Parsing VmtxTable fails for C:\Windows\Fonts\msyi.ttf

[roy-t]

@wldevries the new addition for the VmtxTable seems to have a bug, with the msyi.ttf font it fails because it tries to read past the end of the stream.

Can you try to run WindowsFontsTest.ShouldParseWindowsFonts and try to debug the issue? Unfortunately I'm not familiar with the VmtxTable, but for you its fresh in your head 😄 .

[roy-t]

It looks like it also happens in C:\Windows\Fonts\simsunb.ttf, all other fonts I have installed seem fine. Both are Chinese fonts so maybe its an issue with top->bottom fonts?.

Note that I've just pushed a new version to master, that highlights this issue better.

[wldevries]

After adding sanity checks that compares the expected table length and the actual table length there are problems in a lot of fonts. A similar issue occurs in the HmtxTable table.

I suspect the counters are maybe wrong: VheaTable.NumberOfVMetrics, HheaTable.NumberOfHMetrics and MaxpTable.NumGlyphs. At the moment I'm a bit at a loss though. They seem to be read at the correct table offset and in the correct format. Will investigate more later.

[wldevries]

I found the problem: I used the FontReader.Seek method which seeks from the beginning of the file to skip over reserved space.

https://github.com/roy-t/TrueType/blob/master/RoyT.TrueType/Tables/HheaTable.cs#L28

[roy-t]

Ah yes it uses this.BaseStream.Seek(offset, SeekOrigin.Begin); feel free to add an overload to FontReader.Seek that does a relative seek :)

[wldevries]

Can be closed now. Thanks for the support :-)

Also, readme still states netstandard 1.6 instead of 2.0

[roy-t]

Oops, I'll fix that over the weekend. No problem. Hope it solves your issues!

…

________________________________ From: Wouter de Vries ***@***.***> Sent: Friday, February 24, 2023 4:09:37 PM To: roy-t/TrueType ***@***.***> Cc: Roy Triesscheijn ***@***.***>; Author ***@***.***> Subject: Re: [roy-t/TrueType] Parsing VmtxTable fails for C:\Windows\Fonts\msyi.ttf (Issue #6) Can be closed now. Thanks for the support :-) Also, readme still states netstandard 1.6 instead of 2.0 — Reply to this email directly, view it on GitHub<#6 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAEYIERXWM4NW2S2PQX45XLWZDFLDANCNFSM6AAAAAAU5JJW2Y>. You are receiving this because you authored the thread.Message ID: ***@***.***>