Possible bug in makexindex

[Frank Mittelbach]

Am 07.08.24 um 23:09 schrieb Zdenek Wagner:
> Maybe I am too sleepy but it seems to me that the example on

I think too sleepy :-)


> StackExchange uses \vspace in the horizontal mode, thus it does not
> expand to a simple \vskip but to \vadjust. What happens if you use
> \vskip 0.5em instead of \vspace{0.5em}?


the question is not what is in these strings but which strings are inserted.

the problem in the report that Lars prepared is:  when is item_01 and 
when is item_1  inserted

and it looks as if item_1 is inserted when it should have been item_01

(below I talked about item_01 and _item_x1 but it is really item_01 that 
is replaced by item_1 when it shouldn't -- thus also too sleepy :-)

good night
frank

> 
> st 7. 8. 2024 v 23:04 odesílatel Frank Mittelbach
> <frank.mittelbach at latex-project.org> napsal:
>>
>> Am 07.08.24 um 22:53 schrieb Karl Berry:
>>> Has to be plain TeX (actually initex, but I can probably do that part)
>>> in order to have any chance of running as a compile-time test. lm> That
>>> is not the impression I get from reading TLC3, part II page 358. Does
>>> the makeindex documentation have anything to say about this case?
>>
>>
>> the man page contains more or less what TCL contain here
>>
>> see entry for item_01 and item_x1
>>
>> I think it is pretty much saying what should happen and that doesn't
>> happen in the example
>>
>>    ie if there is
>>
>>     item
>>       subitem
>>
>> item_01 should be used
>>
>> and if it is
>>
>>     item <page-number(s)>
>>       subitem
>>
>> then item_x1 should be used instead (and same for 12 level entries,
>> which may or may nort have the same bug)
>>
>> frank
>>
>> sorry no time to generate a plain tex example (not done with my keynote
>> and close to lave for the US)
>>
>> frank
>