Add pphtml tool

[windymilla]

Just putting a few notes here about pphtml to aid us in deciding what to do and how to link pphtml with Guiguts. Just thoughts that I didn't want to lose, rather than definite suggestions of things to do:

1. Current version of pphtml.py runs on the command line and can redirect output to a file.
2. Presumably a similar technique to that used for pptxt could be used, with StringIO objects passed to a pphtml function, and a small "main" section so it can continue to be run as a standalone script
3. Current version of pphtml puts some HTML in its output. The new version of pptxt either outputs plain text or decorated text (colored using ANSI terminal codes) but not HTML. It would be good for the revised pphtml to be able to do the same, i.e. HTML, plain or decorated.
4. If we are aiming to make the output of the two tools consistent with one another, and potentially with other tools yet to be written, like jeebies, should we consider whether any of the code in the tools ought to be common, particularly to support the 2/3 ways the output is formatted. Maybe something for the future - perhaps pphtml, pptxt, jeebies,... could/should live in a repo together with some shared routines, though I guess that would complicate their use as standalone scripts.

[cpeel]

Looking at pphtml.py there are a few changes we'll want to make to use it as a module where we can pass in data.

• We need to remove all exit()s and have it raise exceptions that can be caught.
• It uses the file command to detect the character set (see charsetCheck()) which won't work on a buffer and won't work on WIndows.
• It uses the input file's path to iterate over all images (self.sdir) so we'll want to pass that into the module instead.
• I think to support non-html output we just need to change the saveReport() post-processing. I'll give some thought about if/how we can standardize across tools

Since this is an existing tool we probably want to do some small iterative PRs with these improvements so we can test them individually. Thumbing through it I saw some other improvements we could make which also means we probably need some meager samples to test against to detect regressions along the way.

[windymilla]

Given the successful implementation of the functionality of pptxt in GG2, I think it will end up being best to do pphtml in a similar way.

[windymilla]

Suggestion: include the functionality of the ppvimage in the new pphtml.

[tangledhelix]

As noted in this forum thread, pphtml attempts to check for unprocessed ppgen commands in the HTML. I think it's a bit too naive as-is. It just checks for a <p> tag followed by a dot, and therefore it is triggered by a paragraph which begins with an ellipsis.

if ( $line =~ /<p>\./ ) {
    printf LOGFILE ( "%d:0 Possible PPG command: %s\n", $count, $line );
}

In the reimplementation it should be improved, I see one of two ways to do it:

1. The specific-exclusion route: add an and to the condition such that <p>\.\.\. doesn't match this check.
2. Or make it more aware of ppgen's command syntax: all ppgen commands should match <p>\.[a-z]{2} (a period followed by two lowercase letters)