Hi!
Applied, thank you.
Thank you too!
How familiar are you with FPVectorial? This package is badly broken, and more contributions to fix the issues are very welcome.
Using it for a few weeks now (with long breaks in between) to draw diagrams of measurements, see
https://github.com/hansiglaser/pas-gpib/blob/master/examples/comparison/vectorialdiagram.pas, diagrams.pas, and to display it in a GUI (mainform.pas) and save SVGs to be included in a LaTeX report (report.pas).
I didn't have any real issues so far. Scaling and text (size&position) to be "beautiful" in the GUI and in SVGs doesn't work properly yet, but this is my own fault, because I have too many magic numbers around.
One problem is that the bounding box of the SVGs is off, often way off, but I don't understand enough. Fortunately, inkscape can use the real objects when converting to PDF.
Another drawback is that it relies on TCanvas to calculate text outlines, but I assume there is no better way. It is however a problem for my command line tool which just generates the SVGs, which adds a dependency which I still don't have provided.
Ah, and I tried TTvFormula to write "10
5" (with a true exponent), and this didn't work ("// this doesn't raise the exponent, it is written in the same line and same height as the base"), but perhaps I just didn't provide the Canvas early enough. I've resorted to hand crafted placement.
I couldn't find a single user of GenerateDebugTree, despite it being a very elaborate infrastructure behind.
The debug tree is displayed in an external application on ccr, fpvviewer (a lot of years ago, fpvectorial was available on ccr, and this is a remnant of it). You see the debug tree after loading an svg file and clicking on "View FPVectorial tokens". You can find the application at https://sourceforge.net/p/lazarus-ccr/svn/HEAD/tree/applications/fpvviewer/
Thanks for the info! Perhaps would be worth mentioning in a README file or (more explicitly) in the great Wiki page. I've cobbled together a tiny implementation which just executes WriteLn (in fpvwritetest.pas but didn't commit). Should I extract it to a separate debugtest.pas? I'll attach the diff, if you want to review my attempt.
Thanks
Hansi