Just put it before base, which is what's expected.

Malayalam failures down from 1559 to 1197 (0.114172%).

BENGALI: 353988 out of 354285 tests passed. 297 failed (0.0838308%)
DEVANAGARI: 693571 out of 693628 tests passed. 57 failed (0.00821766%)
GUJARATI: 366489 out of 366506 tests passed. 17 failed (0.0046384%)
GURMUKHI: 60750 out of 60809 tests passed. 59 failed (0.0970251%)
KANNADA: 950956 out of 951913 tests passed. 957 failed (0.100534%)
KHMER: 299094 out of 299124 tests passed. 30 failed (0.0100293%)
MALAYALAM: 1047219 out of 1048416 tests passed. 1197 failed (0.114172%)
ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%)
SINHALA: 271699 out of 271847 tests passed. 148 failed (0.0544424%)
TAMIL: 1091837 out of 1091837 tests passed. 0 failed (0%)
TELUGU: 970524 out of 970573 tests passed. 49 failed (0.00504856%)

The sequence Ra,H,Ya in Bengali is ambigious and Unicode encoded that to
get Ya-Phalaa, one would place ZWJ before Halant.  Ie. a ZWJ,H sequence
requests subjoining, while a H,ZWJ requests Half form.  Implement that.

Bengali failures go down from 377 to 297 (0.0838308%).
Gujarati is down by 4 to 17 (0.0046384%).
Kannada is down by 226 to 957 (0.100534%).

Current status:

BENGALI: 353988 out of 354285 tests passed. 297 failed (0.0838308%)
DEVANAGARI: 693571 out of 693628 tests passed. 57 failed (0.00821766%)
GUJARATI: 366489 out of 366506 tests passed. 17 failed (0.0046384%)
GURMUKHI: 60750 out of 60809 tests passed. 59 failed (0.0970251%)
KANNADA: 950956 out of 951913 tests passed. 957 failed (0.100534%)
KHMER: 299094 out of 299124 tests passed. 30 failed (0.0100293%)
MALAYALAM: 1046857 out of 1048416 tests passed. 1559 failed (0.148701%)
ORIYA: 42320 out of 42329 tests passed. 9 failed (0.021262%)
SINHALA: 271699 out of 271847 tests passed. 148 failed (0.0544424%)
TAMIL: 1091837 out of 1091837 tests passed. 0 failed (0%)
TELUGU: 970524 out of 970573 tests passed. 49 failed (0.00504856%)

Fixes another 1 Khmer failure.  Down to 30 (0.0100293%) now.

Khmer failures down from 39 to 31 (0.0103636%).

In Khmer coeng model, a V,Ra can go *after* matras.  If it goes after a
split matra, it should be reordered to *before* the left part of such matra.

Khmer failures down from 136 to 39 (0.0130381%).

Fixes another 6 Khmer failures.  Now at 136 (0.0454661%).

Khmer Bindu doesn't like to move to syllable end.  Leave it where it
was.

Brings down Khmer failures from 510 to 142 (0.047572%).

Fixes another 10 Sinhala failures.  Down to 148 (0.0544424%).

Apparently if there is C,V,ZWJ,C, the first C will be base, but if
it's C,ZWJ,V,C, the second one will be.

Note that Uniscribe implements this differently, by breaking syllable in
the case of C,ZWJ,V,C and putting the first consonant in one syllable
and the rest in the next syllable.

Sinhala failures down from 208 to 158 (0.0581209%).  No changes to
Khmer.

Sinhala does not have half forms.  And most (all?) consonants can be
base, except when preceded by ZWJ, which would request a subjoined form.
Hence switch the base algorithm to categorize with Khmer, start search
at start, and stop at a ZWJ.

Also, mark all pos=base consonants after base to be subjoined.  Mark
base itself to have pos=base.

Finally, adjust Sinhala's reph position to after-main.

Brings down Sinhala failures from 455 to 328 (0.120656%).

Only skip actual ligatures, not marks in-between ligature components.

If, say, a H,ZWJ,C ligature was formed, we don't want the code to detec
that as a Halant.  So, ignore ligatures when matching category in
final_reordering.

Sinhala failures down from 514 to 455 (0.167374%).

We now limit syllable lengths in the machine.  No need to match here.

Test case was: <U+0D15,U+0D4D,U+0D15,U+0D4A>.

This should address all possible cluster misformations that I had in
mind.

This should fix any instabilities in cluster formation that we were
speculating may happen with surrounding syllables.  Or most of it
perhaps.

Fixes crashes reported with left matra under
non-uniscribe-bug-compatibilty mode.

Improves Bengali and Gurmukhi.  Malayalam regressed a bit.  We will deal
with that later.

This is a bunch of hacks for now.

Improves Bengali a bit.

Fixes 20 out of 48 failing Oriya tests.  Failure rate down to 0.066% now.

Oriya failures down from 0.65% to 0.20%.

Fixes most Malayalam failures.  Down from 1.6% to 0.38% now.  Fixes a
few more in other scripts too.

Apparently this was approved in Feb 2012.  No font yet.

Now that we break syllables on Halant,ZWNJ, this code can be simplified.

That's really what Uniscribe does, and explains a lot of pecularities of
Halant,ZWNJ before the base.

Sent Telugu from 1% failures to 0.03%.  Improved Kannada and Malayalam
slightly.  Fixed half of Bengali, and did NOT break anything!

Specifically, don't apply 'init' if previous char is a joiner.

Fixes some more of Bengali.

Not tuned, just copied from step 2.  Fixes another 0.5% of Kannada
failures.  1% to go.

Fixes a few Devanagari, half of remaining Kannada failures, quarter for
Telugu, and others slightly improved or unchanged.

Fixes 5 Devanagari failures, and no regressions.

Brings down failures with Lohit-Telugu from 57% to 1.40%.

Kannada failures down from 3.5% to 2.93%.

It's not in IndicSyllabicCategory.txt.  Fixes most of Gurmukhi failures.
Failures down from 7.7% to 0.222%!

For Khmer, all consonants are subjoining.  No need to look in the font.
We were looking in the wrong order anyway.