DBFtrim: MinEx0, MaxIntD, DecCount & pals, and differing sig figs

[yakra]

Relevant to how things got this way: https://github.com/yakra/tmtools/issues/6#issuecomment-370327880 https://github.com/yakra/tmtools/issues/23#issuecomment-372090341

field::GetMax cannot cope with input values with differing sig figs.

Suppose DecCount = 2, and Input values are 0.0 0.00 1.0 1.00 10.0 10.00 12.0 12.00 if the 1st value is 0.0, MinEx0 gets set to 1. This is still < DecCount, so MinEx0 is not increased to account for the '.' itself. if the 2nd value is 0.00, MinEx0 is already 1, so nothing more happens. Output values are 0. 0.0 1. 1.0 10. 10.0 12. 12.0

Simply detecting the '.' when trimming, rather than comparing MinEx0 against DecCount, will not work, come file writing time: outDBF.write(fVal, tDBF.fArr[fNum].len-fI+vlen-oDBF.fArr[fNum].MinEx0); resolves to outDBF.write(fVal, 2-2+3-3); or outDBF.write(fVal, 3-3+3-3); or outDBF.write(fVal, 2*(3-3));... You get the idea.

Goals:

• Preserve max number of decimal places before the minimum extraneous zeros; don't go reducing the number of output sig figs willy-nilly if there were uniform sig figs in original file.
• Cope with differing sig fig amounts in input.
• Trim as much as possible within these constraints. Never increase field size.
• Make sure DecCount stays sane.

new size = MaxIntD+1+MaxDecP or old size, whichever is smaller...

[yakra]

Pseudocode...

// Trimming...
    // bye-bye MinEx0

// Comparing...
if (tDBF.fArr[fNum].DecCount) tDBF.fArr[fNum].len = MaxIntD+1+tDBF.fArr[fNum].DecCount;
else if (tDBF.fArr[fNum].len < MaxIntD) tDBF.fArr[fNum].len = MaxIntD;
if (tDBF.fArr[fNum].len > len) tDBF.fArr[fNum].len = len;

// Saving...
if (vlen > strchr(fVal, '.') - fVal + tDBF.fArr[fNum].DecCount + 1)
    vlen = strchr(fVal, '.') - fVal + tDBF.fArr[fNum].DecCount + 1;
for (fI = tDBF.fArr[fNum].len-vlen; fI < tDBF.fArr[fNum].len; fI++) outDBF.put(' ');
outDBF.write(fVal, vlen);

[yakra]

~/gis/data/tests/dec-ints.dbf ~/gis/data/tests/dec-ints2.dbf ~/gis/data/tests/dec-ints3.dbf all behaving as expected

~/gis/data/tx/txdot-roads_48113/txdot-2015-roadways_48113.dbf only DIFF is DecCount in Shape_Leng field: now corrected from 11 to 10 (see https://github.com/yakra/tmtools/issues/6#issuecomment-370251373) Good!

everything else: no DIFF