay and \fax not working properly with tags in the line

[GoogleCodeExporter]

Using \fax or \fay in a line and using \c, \alpha, \fs, \fsp, \fscx, \fscy, 
\bord and \shad tags in any other place than the start of the line gives a line 
with the right \fax or \fay on the whole line but getting shifted (and can be 
shifted by a lot with high values of \fax or \fay). it isn't much annoying with 
most of these tags but with \c it makes the use of per character color (for 
gradient purposes for example) difficult.

Output using libass/xy-vsfilter is on the attached files (working fine with 
libass, shft with xy-vsfilter)

Test were made with xy-vsfilter 3.0.0.144 (Aegisub 3.0.2) on Windows 8 x64

An exemple of one script where you can see the shifting
{\fay0.2\c&HB7AAAD&\pos(813.943,399.754)}Class{\1c&H00&} 1-1

Original issue reported on code.google.com by sylviort...@gmail.com on 9 Mar 2013 at 5:55

• Merged into: #76

Attachments:

• libass.png
• xy-vsfilter.png

[GoogleCodeExporter]

i forgot to say it also fails with \blur, \k, \t (if you use any of the tags 
that starts this issue in it), \i, \b, \u, \s and also the x and y directional 
\bord and \shad (so basically all the tags that can be defined several times in 
one line)

Original comment by sylviort...@gmail.com on 9 Mar 2013 at 6:07

[GoogleCodeExporter]

This is a known limitation of the \fay implementation which was added to 
VSFilter way back in 2007. You can jiifurusu's post in the duplicate Issue #76 
for a way to workaround this issue somewhat.

For your example of;
{\fay0.2\c&HB7AAAD&\pos(813.943,399.754)}Class{\1c&H00&} 1-1

A near-identical alternative without this VSFilter issue would be:
{\fax0.2\frz-11\c&HB7AAAD&\pos(798.943,423.754)}Class{\1c&H00&} 1-1

This is still something which we'd be reluctant to look into fixing unless a 
few things happened first:

Examples provided of scripts & workarounds in the wild which may break from 
fixing this

Aegisub devs unanimously decide they want this fixed in VSFilter, rather than 
leaving things how they are

Users of Aegisub are strongly in favor of this being fixed, rather than not

Someone with repo access is willing to backport/merge any fix into MPC-HC 
VSFilter as well as VSFilter 2.39 (guliverkli2) repositories

Original comment by cyber.sp...@gmail.com on 9 Mar 2013 at 7:21

• Changed state: Duplicate
• Added labels: Priority-Low, Type-Enhancement
• Removed labels: Priority-Medium, Type-Defect