LABEL_TOO_LONG and malformed URLs

[yakra]

The LABEL_TOO_LONG datacheck is performed in Route::read_wpt after the Waypoint ctor returns. Problem is, the Waypoint ctor performs the MALFORMED_URL, MALFORMED_LAT & MALFORMED_LON datachecks. If any of these 3 coexist with LABEL_TOO_LONG, they'll try to insert a label that's too long to fit into the DB.

• [x] This is already fixed in some pending speed/RAM improvements to the Waypoint ctor.
• [x] Refactoring Waypoint::label_too_long into the ctor will help code brevity. Will it run faster? Yes!

[yakra]

• [x] Python: work in a separate ~branch~ file, pending tests
• C++:
  • [x] hard reset back one commit (from 40c6a64d68451a383050c29581737e3eedd9cc91)
  • [x] mixed reset back another; discard Waypoint.cpp & siteupdate.py changes
• [x] Have a go at making C++ & Python produce identical datacheck.log files for super-long labels with multi-byte chars
  • [x] Does Break1ofEverything have sufficient test cases? ~No!~ Yes! New commit.
• [x] C++ bug: Single-char SINGLE_FIELD_LINE cases appear to not even make it as far as the Waypoint ctor to be flagged as such. Why?
• [x] Web repo updates for SINGLE_FIELD_LINE

[yakra]

Adding

/* DEBUG */
if (root == "me.i295")
{   std::ofstream dump295(Args::logfilepath+"/me.i295.dump");
    dump295.write(wptdata, wptdatasize);
    dump295.close();

    char lstr[2050];
    std::ofstream lines295(Args::logfilepath+"/me.i295.lines");
    for (char*l : lines)
    {   sprintf(lstr, "%X\n", int(l-wptdata));
        lines295 << lstr;
    }
    lines295.close();
}

to Route::read_wpt right before the // per-route datachecks reveals...

• me.i295.dump: The label is still in place, not zeroed out. The newlines on either side are zeroed out. All as expected.
• me.i295.lines is as expected:
  • 32 lines in this file, for 31 lines in the original me.i295.wpt + 1 dummy "past-the-end" element.
  • Line 9 shows 1FA, the location of the single-field exit 7.

---

Adding

/* DEBUG */
if (route->root == "me.i295")
    std::cout << "DEBUG: me.i295 @ " << line << std::endl;

to the Waypoint ctor between route = rte; and // split line into fields yields...

• 30 DEBUG: lines in siteupdate.log
• Exit 7 is missing. The ctor isn't called. What the...?

[yakra]

// process lines
for (unsigned int l = lines[1] < wptdata+2; l < lines.size()-1; l++)
{   // strip whitespace from end...     // l == 8 | lines[l] == 1FA | lines[l+1] == 1FC
    char* endchar = lines[l+1]-2;       // endchar == 1FA
    while (*endchar == 0) endchar--;    // *endchar == '7'; no change to endchar
    if (endchar <= lines[l]) continue;  // endchar == lines[l]; continue; here's the problem

Time to review the git blame and see when this was introduced and why, and what can be changed. Was it the "blank line at the beginning" bug? Yes.

[yakra]

Hopefully changing (endchar <= lines[l]) to (endchar < lines[l]) will work. I'll have to test this out on noreaster per #443.

[yakra]

Now we get:

[yakra@noreaster ~/ytm/DataProcessing/siteupdate/cplusplus]$ ~/diff _master/logs/datacheck.log _debug2/logs/datacheck.log 
1c1
< Log file created at: Tue Oct  5 14:27:05 2021
---
> Log file created at: Tue Oct  5 14:36:17 2021
1221a1222
> ausvic.c156;;;;SINGLE_FIELD_LINE;

An extra line at the end of the file contains a single space.