I think the reasons I originally omitted latex from Test2 were:
- I didn't know anything about latex at the time (I've since learned a bit).
- Running a full process to generate PDFs would require installation of textlive-full or its equivalent on the host machine, which was a requirement I was reluctant to incur.
- I believed at the time that the XSL:FO/FOP approach to generating PDFs would eventually take precedence.
However, it's clear that tex outputs are quite widely used, so I think it's time to add a single comprehensive test for it. Rather than actually generate the PDF (which would be an output that's hard to assess anyway), we should follow the Test approach, just generating the .tex file and diffing it. However, in contrast to Test, we should use a single comprehensive input file and a single output file.
I think the reasons I originally omitted latex from Test2 were:
However, it's clear that tex outputs are quite widely used, so I think it's time to add a single comprehensive test for it. Rather than actually generate the PDF (which would be an output that's hard to assess anyway), we should follow the Test approach, just generating the .tex file and diffing it. However, in contrast to Test, we should use a single comprehensive input file and a single output file.