Front page | perl.perl5.porters |
Postings from December 2001
I just got bit by a change 5.6->5.8 change in what \p{InTibetan} means.
In 5.6, \p{InTibetan} refers to a Unicode _block_.
In bleedperl, \p{InTibetan} now refers to the Unicode _script_.
Unicode scripts are superior to blocks... unless you were expecting block
semantis, as legacy code does. It turns out that
\N{TIBETAN MARK GTER TSHEG}
is not part of the script, so I got bit.
Programming Perl says:
"Note that these 'In' properties are only testing to see if the
character is in the block of characters allocated for that script."
Since scripts are closer to general categories (which use 'Is') than to
blocks, it might be appropriate to keep \p{In...} as the block, while
adding \p{Is...} to refer to the script:
current
perl5.6 bleedperl *proposed*
------- ------------- -------------
InTibetan block script||block block
IsTibetan error script||block script
Tibetan error script||block script||block
Where
"script||block"
means:
"script if available, block otherwise"
This preserves existing semantics, keeps the Is/In distinction consistant,
yet allows easy access to the superior script concept.
Another benefit is that it allows us to get rid of the {TibetanBlock}
names, which are sometimes there and sometimes not (they are there when
there's a script with the same name, so you still have to go to the man
page every time to see which to use.)
Jeffrey
Thread Next