On 8/19/20 5:34 PM, Pavel Sanda wrote: > On Tue, Aug 18, 2020 at 10:21:23PM -0400, Richard Kimberly Heck wrote: > >>> + if (is_bibtex_) { >>> + doi = operator[]("doi"); >>> + if (!doi.empty() && !prefixIs(doi,from_ascii("http"))) >>> + doi = "https://doi.org/" + doi; >>> + url = operator[]("url"); >> I'm confused about what this is for. Comment? > Not sure what line you refer to. A I meant the url variable. I guess all these things are returned to the caller, yes? Maybe a comment somewhere to that effect.+ file = operator[]("file"); >>> + >>> + // Jabref case, field has a format: >>> + // Description:Location:Filetype;Description:Location:Filetype... >>> + // We will grab only first pdf >> Since the second might be an absolute path, it might be worth a loop. >> Shouldn't be too hard. Just split on ; first and loop over the results. >> (Same with kbibtex.) > Makes sense. Do you prefer me to commit it now and continue working on updates > within master or wait for finished patch? (I won't have time to work on this > in next days).
I think it's fine to commit, and we can polish it as time permits. >>> + FileName f(to_utf8(filedest)); >>> + if (f.exists()) >>> + file = "file:///" + filedest; >>> + } >>> + >>> + // kbibtex case, format: >>> + // file1.pdf;file2.pdf >>> + // We will grab only first pdf >>> + docstring kfile; >>> + if (file.empty()) >>> + kfile = operator[]("localfile"); >>> + if (!kfile.empty()) { >>> + docstring filedest, tmp; >>> + tmp = split(kfile, filedest, ';'); >>> + //TODO howto deal with relative directories? >>> + FileName f(to_utf8(filedest)); >>> + if (f.exists()) >>> + file = "file:///" + filedest; >>> + } >> It occurred to me (and maybe I'll do this) that we could possibly have a >> configuration file for this that would look something like: >> >> localfile : ^([^;]+?) >> file ^:([^:]+?) >> >> I.e., for each possible field, we have a regex telling how to get the >> path we want. That could simplify the code, too. > Please feel free to adjust the code for files any way you want once committed. > I coded the whole biblatex + jabref/kbibtex part without testing because > I don't use these tools. It was nice surprise that it actually worked > for both Juergen and you ;) I suspect that there will be bunch of other > tools and other people will fill in those gaps later. Seems reasonable. Nice work on this. It's been a long time we've been trying to figure this out. Riki -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel