two usage questions on `--user-wordlist`

[arrowtype]

Today is my first time trying out diffenator2. It is AWESOME. Really, it’s so useful. Thank you for making it!

I’m currently a little stuck on how to use custom wordlists. One thing I’d like to do is make a wordlist representing all the OT features in my font, and check those for each new version. So, I’ve started from the example in the readme, and updated that a bit:

a,latn,dflt,ss09
1/4,,,frac
10/23,,,frac
0123456789,,,numr,tnum,onum
DES,,,sups,ordn

This gives me two questions:

1. Where does the wordlist output go? My expectation (without digging too deep) is that it would generate a file called something like “user-wordlist-proof.html”, or that it might be put into diffbrowsers_proofer.html or diffenator.html. But, I’m not seeing it anywhere. Do the strings only appear if they are specifically different/changed between versions? This is my current guess, but I kind of wish there were also a proof output to confirm that items in the user wordlist are working as expected, as that alone would be helpful.

2. If I have a set of many words to proof for a given OT feature, do I have to repeat that line in the wordlist file for each of them? The answer in https://github.com/googlefonts/diffenator2/issues/69 seems to say so. But, in my example above, with the last line DES,,,sups,ordn, I’m really hoping to proof a set of ordinals. Ideally, I would want to proof a long string of them, ideally with a word spaced list, like the following: A C D E O R T Ú c d e r t ú AS DE DES DS EME ER ERE ERS ES GR IEME LLE ME ND OS RD RE RES ST STE TH as de des ds eme er ere ers es gr ieme lle me nd os rd re res st ste th. But, putting them onto one per line would be a bit inconvenient, especially because I have quite a few other features to also proof. Is there a better way? If not, that is understandable, and I’m partly just including this a user feedback.

Thanks again!

[m4rc1e]

Hey Stephen. Glad you are enjoying the tool and thank you for your kind words.

1. The output is in diffenator2-report.html under the heading "Misshapen user strings". If there's no results, it means it couldn't find any differences. I may add the full strings to diffbrowsers_text but I need to think about this.

2. do I have to repeat that line in the wordlist file for each of them?

I'm afraid so. Wordlists are delimited by a return, not a space of comma. We do this because we need to test multilingual strings where you may have one word in English and another in Hindi etc.

Thank you so much for the feedback. I plan to do another sprint soon based on everyone's frustrations. Please open as many issues as you like.

[arrowtype]

Hey Mark, thanks so much for the quick and thorough response! This is all helpful to know.

I understand that the idea of generating a full list of wordlists words might not fit with the main spirit of the project... finding diffs. Still, it would be handy, especially if marked as "no diffs," as a confirmation.

But, I can also understand if this is the wrong tool for that job. I could alternatively create a DrawBot script to proof OT features, or something similar.

Thanks also for welcoming feedback. I'll file more issues if I have them. Cheers!

[m4rc1e]

I understand that the idea of generating a full list of wordlists words might not fit with the main spirit of the project

We're using the --user-wordlist cmd with a custom wordlist which has over 400k words. These wordlists are usually generated from a python script though.

I'll add "no diffs found" if diff2 doesn't find any differences.

[arrowtype]

Hmm, could the proof subcommand instead perhaps be a place I can feed a custom list of things to proof? I haven’t tried it or dug into it, but that option isn’t in the readme, so I assume not... but that might be a place that custom proofing text could make sense!

[m4rc1e]

Totally, I think I'm leaning that way in https://github.com/googlefonts/diffenator2/issues/69#issuecomment-1620477412