Fix parsing integers with defined length #195
Conversation
This is only a partial fix, because it doesn't work with integers that start with 0x or 0o.
|
That's an odd one. It seems not to work with the previous parser either and, from the help page, IRAF uses |
|
Could we therefore get rid of the |
|
Yes, this was also my thought: the iraf scanf code does not handle |
While this was implemented in the original PyRAF scanf module, its Python implementation didn't work well. Also it was not implemented in the original IRAF scanf function. Parsing of octal/hexadecimal strings is still possible by using the %o ond %h formats, just like in IRAFs scanf.
|
Yeah, this seems right to me (subject to more testing). It's also getting simpler! Thanks. |
|
I'll take the opportunity and remove also a number of other format options that are not part of IRAF scanf:
|
|
Sounds good (without double checking that list). |
This is only a partial fix, because it doesn't work with integers that start with 0x or 0o. So, the following will not work correctly:
because the first conversion will eat up all (not only the first two digits). This seems however a quite strange case because it is f.e. unclear whether the 0x is valid also for the second part.
I am not even sure whether we shall support
0xor0oprefixes at all; they seem to be artefacts of the old parser. IRAF scanf does not support them.Closes #194